0% found this document useful (0 votes)

513 views477 pages

Seismic Design of Foundations

Uploaded by

Jorge Armando Anyosa Doza

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

513 views477 pages

Seismic Design of Foundations

Uploaded by

Jorge Armando Anyosa Doza

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 477

Seismic Design of Foundations

Seismic Design of Foundations
Seismic Design of Foundations
Concepts and applications

Subhamoy Bhattacharya
University of Surrey, United Kingdom

Rolando P Orense
The University of Auckland, New Zealand

Domenico Lombardi
The University of Manchester, United Kingdom

Seismic Design of Foundations

Published by ICE Publishing, One Great George Street, Westminster, London SW1P
3AA.

Full details of ICE Publishing representatives and distributors can be found at:
www.icebookshop.com/bookshop_contact.asp

Other titles available from ICE Publishing:

Earthquake Design Practice for Buildings, Third edition.
Edmund Dwight Booth. ISBN 978-0-7277-5794-4
Designers’ Guide to Eurocode 8: Design of bridges for earthquake resistance
Basil Kolias, Michael N. Fardis and Alain Pecker. ISBN 978-0-7277-5735-7
Shock Transmission Units in Construction
Dinesh J. Patel. ISBN 978-0-7277-5713-5
www.icebookshop.com

A catalogue record for this book is available from the British Library

ISBN 978-0-7277-6166-8

© Thomas Telford Limited 2019

ICE Publishing is a division of Thomas Telford Ltd, a wholly-owned subsidiary of the

Institution of Civil Engineers (ICE).

All rights, including translation, reserved. Except as permitted by the Copyright,

Designs and Patents Act 1988, no part of this publication may be reproduced, stored
in a retrieval system or transmitted in any form or by any means, electronic, mechanical,
photocopying or otherwise, without the prior written permission of the Publisher,
ICE Publishing, One Great George Street, Westminster, London SW1P 3AA.

This book is published on the understanding that the author is solely responsible for
the statements made and opinions expressed in it and that its publication does not
necessarily imply that such statements and/or opinions are or reﬂect the views or
opinions of the publishers. While every effort has been made to ensure that the
statements made and the opinions expressed in this publication provide a safe and
accurate guide, no liability or responsibility can be accepted in this respect by the
author or publishers.

While every reasonable effort has been undertaken by the author and the publisher
to acknowledge copyright on material reproduced, if there has been an oversight
please contact the publisher and we will endeavour to correct this upon a reprint.

Cover photo: Kobe, Japan; March 2009. Preserved crater from 1995 Great Hanshin
Earthquake.

Commissioning Editor: James Hobbs

Production Editor: Madhubanti Bhattacharyya
Marketing Specialist: April Asta Brodie

Typeset by The Manila Typesetting Company

Index created by Laurence Errington
Printed and bound in Great Britain by TJ International, Padstow

Seismic Design of Foundations

Contents Preface
Acknowledgements
ix
xi

01 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction to earthquake geotechnical engineering in

relation to foundation design 1
1.1. Introduction 1
1.2. Overview of earthquake engineering 1
1.3. What does earthquake geotechnical engineering cover? 1
1.4. Examples of geotechnical damage 4
1.5. Major topics in earthquake geotechnical engineering 4
1.6. Historical perspective: how engineers learnt from earthquakes 13
1.7. Complexity of earthquake geotechnical engineering 29
1.8. Summary 32
References 32
02 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Basic concepts of engineering seismology and seismic hazard
analysis 33
2.1. Introduction 33
2.2. Theories of plate tectonics and elastic rebound 33
2.3. Types of faults 34
2.4. Evaluation of earthquake size 35
2.5. Hazards for the seismic design of foundations 38
2.6. Seismic hazard analysis 39
2.7. Case study: damage to the outlet tunnel of the second
Kakkonda hydropower station (This case study is based on
the seminal work of Prof Konagai, see for further details
Konagai (2005), Johansson and Konagai (2007)) 44
2.8. Case study: quantification of the PGA for locations in India
where strong-motion records are unavailable 46
2.9. Quantification of fault movement 56
2.10. Case study: hazard to pipelines due to permanent ground
deformations 56
2.11. Quantification of the duration of an earthquake 60
2.12. Summary 60
References 60
03 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selection of strong motion for foundation design 63
3.1. Introduction 63
3.2. Background to the problem 63
3.3. Criteria for the selection of earthquake records 65
3.4. Methods for spectral matching 67
3.5. Case study: application of spectrum-compatible input motion 69
3.6. Summary 77
References 77

04 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ground response analysis 79

4.1. Introduction 79
4.2. Seismic wave propagation in layered media 81
4.3. Methods of ground response analysis 83
4.4. Linear ground response analysis procedure 84

Seismic Design of Foundations

4.5. Equivalent linear and non-linear programs for ground
response analysis 94
4.6. Simpliﬁed method to obtain the natural period of the ground 101
4.7. Natural period of the ground for a layered soil 101
4.8. Summary 102
References 102

05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seismic analysis methods related to foundation design 103

5.1. Introduction 103
5.2. Fundamental period of structures 103
5.3. Fundamentals modal analysis 104
5.4. The concept of response spectrum 106
5.5. Application of response spectra analysis in foundation design 110
5.6. Example calculations for the base shear of a three-storey building 115
5.7. Essentials of performance-based design 119
5.8. Beam on a non-linear Winkler foundation (SSI) 125
5.9. Summary 136
References 138

06 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Liquefaction: theoretical aspects 141

6.1. Introduction 141
6.2. Concept of effective stress and dilatancy 141
6.3. Mechanism of pore pressure development 143
6.4. Laboratory undrained tests 145
6.5. Role of numerical modelling 152
6.6. Actual earthquake records as manifestations of in situ soil
liquefaction 161
6.7. Effects on the built environment 164
6.8. Summary 168
References 170

07 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Liquefaction: practical aspects 173

7.1. Introduction 173
7.2. Liquefaction susceptibility 173
7.3. Liquefaction triggering 178
7.4. Evaluation of the effects of liquefaction 187
7.5. Codes of practice-speciﬁed methods 190
7.6. Liquefaction-induced ground displacements 194
7.7. Worked example 202
7.8. Planning and design of a major project in seismic areas 207
7.9. Summary 210
References 210

08 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis and design of shallow foundations 215

8.1. Introduction 215
8.2. Seismic bearing capacity of shallow foundations 215
8.3. Case studies: observed settlements of buildings during
earthquakes 217
8.4. Case studies: settlements of footings in physical model tests 223

Seismic Design of Foundations

8.5. Factors affecting settlements of shallow foundations in
liqueﬁed ground 230
8.6. Estimation of total and differential settlements 233
8.7. Seismic design of mat foundations 238
8.8. Summary 238
References 238

09 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pile foundations 241

9.1. Introduction 241
9.2. Why piled foundation are used 241
9.3. Pile foundations in seismic areas 241
9.4. Failure mechanisms of pile-supported structures 245
9.5. Criteria for the design of piles in seismic areas 253
9.6. Design checks and calculations required 255
9.7. Design codes of practice 261
9.8. Modelling of the problem 263
9.9. Raked or battered piles 282
9.10. Example: pile-supported LNG tanks 285
9.11. Example: Kandla Port tower building on piles 290
9.12. Summary 291
References 291

10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis of foundations for major bridges 297

10.1. Introduction 297
10.2. Survey of the performance of bridge foundations in
seismic areas 297
10.3. Bridge foundations 305
10.4. Caisson foundations 306
10.5. Stiffness of deep foundations 308
10.6. Methodology for fundamental frequency estimations 316
10.7. Dynamic behaviour of pile-supported bridges 320
10.8. Performance comparison of bridges supported on pile and
caisson foundations in liqueﬁable soils 328
10.9. Summary 329
References 329

11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Foundations in slopes and for retaining walls 331

11.1. Introduction 331
11.2. Concept of seismic coefﬁcients for slopes and retaining walls 331
11.3. Foundation considerations for buildings on slopes 334
11.4. Case studies: seismic performance of buildings on slopes 336
11.5. Earth pressure theory 341
11.6. Foundations for retaining walls 347
11.7. Worked example 350
11.8. Case studies: seismic performance of retaining walls 354
11.9. Summary 362
References 362

vii

Seismic Design of Foundations

12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Liquefaction countermeasures 365
12.1. Introduction 365
12.2. Liquefaction countermeasures 365
12.3. Ground improvement methods 369
12.4. Analysis and design of typical ground improvement methods 377
12.5. Worked examples for typical methods 386
12.6. Case studies: ground improvement method application 389
12.7. Case studies: seismic performance of improved-ground sites 395
12.8. Methods to strengthen shallow foundations 409
12.9. Methods to strengthen pile foundations 414
12.10. Summary 417
References 417

Appendix ........................ Engineering correlations for the design of foundations 421

Index 451

viii

Seismic Design of Foundations

Preface In order to carry out design of foundations in seismically active areas,
one requires some level of expertise in seismic hazard assessment, soil
behaviour under cyclic loading, structural dynamics and soil-structure
interaction, in addition to standard foundation analysis and design concepts.
As a result, this subject initially emerged as a specialist area for only
a few researchers as well as a handful of professional engineering
consultants, although over the years, it has gradually appeared on
specialist undergraduate and postgraduate courses at many universities.

The ﬂow (that is, the structure) of the book follows a pattern that would
be typical of a design process: assessment and quantiﬁcation of seismic
hazards that would affect foundation design (Chapters 2 to 7), followed
by design and mitigation aspects. The design of foundations requires a
fundamental understanding of probabilistic concepts of seismic hazards,
ground response analysis and structural analysis, and these subjects can
themselves be the subjects of an entire book. Therefore, only the necessary
analysis and design concepts – including, but not limited to, response
spectra, modal analysis, beam on Winkler foundations, and ground response
analysis – are covered in brief, and only the fundamentals are explained,
along with some examples. Wherever possible, references are provided for
further reading. The authors have sincerely attempted to make the book as
useful and practicable as possible by providing case studies, example
applications, and solved examples from the real world.

The 11th March 2011 Tōhoku earthquake (Japan) was a rare and a landmark
event which can be described as an act of war between nature and earthquake
engineers. It is fair to say that if the effects of the tsunami are discounted,
the overall performance of the structures under the action of earthquake
(together with liquefaction) shows that there is a lot to learn from Japanese
earthquake geotechnical engineers. The ﬁrst author (Professor Bhattacharya)
was in Japan on that day and had ﬁrst-hand experience of the performance of
these structures. Following the 2010–2011 Canterbury earthquake sequence
(New Zealand), the second author (Associate Professor Orense) worked
over a period of more than 18 months to investigate the impact of ground
shaking and soil liquefaction on the built environment of Christchurch
and surrounding areas. This book heavily draws upon the experiences of
these two recent earthquakes, as well as from lessons learned in other recent
large-scale earthquakes.

The book is born out of research by the authors as well as postgraduate

and CPD (Continuing Professional Development) teaching and specialist
lectures they delivered at various universities and organisations. Examples
include: Professor Bhattacharya’s courses on earthquake engineering
at the University of Surrey (MSc course) and University of Bristol
(MEng course) in the UK, Indian Institutes of Technologies (IIT Kharagpur

Seismic Design of Foundations

and IIT Bhubaneswar as a part of ISWT (International Summer and
Winter Term) and GIAN (Global Initiative for Academic Networks) courses
respectively), ASRANet CPD courses in Delft, London and Brussels as well
as at Yamaguchi University and Tokyo Institute of Technology (Japan),
Zhejiang University (China), Tongji University (China) and Harbin Institute
of Technology (China). Associate Professor Orense (second author)
provided lectures on the same topic at the University of Auckland
(New Zealand), University of Tokyo and Yamaguchi University (Japan),
the University of the Philippines and through technical workshops and
seminars sponsored by the Engineering New Zealand (EngNZ) held in
various cities in New Zealand.

It is hoped that this book will be ideal for any student studying postgraduate
geotechnical engineering or structural engineering, as well as researchers
and practitioners working in the field of earthquake geotechnical
engineering. Due to the changing requirements in the current technological
age and market demands of the modern educational system, ICE Publishing
wishes to make individual chapters of the book available to readers through
its e-book platform ICE Virtual Library. Therefore, to make each chapter
self-sufficient, there are some inevitable repetitions of some figures and
textual analysis.

This ﬁeld of engineering is constantly evolving as each major earthquake

teaches us something new. Many of the methodologies or calculation
procedures presented here may become outdated or be modiﬁed in the
next few years, at which point this book may require a new edition.
There may also be errors and omissions in the book and we would be
grateful if those were brought to our attention. Please email the authors
at [email protected] or [email protected] and
the comments will be duly acknowledged in the next edition.

Subhamoy Bhattacharya
Rolando P Orense
Domenico Lombardi

Seismic Design of Foundations

Acknowledgements The authors are grateful for the contributions of the many people who
helped make this book a reality. Individually and collectively, they
beneﬁtted from close association with various researchers in drawing
experiences for this book: Gopal Madabhushi and Malcolm Bolton
(University of Cambridge), Tony Blakeborough, Harvey Burd and
Martin Williams (University of Oxford), Kazuo Konagai (Institute of
Industrial Science, Tokyo), Jorgen Johansson and Amir Kaynia (NGI),
Nick Alexander, George Mylonakis, Colin Taylor, D Muir Wood,
Erdin Ibraim (University of Bristol), Masayuki Hyodo and Gakuo Watanabe
(Yamaguchi University), George Nikitas, Stefan Szyniszewski and
Stergios Mitoulis (University of Surrey), Sudhir Jain (IIT Gandhinagar),
Vinay Gupta, Samit Roychowdhury, Prishati Roychowdhury (IIT Kanpur),
Dilip Baidya and Nilanjan Mitra (IIT Kharagpur), Indrajit Chowdhury
(ex-Petrofac), Murali Krishna, Pradeep Dammala and Kaustubh Dasgupta
(IIT Guwahati), T Sitharam and GLS Sivakumar Babu (IISc Bangalore),
L Gobindraju (NIT, India), Sekhar Dutta, Rajib Sarkar and Sankey Nayak
(IIT Dhanbad), Armando Simonelli and Luigi Di Sarno (University of
Sannio, Italy), Kenji Ishihara (Chuo University), Ikuo Towhata (ex-University
of Tokyo), Kohji Tokimatsu (ex-Tokyo Institute of Technology),
Deepankar Chowdhury (IIT Mumbai), Nozomu Yoshida (Tohoku Gakuin
University), Mehdi Rouholamin (University of Portsmouth), Josh Macabuag
(SCOR, London), Suresh Dash and Sumanta Halder (IIT Bhubaneswar),
A Boominathan (IIT Madras), Rajiv Saha (NIT, Agartala), Katsu Goda
(University of Western Ontario), Barnali Ghosh (Motts, UK), Ziggy Lubkowski
(Arup, London), Fumio Tatsuoka (Tokyo University of Science),
Susumu Yasuda (Tokyo Denki University), Susumu Iai (ex-Kyoto
University), Junichi Koseki (University of Tokyo), Kenji Harada
(Fudo Tetra Corporation), Tao Wang (IEM, Beijing), Yu Huang
(Tongji University), the late Takashi Tazoh (Shimizu Corporation),
Kenichi Soga (UC Berkeley), Andrew Chan (University of Tasmania),
Michael Pender, Liam Wotherspoon and Tam Larkin (University of
Auckland), Sjoerd van Balleegooy (Tonkin & Taylor), Misko Cubrinovski
and Brendon Bradley (University of Canterbury), Russell Green (Virginia
Tech), Jonathan Bray (University of California Berkeley), Brady Cox and
Ken Stokoe (University of Texas, Austin), Ahmed Elgamal (University of
California, San Diego), Ross Boulanger (University of California, Davis),
George Gazetas (NTU, Athens), Nitish Puri (NIT, India), Hironori Suzuki
(Shimizu Corporation), and Amar Bouzid Djillai (Algeria).

Moreover, they would like to collectively acknowledge their close

collaboration with various research institutes, funding bodies and
organisations: Cambridge Commonwealth Trust, CVCP (Committee
of Vice Chancellors and Principals for ORS Fellowship), Nehru Trust
for Cambridge University, Junior Research Fellowship of Somerville
College (Oxford University), University Research Fellowship

Seismic Design of Foundations

(University of Bristol) NICEE and NPEEE (IIT Kanpur), and EERC
(Earthquake Engineering Research Centre, University of Bristol), RELUIS
project (Department of Civil Protection, Italian Government), SERIES
Project (REA, European Commission), EPSRC (Engineering and Physical
Science Research Council, UK), Royal Society (UK), NSFC (National
Science Foundation China), Commonwealth Spilt-Site Fellowship,
Commonwealth PhD fellowship, Newton Funds (Newton Bhabha
fellowship through British Council), UKIERI (UK India Education and
Research Initiative) project, Shimizu Corporation (Japan), Hokiroku
Corporation (Japan), VJTech (UK), BLADE (Bristol Laboratories for
Advanced Dynamics Engineering) facilities, ARUP (UK), Motts (UK),
GEER (USA), ISET (Indian Society For Earthquake Technologists), IGS
(Indian Geotechnical Society), Japanese Geotechnical Society (JGS),
New Zealand Geotechnical Society (NZGS), Institution of Civil Engineers
(ICE), Institution of Structural Engineers (IStructE), Kiso-jiban Consultants
Ltd., Japan Home Shield Corp. and Fudo Tetra Corp.. They would also like
to thank their current and former postgraduate students who, in one or
another, contributed to this book. Special thanks to current PhD students
Hasan Emre Demirci, Rajesh Rele and Piyush Mohanty.

Finally, they would like to thank Amber Thomas (ex-ICE), James Hobbs,
Madhubanti Bhattacharyya, and others on the ICE Publishing editorial staff
who made this book possible and reﬁned the rough draft into this ﬁnished
product.

xii

Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

ISBN 978-0-7277-6166-8
https://doi.org/10.1680/sdof.61668.001
ICE Publishing: All rights reserved

Chapter 1
Introduction to earthquake geotechnical
engineering in relation to foundation
design
1.1. Introduction
Earthquakes cause damage to engineering structures and often result in loss of life. Forecasting the exact time
of an earthquake can at best reduce casualties, and at present also appears to be an impossible task. Therefore,
structures need to be designed to withstand the impact of an earthquake and prevent collapse, as it is buildings
and soil structure (dams, embankments, retaining walls) that kill people, not earthquakes. This chapter
provides a brief overview of the ﬁeld of earthquake geotechnical engineering; that is, where the ground is the
main contributor to the damage.

1.2. Overview of earthquake engineering

Earthquakes cause damage to the built environment and result in fatalities and financial loss. During earth-
quakes, structures may collapse or be severely damaged as a result of structural inadequacies, foundation
collapse/damage, ground failure or, more frequently, a combination of these effects. In addition, earthquakes
may trigger a series of cascading catastrophic events (e.g. tsunami, fire, landslide, debris flow, flooding and
nuclear contamination) that may lead to shortage of food and water. Furthermore, recent seismic events have
shown that secondary events such as tsunami or fire are responsible for the vast majority of casualties and
substantial economic damage. The impact of earthquakes in terms of fatalities and monetary loss may exhibit
significant disparities between different geographical regions of the world, as may be observed from Figure 1.1.
This graph is based on the work of Goda and Tesfamariam (2015), who showed that the impact of earthquakes
in developed countries results in relatively few fatalities but may have considerable monetary loss. In contrast,
emerging economies (in South America, Asia and Africa) had higher fatalities but lower monetary loss.
However, two exceptions can be noted from Figure 1.1: the 1995 Kobe and 2011 Tōhoku earthquakes, both in
Japan, in which the post-earthquake fire and tsunami, respectively, were responsible for the high fatality rate –
not the earthquake itself. It is therefore evident that a comprehensive assessment of the impact of earthquakes on
infrastructure and people requires a multidisciplinary approach and professionals with different expertise.

This book covers the fundamental principles of seismic design practice where foundations and ground failure
are major concerns. This specialised area is known as earthquake geotechnical engineering. The book is
intended for use in undergraduate and postgraduate courses where design practice is combined with theory,
but can also be used by practising engineers as a reference book.

1.3. What does earthquake geotechnical engineering cover?

Normally, the failure of structures during earthquakes is the result of structural inadequacies or foundation
failure, or a combination of both. Figure 1.2(a) shows the failure of a residential building predominantly due to
structural inadequacies such as poor ductility and improper beam–column detailing. On the other hand, the
failure shown in Figure 1.2(b) and 1.2(c) is not due to structural inadequacies but due to foundation failure.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

In such failures, the soil supporting the foundation plays an important role. The behaviour of foundations
during earthquakes is often dictated by the response of their supporting soil due to the ground shaking. This
subject covers those aspects of earthquake geotechnical engineering where soil is the major cause of damage.

Figure 1.1 Total fatality versus monetary loss for worldwide earthquakes for 1980–2011 (Goda and Tesfamariam,
2015)

106
2011 Tohoku
1995 Kobe
105
Monetary loss: millions US $

South America
104 Africa
Asia
Turkey
103
USA
Europe
102 New Zealand/Australia
Japan
101

100
100 101 102 103 104 105 106
Total fatalities

Figure 1.2 (a) Failure of a residential building during the 2001 Bhuj earthquake, India (courtesy of the National
Information Centre of Earthquake Engineering – NICEE). (b) Complete uproot of a building during the 2011 Tōhoku
earthquake, Japan. (c) Example of manhole uplift failure caused by soil liquefaction

(a)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Introduction to earthquake geotechnical engineering in relation to foundation design

Figure 1.2 (Continued)

(b)

(c)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

1.4. Examples of geotechnical damage

Some examples of geotechnical damage to the built environment are

■ Failures of earth structures such as dams, embankments, landfill and waste disposal sites. Failure of
dams often cause floods. During the 2001 Bhuj earthquake (India), four dams – Fatehgarh, Kaswati,
Suvi and Tapar – suffered severe damage. The reservoirs were empty at the time of the earthquake
and there were no reported casualties and the economical loses were limited. Figure 1.3(a) shows the
longitudinal cracks at Fatehgarh dam following the 2001 Bhuj earthquake.
■ Widespread destruction to road networks and foundations from ground liquefaction – when the soil
momentarily becomes a fluid-like material (change in state from solid to fluid). It is often observed
that raft and piled foundations collapse in liquefied ground without any damage to the superstructure:
see Figure 1.3(b). An example of damage to a road network can be seen in Figure 1.3(c). Further
effects of liquefaction are shown in Figures 1.3(d) and 1.3(e). Chapters 6 and 7 covers various aspects
of the liquefaction.
■ Damage to underground or buried structures such as tunnels, box culverts, underground storage
facilities and pipelines (Figure 1.3f). During the 1999 Turkey earthquake, the Bolu tunnel suffered
severe damage.
■ Damage to foundations due to large ground displacements owing to liquefaction-induced lateral
spreading or fault displacement (Figures 1.3(h)–1.3(k)). Such foundations include underground
retaining walls, quay walls and pile-supported wharfs.
■ Damage to foundations due to fault movement. An example is shown in Figure 1.3(g), where a bridge
collapsed due to fault movement. Fault movement can be particularly damaging to long gas or oil
pipelines spanning continents.
■ Overturning/uprooting of buildings due to the combined effects of a tsunami and an earthquake.
Figure 1.3(l) shows an uprooted building in Onagawa, Japan, supported on pile foundations that
collapsed following the tsunami induced by the Tōhoku earthquake.
■ Severe damage to infrastructure from landslides – permanent ground deformations often triggered by
earthquakes. Figure 1.3(m) shows the landslide-induced failure of Baihua Bridge (China) during the
Wenchuan earthquake (2008).

1.5. Major topics in earthquake geotechnical engineering

Earthquake geotechnical engineering encompasses the following major topics.

Wave propagation. Seismic waves are generated during fault rupture, and this process produces mainly two
types of waves: (a) secondary (S) waves and; (b) primary (P) waves. S waves are slower but often more
destructive. They shake the soil from side to side and generate inertial force on the structure. These waves are
responsible for the onset of soil liquefaction wherein the soil loses much of its strength and stiffness,
behaving like a heavy viscous ﬂuid, see Chapter 6. Typical S wave velocities are given in Table 1.1. In
contrast, P waves are faster and are transmitted by compression and rarefaction, very similar to sound waves.
Typical velocities of P waves in soft soils are 1400–1500 m/s. At ground level, P waves and S waves produce
surface waves (Rayleigh waves and Love waves) that often crack window glasses.

Figure 1.3 Examples of geotechnical damage. (a) Longitudinal cracks in Fatehgarh Dam after the 2001 Bhuj
earthquake, India (by courtesy of NICEE). Examples of liquefaction-induced damage: (b) tilting of raft foundations
(by courtesy of JGS); (c) subsidence of a road during the 2016 Kumamoto earthquake, Japan. Liquefied soil effects
during the 2011 Tōhoku earthquake, Japan: (d) car completely swallowed; (e) manhole uplifted. (f) Lifting of a
manhole in a Japanese earthquake. (g) Collapse of a bridge due to fault movement (by courtesy of the Earthquake
Engineering Research Institute). (h) Subsidence due to fault movement close to buildings during the 2016
Kumamoto earthquake. (i) Permanent ground displacement causing a railway line to buckle (image courtesy Pacific
Press Service/Alamy Stock Photo). (j) Severe damage to the quay wall in Navalakhi Port (2001 Bhuj earthquake)
(by courtesy of NICEE). (k) Permanent displacement of the deck of Surajbari Bridge, possibly due to the tilting of the
bridge pier (by courtesy of NICEE). (l) Completely uprooted building in Onagawa city during the Tōhoku earthquake.
(m) Baihua Bridge collapsed during the 2008 Great Wenchuan earthquake, China, as a result of structural
inadequacy, exacerbated by the damage of several piers due to large soil movement and debris flow (image
courtesy Prof Yu Huang)

(a)

(b)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 1.3 (Continued)

(c)

(d)

Figure 1.3 (Continued)

(e)

(f)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 1.3 (Continued)

(g)

1.5 m

(h)

Figure 1.3 (Continued)

(i)

(j)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 1.3 (Continued)

(k)

(l)

Figure 1.3 (Continued)

(m)

Table 1.1 Typical velocity of S waves (Vs)

Ground type Vs: m/s

Very soft soil 100

Stiff soil 300

Hard rock 3000

Fault movement and its destructive action. The effect of fault movement can be very destructive if there are
structures passing through these boundaries, as observed during the 1999 Chi Chi (Taiwan) and the 1999
Kocaeli (Turkey) earthquakes. Faults are usually deﬁned as a form of discontinuity in the bedrock and are
associated with the relative displacement of two large blocks of rock masses. Faults are broadly subdivided
into three categories depending on their relative movement

■ Normal fault: one block (often termed the ‘hanging wall block’) moves down relative to the other
block (often termed the ‘footwall block’). The fault plane usually makes a high angle with the surface
(Figure 1.4(a)).
■ Reverse fault: the hanging wall block moves up relative to the footwall block. The fault plane usually
makes a low angle with the surface (Figure 1.4(b)).
■ Strike-slip fault: the two blocks move either to the left or to the right relative to one another (Figure 1.4(c)).

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 1.4 (a) Normal fault, (b) reverse fault and (c) strike-slip fault

Fault plane Compressive Shear force

force

Tensile force Fault plane

(b) (c)
(a) Fault plane

The damage caused by the fault movement can be destructive: see, for example, Figure 1.3(g), where the
bridge was located near the fault boundary. Tunnels and pipelines (if crossing a fault line) may become
damaged during an earthquake.

Prediction of ground motion characteristics at the site. Peak ground acceleration (PGA) is often required
by designers to evaluate the response of a structure due to seismic shaking. The motion at a site is a function of
various parameters such as the geological formation of the site, location of the fault, dimension of the fault
and the source mechanism. This is primarily dealt by seismologists. Geotechnical or structural engineers are
more speciﬁcally interested in the magnitude and frequency of the earthquake motion – which are required to
undertake ground response studies (covered in detail in Chapters 3 and 4).

Prediction of the duration of the earthquake motion at a particular site. Earthquakes produce cyclic
loading and may cause loose to medium-dense sand to liquefy (i.e. the soil behaves like quick sand). The
behaviour of soil under cyclic loading depends not only on the magnitude and duration but also on the number
of cycles of loading. Certain type of hazards may occur if the soil remains liqueﬁed for a longer duration. It
may thus be important to predict the duration of the earthquake motion or the number of cycles of loading
while evaluating the response of a structure. Details about ground motion selection for foundation design will
be provided in Chapter 3.

Prediction of the cyclic behaviour of the soil at the site. Different soils behave differently under the action
of cyclic loading. One of the main challenges is to predict whether or not liquefaction will be triggered at a
site. Normally, loose to medium-dense sandy soils liquefy under moderate shaking. If liquefaction were to
happen at that site, to what depth would it liquefy? If the ground is composed of clay, how would it behave
under the action of cyclic loading? For example, it was observed during the Mexico City earthquake that the
earthquake motion was severely ampliﬁed by the soil. Prediction of the ground response at a given site is
covered in Chapter 4.

Generation and dissipation of pore water pressure. Moderate shaking generates pore water pressure in
most sandy soils. In some cases, the pore pressure developed may be adequate to cause liquefaction, but in
any event it would certainly decrease the effective stress of the soil. As most of the soil properties are related
to effective stress, it is clear that the foundation response will be affected. In short, an earthquake generates
excess pore pressure, and when the shaking stops, the pore pressure has to dissipate. As the pore pressure
dissipates, the soil regains its strength. The concepts and theory of pore pressure development and dissipation
will be discussed in Chapter 6.

Prediction of permanent ground deformation after the earthquake. Soil liquefaction often leads to
permanent lateral ground displacements in sloping grounds, commonly referred to as ﬂow failure or lateral
spread. Flow failures are associated with very large surface displacements of the order of hundreds of metres.
Conversely, lateral spreading displacements are typically of the order of centimetres or metres. It is often
necessary to predict the ground displacement that will be induced by liquefaction.

Post-liquefied strength of soil for stability analysis of various structures. After liquefaction, the stiffness
of the liquefied soil drops to a near-zero value. It is often necessary to assess the stability of structures
supported on such soils. The liquefied and post-liquefied strength of soils for practical aspects is discussed in
Chapter 7.

Seismic earth pressure. Earthquake geotechnical engineers often have to compute the seismic earth pres-
sures acting on foundations, retaining walls, abutments and dams.

1.6. Historical perspective: how engineers learnt from earthquakes

Earthquakes in the past have shown the shortcomings of current design methodologies and construction
practices, at the cost of structural failures and loss of lives. Post-earthquake investigations have led to
improvements in engineering analysis, design and construction practices. A brief historical development of
earthquake engineering practice showing how earthquake engineers have learned from failures in the past is
outlined in Table 1.2.

Table 1.2 Historical development of earthquake engineering practice and major lessons learnt

Earthquake Remarks Post-earthquake developments

63 AD Campania The philosopher Seneca witnessed the The philosopher attributed the origin of
earthquake (Italy) earthquake and narrated his account of earthquakes to underground movements of
the events in his manuscript Naturales water and trapped air in the Earth’s crust.
Quaestiones.

1707 Hōei A large tsunami killed thousands of Stress changes induced by the fault rupture
earthquake (Japan) people. may have been responsible for subsequent
volcanic eruptions, notably the last ever
recorded eruption of Mount Fuji in 1707.

1755 Lisbon The earthquake promoted debate among Systematic observations and interviews
earthquake European philosophers, including Voltaire, carried out in the aftermath of the earth-
(Portugal) Rousseau and Kant. At the time, the origin quake allowed the epicentral intensity of the
of earthquakes was uncertain. Philoso- earthquake to be determined, signalling the
phers such as Kant attributed earthquakes birth of modern seismology.
to circulation of hot gas in the crust of the
Earth. In contrast, theologians claimed that
earthquakes were a punishment by God
for the sins of the world.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Table 1.2 (Continued)

Earthquake Remarks Post-earthquake developments

1906 San Francisco A high concentration of damage was Development of earthquake-resistant

earthquake (USA) observed in the reclamation area in San construction initiated.
Francisco Bay. Most of the damage to
buildings was caused by subsequent ﬁres
that could not be promptly extinguished
owing to broken water mains.

1908 Reggio 120 000 fatalities. A committee of nine The base shear equation was derived,
Messina earthquake practising engineers and ﬁve professors were i.e. the lateral force exerted on the structure
(Italy) appointed by the Italian government to study is some percentage of the dead weight of
the failures and to set design guidelines. the structure (typically 5–15%).

1923 Kanto Destruction of bridges and buildings. The seismic coefficient method (equivalent
earthquake (Japan) Foundations settled, tilted and moved. static force method using a seismic
coefficient of 0.1–0.3) was first used in
the design of highway bridges in Japan.
Essentially, the peak lateral inertial load on
the structure were taken as 10–30% of the
dead weight.

1933 Long Beach Destruction of buildings, especially school This is the ﬁrst earthquake for which
earthquake (USA) buildings. acceleration records were obtained from
the recently developed strong motion
accelerograph.

1964 Niigata The ground can also be a major Soil liquefaction studies started. Seed
earthquake (Japan) contributor of damage. Extensive damage and Lee (1966) published the ﬁrst paper
due to liquefaction was observed. For on the liquefaction behaviour of sand
example, the newly built Showa Bridge using cyclic triaxial and cyclic shear tests.
(just 15 days after construction) collapsed Procedures for estimating the depth of
and middle spans of the bridge fell into liquefaction were established by Seed and
the river. This earthquake is very well Idriss (1971), which were subsequently
researched in the earthquake engineering updated and adopted in codes of practice
community and is discussed in detail in worldwide.
Chapter 10.

1971 San Fernando Bridges collapsed and dams failed, causing Liquefaction studies intensified. Bridge
earthquake (USA) flood. retrofit studies started. Analysis of dams due
to earthquake effects was initiated.

Table 1.2 (Continued)

Earthquake Remarks Post-earthquake developments

1985 Mexico City Building damage and landslides were Seismic ground response analysis studies
earthquake (Mexico) widespread throughout Mexico although intensified and were subsequently
most of the effects were felt in Mexico City, translated into codes of practice.
350 km from the epicentre. The worst This earthquake showed that the ground
damage within Mexico City occurred to can become tuned to an earthquake,
buildings located on a former lake bed, causing amplification of the motion.
where soft sediments were present.
Buildings with 5–15 stories suffered
greater damage due to the natural period
of the ground being close to that of the
input ground motion at the site. These
observations prompted engineers to con-
sider the effects of local site conditions on
ground motion. This is a ‘resonance’-type
effect of soft soil with the ground motion.
Effectively, the period of the ground
matched the predominant period of the
earthquake, which resulted in amplifica-
tion of the ground motion.

1994 Northridge Steel connections failed in bridges. The importance of ductility in construction
earthquake (USA) realised and implemented.

1995 Kobe Foundation failure observed together with Lateral spreading is believed to be one of the
earthquake (Japan) lateral spreading (downward movement of main causes of foundation failure. Japanese
a slope) of the ground. Good performance standards such as the Japanese Road Asso-
of seismic isolated bridges was noted. ciation (1996) code were modiﬁed to take
into account lateral spreading effects.
Seismic isolation studies intensiﬁed.

1999 Chi-Chi Many bridges collapsed because they were Considerations for the design of important
earthquake (Taiwan) located close to or just above the faults. structures in the vicinity of plate boundaries
and faults were implemented in practice.

1999 Koceli Damage to Bolu Tunnel as a result of fault Cyclic failure of intermediate soils (non-
earthquake (Turkey) movement and collapse of buildings due standard soils, e.g. sandy silt) intensiﬁed.
to liquefaction. Buildings conforming to
design codes performed well. The perfor-
mance of intermediate soils, e.g. clayey
sand, sandy silt or silty sand, was noted.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Table 1.2 (Continued)

Earthquake Remarks Post-earthquake developments

2001 Bhuj Large-scale destruction of buildings and The National Program for Earthquake
earthquake (India) infrastructure occurred. Liquefaction of Engineering Education (NPEEE) was formed
earthen dams were observed. This is one of in India, creating widespread awareness
the major intraplate earthquakes that have of earthquake engineering, affecting the
occurred after 1811. 1.25 billion people in the country.

2004 Sumatra Destruction of the built environment due New research focused on tsunami warning
earthquake and to earthquake and giant tsunami waves. systems.
tsunami (Indonesia)

2009 L’Aquila Seven public ofﬁcials of the national Com- Importance of using appropriate words to
earthquake (Italy) mission for the Forecast and Prevention of communicate seismic risk to the public.
Major Risks were charged and found guilty Probabilistic wording (similar to weather
of manslaughter for failing to adequately forecasts) is more suitable than deterministic
warn local residents and giving ‘inaccurate, phrasing.
incomplete and contradictory’ statements
prior to the Mw 6.3 mainshock. The defen-
dants were sentenced by a regional Italian
court to 6 years’ imprisonment and ordered
to pay €7.8 million in damages and costs.
In November 2014, the appeals court over-
turned the manslaughter convictions for six
ofﬁcials but endorsed a 2-year sentence on
one of the defendants for ‘negligence and
imprudence’ in making a series of reassuring
comments to a television journalist ahead of
the experts’ meeting.

2011 Tōhoku Cascade failure observed: earthquakes can Multi-hazard studies initiated globally.
earthquake (Japan) cause intense shaking, trigger tsunamis Studies to evaluate the resilience of critical
and cause liquefaction. The combined infrastructure against multi-hazards initiated.
effects can be very damaging even if a
structure is safe for each of these individual The International Nuclear Energy Academy
hazards. One form of disaster may lead to (INEA) proposed the reassessment of the
the next one. seismic safety of nuclear power plants
globally against cascading natural hazards.
Collapse of the Fukushima Daiichi Nuclear Construction of nuclear power plants in
Power Plant. Strong shaking due to the seismic zones is considered a risky enterprise.
earthquake destroyed the grid and put the A shift began towards offshore wind power
primary cooling system of the power plant and decommissioning of nuclear power
out of action. This led to operation of the plants. A ﬂoating offshore wind farm was
secondary cooling system (diesel generator developed offshore from Fukushima.
set). However, this and the battery power

Table 1.2 (Continued)

Earthquake Remarks Post-earthquake developments

supply were destroyed by 14.5 m tsunami Ageing effects of soil (i.e. the time from
waves, causing the triple melt down of the reclamation) on liquefaction susceptibility
Fukushima Power Plant. initiated.

2011 Christchurch Severe and extensive liquefaction was Liquefaction susceptibility maps proposed.
earthquake observed. Liquefaction can be inﬂuenced Liquefaction mitigation measures suggested
(New Zealand) by previous land use and the presence of to be incorporated in standards.
rivers and streams.

2015 Gorkha Ground motion amplification occurred at Disaster preparedness in urban and rural
earthquake (Nepal) soft soil sites in Kathmandu Valley. Severe parts of Nepal suggested. Retrofitted
liquefaction footprints were observed. Heri- non-engineered masonry structures per-
tage buildings and non- engineered masonry formed well without significant damage.
structures experienced severe damage. Further details can be found in Macabuag
et al. (2012) and Heydariha et al. (2018)

2016 Kumamoto Migrating seismicity (progressive earth- Research to consider fault rupture
earthquake (Japan) quake surface rupture) with foreshock– mechanisms and travelling seismicity
mainshock–aftershock sequence intensiﬁed.
observed. A total of 72 seismic events were
recorded between the foreshock and the
mainshock. Similarly, 248 events after the
mainshock took place within a span of
6 days (between 13 and 18 April 2016).

2016 Amatrice Topographical effects were observed. Earthquake effects being studied to improve
earthquake (Japan) Many retroﬁtted buildings collapsed. the Italian seismic codes.

1.6.1 Case studies

Some of the significant findings from recent earthquakes are briefly described here.

1.6.1.1 1964 Niigata earthquake, Japan

The Niigata earthquake occurred on the 14 June 1964 and registered a moment magnitude (Mw) of 7.6.
Located some 55 km from the epicentre, crossing the Shinano River, Showa Bridge was one of the bridges
that collapsed as a result of the earthquake, primarily due to liquefaction-induced dynamic changes in the
bridge structure. Construction of the bridge had been completed only 15 days before the earthquake struck.
The collapsed view of the bridge is shown in Figure 1.5. Similarly, the Kawagishi-cho apartment complex
(Figure 1.6) tilted due to foundation failure caused by the liquefaction of the underlying soil deposit during
the earthquake. These two instances of damage have been iconic case studies for liquefaction studies globally.
Earthquake geotechnical engineers/scientists started working on liquefaction modelling after this event (Seed
and Lee, 1966), and it took almost seven more years to establish a method to calculate the depth of liqueﬁable
layers (Seed and Idriss, 1971). Further details on liquefaction are provided in Chapters 6 and 7.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 1.5 Aerial photograph of Showa Bridge after the 1964 Niigata earthquake, Japan (photo courtesy of JGS)

Figure 1.6 View of the Kawagishi-cho apartment complex after the 1964 Niigata earthquake, Japan (photo
courtesy of JGS)

1.6.1.2 1985 Mexico earthquake

During the 19 September 1985 Mexico earthquake, engineers witnessed a remarkable feature – ground
motion amplification due to resonance between the natural frequency of the soft soil deposit and the
frequency of the ground motion (Figure 1.7). This was an unexplored subject in earthquake geotechnical
engineering, and initiated investigation of the seismic ground response. Flores et al. (1987) were the first
to publish an article describing ground motion amplification and the factors contributing to it. An
amplification factor of almost 3 was observed near a building in the soft mud, confirming reasons for the

Figure 1.7 Mexico City map showing 1985 earthquake-damaged regions and recording stations (Anderson et al., 1986)

Lake zone
e
zon o
on Rio
mp
n siti Co
a

Tra nsu
Melchor Oc

A lad
o
Hills zone

Eje Central

ec Airport
ultep
Chap
Chabacano
rnia
Baja Califo
o

sco
Patriotism

Xola
s

Rio Churubu
gente

SCT
(98,168,38)
Insur
Revolicion

Tlalpan

CDA
ad
sid

(81,95,27)
r

Calz. de
ive

A
Un

VIV
(44,42,18)

UNAM
(28,34,21)

Accelerograph 0 1000 m
Severely damaged building
Collapsed building
Zone with many collapsed
one- and two-storey houses
(brick and adobe)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

failure. This laid the foundation for ground response studies worldwide, and the ﬁrst computer pro-
gramme to estimate the seismic ground response was released in 1991 (SHAKE 1991). This aspect is
covered in Chapters 3 and 4.

1.6.1.3 1995 Kobe earthquake, Japan

As many as 60% of the bridges (Figure 1.8) in the Kobe area were damaged during the Kobe 1995 earth-
quake. Foundation failure was identified as the most significant contributor to the damage. A new failure
mechanism for foundations known as lateral spreading was observed. Essentially, lateral spreading is the
downward movement of the ground when soil loses its strength and exerts a large lateral earth pressure on
structures opposing their flow. This feature is mostly seen near retaining walls, abutments of bridges and so
on. The Japanese code for the earthquake-resistant design of structures (Japanese Road Association, 1996)
incorporated calculation methods to include lateral spreading effects. Many lateral-spreading-induced
infrastructural failures were reported worldwide after this event.

1.6.1.4 2001 Bhuj earthquake, India

Occurring on 26 January 2001 with an Mw of 7.7 and the epicentre located nearly 50 km to the northeast
of Bhuj, Gujarat (India), this event was the largest intraplate earthquake in Indian history, resulting in a
35 000 death toll and 100 000–150 000 casualties. The epicentral region was classiﬁed as a stable conti-
nental region (SCR) according to seismologists and was not expected to witness such an intense event.
However, the proximity of the Himalayan front and other active geological structures suggests that this
Figure 1.8 Collapsed Hanshin Expressway Bridge in Japan after the 1995 Kobe earthquake (image courtesy of
Paciﬁc Press Service/Alamy Stock Photo)

area may be transitional between an SCR and a plate boundary. Widespread liquefaction was observed in
many earthen dams such as Chang, Fatehgadh (Figure 1.9), Kaswati, Rudramata, Shivlakha, Suvi and
Tapar. Also, large-scale infrastructure damage near the region occurred due to liquefaction-induced lateral
spreading. For example, the Kandla port tower, supported by a mat–pile foundation (Figure 1.10), tilted.
This earthquake helped to accelerate geotechnical earthquake engineering studies in India.

1.6.1.5 2011 Tōhoku earthquake, Japan

The Tōhoku earthquake, occurring on 11 March 2011, is the fourth largest earthquake ever recorded in human
history, with an Mw of 9.0. Geologists believe that the massive earthquake caused Japan’s main island of
Honshu to move 2.4 m eastward. The unexpected tsunami with a peak wave height of about 14 m due to the
earthquake caused most damage to the nation and led to one of the worst ever nuclear disasters, from the triple
meltdown of the Fukushima Daiichi Nuclear Power Plant. Figure 1.11 shows the rescue operation in the
aftermath of the event. One of the new observations in this event was the complete uprooting of buildings
supported on deep foundations: see Figures 1.2(b), 1.3 and 1.12. Figure 1.13 shows damage to a town that
was built in an artiﬁcial valley whose ﬁll material consisted of weakly consolidated sandy soil, originated
from the weathering of tuffaceous sandstone. The landslide induced the tilting of several low-rise houses, and
severe damage to the road. The house depicted in Figure 1.13 experienced vertical settlement of the order of
20–30 cm, exposing its foundation and leaving it balancing precariously.

1.6.1.5.1 Aside: Fukushima Daiichi Nuclear Power Plant. In the 2011 Tōhoku earthquake, a com-
bination of ground shaking due to the Mw 9.0 earthquake, subsurface liquefaction and the tsunami caused a
triple meltdown of the Fukushima Diichi Nuclear Power Plant. Earthquake shaking and liquefaction damaged
the national power grid but the nuclear power plant was within safety limits. However, the tsunami, which
arrived around 50 min after the mainshock, was about 14 m high, which overwhelmed the 10 m-high sea
walls and resulted in ﬂooding of the emergency generator rooms, causing power failure of the reactor cooling

Figure 1.9 Large open ﬁssures caused by localised liquefaction in San Carlo municipality after the 2012 Northern
Italy earthquake sequence (Italy)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 1.10 Location of the epicentre of the Bhuj earthquake and the tilted Kandla port tower (Dash et al., 2009)

PAKISTAN INDIA

Mw 7.7 Earthquake Rajasthan

(26th January 2001)
★
Kutch Kandla Port
Gandhi Nagar
Ahmadabad
Gujarat

Jamnagar Rajkot

N Madhya
Porbandar Pradesh

Surat
Diu
ARABIAN SEA
Daman
Maharastra

Quay ARABIAN SEA

wall Port and
customs tower
Passenger
jetty

Port and
customs
Berthing jetty

Figure 1.11 The damaged Fukushima Daiichi Nuclear Power Plant in Japan – a result of the giant tsunami caused by
the 2011 Tōhoku earthquake (image courtesy of Photo 12/Alamy Stock Photo)

Figure 1.12 Uprooted pile-supported building near Onagawa port in Japan due to the 2011 Tōhoku earthquake
and the subsequent tsunami

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 1.13 Landslide-induced damage in Taiyo New Town (Japan) following the 2011 Tōhoku earthquake

systems. The loss of the cooling systems led to the reactor heating up and subsequent meltdown; conse-
quently, harmful radioactive materials were released to the environment. The power failure also meant that
many of the safety control systems were not operational. The release of the radioactive materials led to the
large-scale evacuation of over 300 000 people near the plant, and the clean-up costs are estimated to be of the
order of hundreds of billions of dollars.

Had there been emergency cooling power, the accident may have been averted: during the earthquake the
wind turbines automatically shut down (like all escalators and lifts). Following an inspection after the
earthquake, they were restarted. Figure 1.14 shows the location of the earthquake and the operating wind
farms. Figure 1.15 shows photographs of the wind farm at Kamisu (Hasaki) after the earthquake.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Figure 1.14 Details of the 2011 Tōhoku earthquake and locations of the wind farms. PGA, peak ground acceleration; Rrup, distance from the fault rupture;
VS30, average seismic shear-wave velocity from the surface to a depth of 30 m

139°E 140°E 141°E 142°E 143°E 144°E 145°E Hiyama wind farm
Wind farm 0 50 100 150 200 FKSH19: Rrup = 43.1 km and VS30 = 338 m/s
Typical range
41°N Strong motion km
for wind turbines
station

NS component
40°N PGA = 569 cm/s2 1000
500 cm/s2

39°N EW component
100 s PGA = 824 cm/s2

38°N 100
Kamisu/Hasaki wind farm
IBRH20: Rrup = 38.3 km and VS30 = 244 m/s
37°N FKSH19 – NS
Spectral acceleration: cm/s2

FKSH19 – EW
NS component
IBRH20 – NS
36°N PGA = 215 cm/s2 10
500 cm/s2 IBRH20 – EW
Epicentre 0.01 0.1 1 10
35°N EW component
Fault plane model

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
100 s PGA = 186 cm/s2 Natural vibration period: s
Introduction to earthquake geotechnical engineering in relation to foundation design

25
Seismic Design of Foundations

Figure 1.15 The Kamisu (Hasaki) wind farm following the 2011 Tōhoku earthquake

But why did the wind farm remain standing? The recorded ground acceleration time-series data are presented
in Figure 1.16 for the Kamisu and Hiyama wind farms (FKSH19 and IBRH20) in the frequency domain for
two directions (north–south (NS) and east–west (EW)). The dominant-period ranges of the recorded ground
motion at the wind farm sites were around 0.07–1 s, whereas the period for offshore wind turbine systems is in
the range of 3 s. Due to non-overlapping, these structures will not become tuned, and as a result they are
relatively insensitive to earthquake shaking. However, earthquake-induced effects such as liquefaction could
still cause damage.

It could be argued that had there been a few offshore wind turbines operating, the disaster may have been averted
or at least the scale of damage reduced. The wind turbines could run the emergency cooling system and prevent the
reactor meltdown. In this context, it is interesting to note that there are plans to replace the Fukushima Daiichi
Nuclear Power Plant by a floating wind farm. The project is in advanced stage, and a 2 MW semi-submersible
floating turbine will be operational in a few years. An innovative 7 MW oil-pressure-drive-type wind turbine on a
three-column semi-submersible floater has recently been tested.

1.6.1.6 2015 Gorkha earthquake, Nepal

The Nepal earthquake on 25 April 2015 (also called the Gorkha earthquake) with an Mw of 7.8 affected
central Nepal and neighbouring areas. The Kathmandu Valley witnessed severe structural damage,
including heritage structures, such as the complete collapse of Dharahara Tower (Figure 1.17). Severe
liquefaction footprints and basin effects were identiﬁed in the region (Gautam, 2017). A unique feature of
this earthquake was the signiﬁcant amplitude of vertical components in the valley, which is considered a
rare seismic feature.

Figure 1.16 Ground related failures from 2016 AMATRICE Earthquake (Italy)

(a)

(b)

(c)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 1.17 Dharahara Tower in Kathmandu before and after the 2015 Nepal earthquake

1.6.1.7 2016 Kumamoto earthquake, Japan

The recent 2016 Kumamoto earthquake in Japan (16 April 2016) is an example of a cascading event
(i.e. travelling seismicity/migrating seismicity). Travelling seismicity is a very unusual/rare seismic hazard
where one earthquake triggers the next. This essentially mean the progressive increase of seismicity with
fault rupture. Figure 1.18 shows the number of events before and after the mainshock, typically called
foreshocks and aftershocks, respectively. As seen in Figure 1.18, a foreshock with a Japan Meteorological
Agency magnitude (Mj) of 6.5 was followed by a mainshock with an Mj of 7.3. Goda et al. (2016) reported
that a total of 72 foreshocks and 248 aftershocks with Mj > 3.5 took place during the sequence. After the
Kumamoto earthquake, seismologists suggested the inclusion of fault rupture mechanisms in the code
provisions, especially for highly active seismic regions. Figure 1.19 shows the subsidence of a road and
reconstruction following the earthquake.

1.7. Complexity of earthquake geotechnical engineering

The subject of earthquake geotechnical engineering is relatively new. It would not be an overstatement to say
that this subject started following the 1964 Niigata and Alaska earthquakes where widespread liquefaction
was observed. Since then, it has been rapidly developing. This section discusses some of the essential
complexities in the subject.

Earthquakes are, by nature, random and unpredictable. The time of an earthquake at any place can be pre-
dicted only in the order of geological years, but areas prone to earthquakes can be mapped based on fault
geometry, and this is known as micro-zonation where very localised seismic hazards can be scrutinised. Each
earthquake is also unique in its magnitude, duration and frequency. Figure 1.20 shows the recorded rock
outcrop motion of the 2011 Sikkim earthquake (India) in three orthogonal directions (north–south, east–west
and up–down vertically) at a location in Gangtok (Sikkim) in the frequency and time domains. The earth-
quake had a duration of 152 s, and the peak ground acceleration (PGA) was measured as 0.161g for the east–
west components, which reduced to 0.152g for the north–south components. From the frequency domain
plots, it is clear that the motion had multiple frequencies of varying magnitudes. It has been widely accepted
that two consecutive earthquakes at the same location may be very dissimilar. Therefore, what should an

Figure 1.18 Foreshock, mainshock and aftershocks for the 2016 Kumamoto earthquake, Japan (Goda et al., 2016)

100
MJ 7.3 1 April 2016 to 31 May 2016
90 mainshock 33.0°N

80
MJ 6.5 Mainshock
70 32.9°N
foreshock
Number of events

50 32.8°N

40
32.7°N Foreshock
30
MJ < 5
20
5 ≤ MJ < 6
32.6°N
10 MJ ≥ 6

0
13/04 14/04 15/04 16/04 17/04 18/04 130.5°E 130.6°E 130.7°E 130.8°E 130.9°E 131.0°E 131.1°E
Date

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 1.19 A road (note the subsidence) (a) before the 2016 Kumamoto earthquake and (b) after reconstruction

(a)

(b)

Figure 1.20 Recorded ground motion of the 2011 Sikkim earthquake at Gangtok and the corresponding frequency
distribution (source: PESMOS (Earthquake Engineering Department, IIT Roorkee, 2018))

50 100 150 0.1 1 10

0.2 0.100
PGA = 0.135g Gangtok station − 27.325° N88.627°E Vertical component 0.083 @ 15.06 Hz

0.1 Vertical component 0.075

0.0 0.050

–0.1 0.025

–0.2 0.00
North – South component 0.132 @ 5.55 Hz
North – South component
0.12

Fourier amplitude
0.1
Acceleration: g

0.09
0.0
0.06

–0.1
0.03
PGA = 0.161g
–0.2 0.00
East – West component 0.134 @ 7.142 Hz
East – West component
0.1
0.10

0.0

0.05
–0.1
PGA = 0.152g

–0.2 0.00
50 100 150 0.1 1 10
Time: s Frequency: Hz

engineer choose while designing structures and foundations? In practice, the design parameters for ground
motion at a particular site (i.e. frequency, magnitude and duration) are obtained from statistical and seis-
mological analysis of past recorded data.

The design life of most structures varies between 30 years (oil platforms, offshore wind turbines, liquid gas
tanks and pipelines) and few thousand years (heritage structures and important infrastructure projects such as
dams and nuclear power plants (e.g. safety structures for the latter are typically designed for a design life of
10 000 years)). The probability of occurrence of earthquakes at a particular site can be less than a few per cent
over 30 years. Implementing best seismic practice in the design of short-life structures will increase the cost
substantially and is often ignored in the design of certain unmanned structures. But if the economic risk for a
structure is high, engineers prefer not to ignore seismic considerations, and apply some judgement. These
engineering judgements are often considered based on probabilistic methods.

Soil by its nature is variable and heterogeneous. In general, there are two types of ground response that are
damaging to structures. In one, the soil fails typically by liquefaction, such as in the 1995 Kobe earthquake. In
the other, the soil ampliﬁes the ground motion (as in, for example, the 1989 Loma Prieta earthquake in
California). Often, laboratory and model tests for earthquake geotechnical engineering problems are carried
out on pure sands or clays or other well-deﬁned materials. This is necessary to understand the fundamental

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

mechanics or processes. It is very well known from laboratory tests that loose to medium-dense sands liquefy
under moderate shaking, but how do we assess the liquefaction potential for mixed soils such as clayey sand
or sandy clay? We rely on ﬁeld case records and observations. Often, soils at sites are not uniform, which adds
further complexity, and results obtained in laboratory environments are often not applied due to lack of
conﬁdence. Therefore, design procedures in earthquake geotechnical engineering are often based on
empirical or semi-empirical approaches: for example, liquefaction assessment and evaluation, and the pre-
diction of permanent ground displacement.

Based on the above discussion, it can be inferred that designing foundations in seismic areas is a multi-
disciplinary process and involves many professionals such as statisticians, seismologists, geotechnical
engineers and structural engineers.

1.8. Summary
The ground can be a signiﬁcant contributor to the damage of infrastructure during earthquakes. Various
ground and structural failure mechanisms have been observed, and this is how the profession and subject area
of earthquake geotechnical engineering developed. Noteworthy ground failures include liquefaction, ground
motion ampliﬁcation in soft soil deposits and lateral spreading.

This chapter introduced geotechnical damage observed during recent seismic events with a focus on
summarising the geotechnical earthquake engineering aspects for practising civil and structural engineers.

REFERENCES
Anderson JG, Bodin P, Brune JN et al. (1986) Strong ground motion from the Michoacan, Mexico, earthquake.
Science 233(4768): 1043–1049.
Dash SR, Govindaraju L and Bhattacharya S (2009) A case study of damages of the Kandla Port and Customs
Office tower supported on a mat–pile foundation in liquefied soils under the 2001 Bhuj earthquake. Soil
Dynamics and Earthquake Engineering 29(2): 333–346.
Earthquake Engineering Department, IIT Roorkee (2018) PESMOS. http://pesmos.in (accessed 24/08/2018).
EERI (Earthquake Engineering Research Institute) (2001) Bhuj, India Republic Day January 26, 2001. Earthquake
Reconnaissance Report. EERI, Oakland, CA, USA.
Flores J, Novaro O and Seligman TH (1987) Possible resonance effect in the distribution of earthquake damage in
Mexico City. Nature 326(6115): 783.
Gautam D (2017) Unearthed lessons of 25 April 2015 Gorkha earthquake (MW 7.8): geotechnical earthquake
engineering perspectives. Geomatics, Natural Hazards and Risk 8(2): 1358–1382.
Goda K and Tesfamariam S (2015) Seismic risk management of existing reinforced concrete buildings in the
Cascadia subduction zone. Natural Hazards Review 18(1), 10.1061/(ASCE)NH.1527-6996.0000206.
Goda K, Campbell G, Hulme L, Ismael B et al. (2016) The 2016 Kumamoto earthquakes: cascading geological
hazards and compounding risks. Frontiers in Built Environment 2(19): 1–23.
Heydariha JZ, Hossein G, Nayak S et al. (2018) Experimental and field performance of PP Band - retrofitted
masonry: evaluation of seismic behavior. Journal of Performance of Constructed Facilities 33(1): 04018086.
Japanese Road Association (1996) Specification for Highway Bridges, part V. Seismic Design. Japanese Road
Association, Tokyo, Japan.
Macabuag J, Guragain R, Bhattacharya S (2012) Seismic retrofitting of non-engineered masonry in rural Nepal.
Proceedings of the Institution of Civil Engineers - Structures and Buildings 165(6): 273–286.
Seed HB and Idriss IM (1971) Simplified procedure for evaluating soil liquefaction potential. Journal of the
Geotechnical Engineering Division, ASCE 97(9): 1249–1273.
Seed HB and Lee KL (1966) Liquefaction of saturated sands during cyclic loading. Journal of the Soil Mechanics
and Foundation Division, ASCE 92(6): 105–134.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

ISBN 978-0-7277-6166-8
https://doi.org/10.1680/sdof.61668.033
ICE Publishing: All rights reserved

Chapter 2
Basic concepts of engineering seismology
and seismic hazard analysis
2.1. Introduction
The goal of earthquake-resistant design is to minimise seismic risk, deﬁned as the possibility of casualties
and economic and cultural loss due to seismic events. This is achieved by ensuring that a structure can
withstand the expected level of ground shaking while maintaining a desired level of performance, or, in
other words, that the seismic capacity of the structure is equal to or larger than the expected seismic
demand. While the seismic capacity depends on the level of exposure and vulnerability of the structure,
which can be controlled and predicted by designers and planners, the seismic demand, or prediction of the
expected level of ground shaking, requires characterisation of the ground shaking hazard due to future
earthquakes for the site of interest. This chapter focuses on the deﬁnition of the expected level of the
ground at the site of interest.

2.2. Theories of plate tectonics and elastic rebound

According to the theory of plate tectonics (Wegener, 1915), the outermost layer of the Earth (i.e. the crust)
consists of number of plates that move relative to each other. The greatest level of seismic activity occurs at
the boundary between two or more plates, where these converge, diverge or move laterally past one another.
Depending on the type of movement, plate boundaries are classiﬁed into three types

■ diverging plate boundaries, also known as spreading ridge boundaries, where plates move away from
each other due to rise of magma
■ convergent plate boundaries, also referred to as subduction zone boundaries, where two plates
converge with one another, such as at the edges of continents, wherein the denser ocean plate
subducts beneath the lighter crustal plate
■ transform fault boundaries, which are found when plates move past each other, such as in the San
Andreas fault in California.

Intraplate regions are characterised by lower seismic activity.

Earthquakes originate by an abrupt release of stored elastic strain energy in the form of seismic waves that
propagate through the Earth’s crust. The size of the earthquake is proportional to the amount of stored energy
released by the fault rupture, which depends on the length of the rupture and slip on the fault. The fault rupture
may extend for several kilometres along the fault plane. The origin of the rupture is often idealised by a single
point, referred to as the hypocentre or focus. The depth of the hypocentre is known as the focal depth; the
vertical projection of the hypocentre on the ground surface is known as the epicentre.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 2.1 Geometric deﬁnition of strike and dip angle

Fault plane

Horizontal plane

e
rik
St
Dip angle

2.3. Types of faults

The geometry of the fault plane is described in terms of its orientation with respect to a horizontal plane
(Figure 2.1). The strike denotes the horizontal line formed by the intersection of the fault plane and the
horizontal plane. Its orientation is defined by the azimuth angle with respect to due north. The angle between
the fault plane and the horizontal plane, measured perpendicular to the strike, is defined as the dip angle of the
fault. Depending on their movement, faults are classified into two categories: dip-slip faults and strike-slip
faults. A simple classification system to categorise the style of faulting consists of calculating the ratio of the
horizontal component to the vertical component of the slip, known as the horizontal-to-vertical slip ratio.
Earthquakes with slip ratios greater than unity are associated with strike-slip faulting, while those with slip
ratios less than or equal to one are associated with normal or reverse faulting, depending on the direction of
the vertical displacement.

2.3.1 Dip-slip faults

In dip-slip faults the movement occurs perpendicular to the strike along the dip. Dip-slip faults are further
classiﬁed into normal faults and reverse or thrust faults. In normal faults the horizontal component of the
movement is extensional, and the block above the fault plane – referred to as the hanging wall, moves
downwards. Earthquakes associated with normal faulting include the 1980 Irpinia earthquake (Italy), the
1995 Kozani earthquake (Greece), the 1995 Dinar earthquake (Turkey), the 2009 L’Aquila earthquake (Italy)
and the 2016 central Italy Earthquake. In reverse faults the horizontal component of the movement is
compressional, and the block above the fault plane moves upwards. Earthquakes associated with reverse
faulting include the 1971 San Fernando earthquake (USA), the 1976 Friuli earthquake (Italy) and the 1978
Tabas earthquake (Iran). Thrust faults are special cases of reverse faults characterised by a small dip angle.
Thrust faults have triggered major past earthquakes, including the 2008 Sichuan earthquake (China) and the
2011 Tōhoku earthquake (Japan).

2.3.2 Strike-slip faults

In strike-slip faults the movement is parallel to the strike and nearly vertical. Strike-slip faults can be further
classiﬁed based on their relative movement into right lateral strike-slip faults, when the block opposite an
observer looking across the fault moves to the right, and left lateral strike-slip faults, if the observer see the
block moving to the left (Figure 2.2). The San Andreas fault in California is a typical example of right lateral
strike-slip fault.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

Figure 2.2 Types of faulting: (a) normal fault; (b) reverse fault

Normal fault Reverse fault

(a) (b)

2.4. Evaluation of earthquake size

The oldest measure of earthquake size is earthquake intensity. According to the original definition of
intensity, the severity of an earthquake can be determined based on a qualitative description of the damage to
both the built and natural environment and of human reactions caused by the earthquake at a particular site.
However, the effects of earthquakes on the natural environment are often disregarded for the assessment of
earthquake intensity, arguably due to the lack of a systematic classification and quantification of geological,
hydrological and geomorphic features for different degrees of intensity. Yet, environmental effects are less
sensitive to the quality of the building stock, thus they can provide a complementary measure of the epicentral
intensity, especially when effects on artificial structures are scarce and/or quickly saturate due to poor
structural quality.

2.4.1 Modiﬁed Mercalli intensity (MMI) scale

The most widely used intensity scale is the modiﬁed Mercalli intensity (MMI) scale, initially proposed by
Mercalli and subsequently modiﬁed by Richter in 1958. The MMI scale consists of 12 intensity degrees
denoted by Roman numerals whose qualitative description is given in Table 2.1.

2.4.2 Moment magnitude (Mw) scale

A more objective measure of the size of an earthquake can be obtained by estimating the total amount of
elastic strain energy released by the fault rupture. Different magnitude scales have been proposed depending
on the quantities used for the determination of the energy released by an earthquake. In most historical
earthquake databases, earthquake size is expressed in terms of the Richter local magnitude scale, denoted by
ML, or the surface wave magnitude, denoted by Ms. Although these are the most commonly cited magnitudes
for older earthquakes, because of their deﬁciency to saturate at magnitudes higher than 6, they have been
superseded by the moment magnitude scale, denoted by Mw, which does not suffer from saturation. It is worth
noting that that there are no signiﬁcant differences between the different magnitudes for values ranging from
5.0 to 7.5 (Kanamori, 1983).

The moment magnitude scale is considered a more accurate measure of the size of an earthquake since it is a
directly related to fault rupture parameters through the seismic moment M0 deﬁned by

M0 = μAD (2:1)

where μ denotes the shear modulus of the material along the fault (typically 32 GPa in the crust and 75 GPa in
the mantle), A is the area of the fault that ruptured (given by the product of fault length times fault width) and

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Table 2.1 Qualitative description of intensity degrees in the MMI scale

Intensity Shaking Description/damage

I Not felt Not felt except by a very few people under especially favourable conditions.
II Weak Felt only by a few people at rest, especially on the upper floors of buildings.
III Weak Felt quite noticeably by people indoors, especially on the upper floors of
buildings. Many people do not recognise it as an earthquake. Standing motor
cars may rock slightly. Vibrations similar to the passing of a truck. Duration
estimated.
IV Light During the day, felt indoors by many people, but felt outdoors by few. At night,
some people are awakened. Dishes, windows and doors disturbed; walls make
a cracking sound. The sensation is like a heavy truck striking the building.
Standing motor cars rocked noticeably.
V Moderate Felt by nearly everyone; many awakened. Some dishes and windows broken.
Unstable objects overturned.
VI Strong Felt by all, many frightened. Some heavy furniture moved; a few instances of
fallen plaster. Damage slight.
VII Very strong Damage negligible in buildings of good design and construction; slight to
moderate damage in well-built ordinary structures; considerable damage in
poorly built or badly designed structures. Some chimneys broken. Noticed by
people driving cars.
VIII Severe Damage slight in specially designed structures; considerable damage in ordi-
nary substantial buildings, with partial collapse; damage great in poorly built
structures. Fall of chimneys, factory stacks, columns, monuments and walls.
Heavy furniture overturned. Changes in well waters. People driving cars may
be disturbed.
IX Severe Damage considerable in specially designed structures; well-designed frame
structures thrown out of plumb; damage great in substantial buildings, with
partial collapse. Buildings shifted off foundations. Underground pipes broken.
X Violent Some well-built wooden structures destroyed; most masonry and frame
structures destroyed with their foundations. Rail tracks bent. Considerable
landslides from river banks and steep slopes.
XI Violent Few, if any (masonry), structures remain standing. Bridges destroyed. Broad
fissures in the ground; earth slumps and land slips in soft ground. Underground
pipelines out of service. Rail tracks severely bent.
XII Extreme Damage total. Practically all works of construction are severely damaged or
destroyed. Waves seen on the ground surface. Lines of sight and level are
distorted. Objects thrown into the air.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

D is the average displacement across the fault surface. The seismic moment has units of force times
displacement, hence it is physically the work done by an earthquake, which is a direct measure of the
energy released during the earthquake. Practically, M0 is computed from source spectra of body and
surface waves (Hanks et al., 1975; Kanamori and Anderson, 1975) or derived from a moment tensor
solution (Dziewonski et al., 1981). According to Hanks, Kanamori and their co-workers, the moment
magnitude Mw can be computed by

log M0
Mw = − 10:7 (2:2)
1:5
Empirical relationships between the moment magnitude and rupture parameters can be used to estimate Mw.
Although different authors have proposed different empirical relationships, those proposed by Wells and
Coppersmith (1994), known as the WC94 relations, are normally used in practice for seismic hazard analyses.
WC94 relations are based on a least-square regression model that relates Mw with surface and subsurface
rupture lengths, the rupture area and the rupture surface displacement. From Figure 2.3, it can be seen that
these rupture parameters increase exponentially with increasing Mw.

Figure 2.3 Empirical relationship between Mw, rupture length, rupture area and rupture displacement from Wells and
Coppersmith (1994): (a) Mw versus rupture area; (b) Mw versus rupture length; (c) Mw versus rupture displacement

10000 350 10
Mw = 4.07 + 0.98log(A) Surface Maximum
Mw = 5.08 + 1.16log(SRL) Mw = 6.69 + 0.74log(MD)
9000 9
300 Average
Subsurface Mw = 6.93 + 0.82log(AD)
8000 8
Mw = 4.38 + 1.49log(SsRL)

250 7
7000
Rupture surface displacement: m

6000 6
Rupture length: km
Rupture area: km2

200

5000 5

150
4000 4

3000 3
100

2000 2
50
1000 1

0 0 0
5 6 7 8 5 6 7 8 5 6 7 8
Mw Mw Mw
(a) (b) (c)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

2.5. Hazards for the seismic design of foundations

It is necessary to quantify seismic hazards when designing a structure to withstand earthquakes. In general,
there are two major seismic hazards associated with the design of structures and foundations: (a) ground
shaking; (b) ground deformation. Descriptions of these hazards are given below.

■ Ground shaking. Every earthquake is associated with ground motion, which generally includes body
waves and surface waves. The ground motion is often characterised in terms of the peak ground
acceleration (PGA) in both the horizontal and vertical directions.
■ Ground deformation. Earthquakes often induce permanent ground deformation. This may be due to
faulting, landslide, liquefaction, tectonic uplift and subsidence and so on. If the structural system
encounters a potential ground deformation area, the amount of ground deformation has to be
quantified and should be considered in the design process. Ground deformation in an earthquake may
occur due to
– Faulting. Fault movement causes localised ground displacement. The amount of fault displacement
is characterised by ground displacement along or normal to the direction of the fault line.
– Liquefaction. Liquefaction is a phenomenon restricted to saturated cohesionless soil, which losses
its strength and stiffness during earthquake shaking. This happens as a result of increased pore
water pressure and loss of shearing strength due to seismic vibration. If the site is found to be
susceptible to liquefaction, the design of structures and their foundations should focus on the mode
and magnitude of the ground failure that might occur. The major types of liquefaction-induced
ground failure and their effects on structures and foundations are listed in Table 2.2.
– Landslide. Landslide is the permanent deformation of soil, which can be very damaging to the built
environment over a particular area. The amount of landslide movement is characterised in terms of
the permanent ground displacement.
– Other causes. Other seismic hazards, such as tectonic uplift and tectonic subsidence, may cause
significant amounts of damage to the structural system in certain seismic events, but these
situations are not very common. Specified literature can be referred to for quantification of these
hazards.

Table 2.2 Liquefaction-induced ground failure

Liquefaction-induced Effect on structures and foundations

ground failure

Failure of slopes Sometimes the main cause of the collapse of a structure due to very large downslope
movement of a soil mass
Lateral spreading Occurs in gently sloping ground. Sometimes leads to the failure of foundations by
imposing a large lateral soil pressure
Loss of bearing capacity Predominantly causes foundation failure by sinking
Buoyancy Floating of buried structures such as tanks and rafts may occur
Ground settlement Settlement of the ground

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

Different seismic hazards are quantiﬁed differently. A designer will want to know the following

■ the PGA (i.e. the amplitude of vibration)

■ the earthquake-induced ground displacement
■ the duration of the earthquake (this is important from the point of view of how long a soil will remain
liqueﬁed)
■ the frequency content of the ground motion (the predominant frequency content of most earthquakes
lies in the range 0.5–10 Hz)
■ the fault rupture length or fault movement (for the design of long linear structures such as tunnels or
pipelines).

Ground shaking hazards are deﬁned through hazard maps, which show the geographical distribution of
earthquake shaking levels for a certain probability of exceedance in a given period of time. For ordinary
buildings, 10% probability of exceedance in 50 years –corresponding to a return period of 475 – is normally
considered for design.

2.6. Seismic hazard analysis

The quantitative assessment of the ground shaking hazard due to future earthquakes at a particular site
requires the development of a seismicity model that specifies the spatial distribution of all seismic sources, by
means of lines or zones at the locations of active geological faults, and their potential size. Evidently, there are
uncertainties related to the size of an earthquake and temporal and spatial occurrence of future earthquakes.
Depending on how these uncertainties are accounted for, seismic hazard analyses can be classified into two
main categories: deterministic seismic hazard analysis (DSHA) and probabilistic seismic hazard analysis
(PSHA). It is worth noting that rather than viewing these as two different types of analysis, DSHA is in fact a
special case of PSHA where a single earthquake scenario, normally the worst-case scenario, is considered for
the definition of the seismic hazard at a site.

2.6.1 Deterministic seismic hazard analysis (DSHA)

A typical DSHA can be conducted in four steps (Figure 2.4). Deterministic seismic hazard analysis is one of
the simple straightforward approaches for assessing the seismic hazard by using discrete, single-valued
models. This method is based on the cause of the earthquake rather than its effect. For example, the magnitude
of the design earthquake can be predicted to some approximation from the length of faults and its proximity
from the site. It is of course assumed that a certain portion of the fault is going to rupture and is carried out
using some established empirical relationships which are based on data obtained from previous earthquakes.
The basic steps involved in the deterministic seismic hazard assessment are

Step 1. The faults in the vicinity of the site that can cause earthquakes are identiﬁed.
Step 2. The magnitude of earthquakes from each of these sources are predicted. The corresponding
distances from the sources are computed. Attenuation relationships are used to predict the
resulting ground motion at the site.
Step 3. The PGA is plotted against the distance from the earthquake source.
Step 4. The most severe case is considered as the design earthquake.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 2.4 Steps involved in DSHA

Fault Mw = ?
(line source) Mw = ? PGA = ?
Fault PGA = ?
(line source)

Area Mw = ?
source
Mw = ? PGA = ?
Fault
(line source) PGA = ?

Step 1. Finding the nearest fault Step 2. Calculate the maximum earthquake
and area source of earthquakes that could trigger from each source

Mw
PGA

PGA = ? or PGV = ?

Distance
Step 3. Select the controlling earthquake for the Step 4. Specify the PGA, peak ground velocity or
site, i.e. which earthquake has the most severe other measure that describes the earthquake
effect (generally peak ground motion) at the site effect for the site under consideration

An attenuation relationship is the empirical equation used to estimate ground motion at any site for a given
distance from an earthquake of speciﬁc magnitude. Such equations are developed by correlating the past
seismic data for a particular region. A basic form of the attenuation model generally used is (Joyner and
Boore, 1981)

logðPGAÞ = A + BðMw Þ + Cðr Þ + D logðrÞ + E ðPÞ (2:3)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

where PGA is in g; r = √(d2 + 7.32), where d is the shortest distance from the site to the surface projection of
the fault rupture in kilometres; and A, B, C, D and E are constants that depend upon the seismic source
characterisation.

The reader should look at the latest literature on these attenuations relations.

The drawback with DSHA is that the level of conservatism in the design earthquake is unknown. Also, this
method does not incorporate the effect of the return period of an earthquake. Moreover, some faults may yet to
be discovered – often termed ‘hidden faults’. However, the advantages of this method are that (a) it is rel-
atively straightforward and (b) it gives a conservative answer.

DSHA has been extensively used for the seismic hazard analysis of sites for critical infrastructure (e.g.
nuclear power plants) where the worst-case earthquake scenario governs the overall design due to the
potential catastrophic consequences of failure. However, the method presents a number of limitations due to
its inability to account for uncertainties in the size, time of occurrence and location of fault ruptures.

2.6.2 Probabilistic seismic hazard analysis (PSHA)

PSHA determines the seismic hazard at a site in the form of hazard curves expressing the probability of
exceedance of a ground motion parameter within a speciﬁed period of time. Probabilistic seismic hazard
analysis allows for the quantiﬁcation of uncertainties in size, and spatial and temporal distributions of future
earthquakes, and makes it possible to quantify these uncertainties in a rational manner by means of proba-
bility distributions. A typical PSHA can be conducted in four steps (Figure 2.5).

Step 1. Similarly to DSHA, the ﬁrst step of PSHA consists of the deﬁnition of the seismicity model of the
region of interest by means of source lines or zones. However, as the fault rupture may occur at any point
within the seismic zone, the spatial uncertainty in the earthquake location is taken into account through
probability density functions of distance of the site from the point of rupture. Although different probability
functions can be used, it is customary to adopt a uniform probability distribution in which all points within
the seismic source have same probability to rupture. This is in contrast to DSHA, where it is implicitly
assumed that fault rupture will occur on the location of the source closest to the site, i.e. one with the
shortest source-to-site distance.

Step 2. The seismic hazard analyst now needs to quantify and take into account the uncertainty in the size of
earthquakes – normally expressed in terms of Mw – that each seismic source can produce. It is customary to
express the size of future earthquakes through recurrence laws that provide the rate of occurrence of earth-
quakes with a magnitude greater than or equal to a certain threshold value. Although several recurrence laws
are available, the Gutenberg–Richter model (Gutenberg and Richter, 1944) is normally used to quantify the
mean annual rate of exceedance lm of an earthquake of magnitude M through the power law given by
log lm = a − bM (2:4)

or, alternatively,

lm = 10a − bM (2:5)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 2.5 Steps involved in PSHA

Area source

Annual number of earthquakes

Line source

Site

Area source

Line source

Earthquake magnitude, M

Step 1. Identify the source of potential Step 2. Evaluate the seismicity of earthquake
earthquakes and represent graphically sources from the recurrence relationship

M-1
M-2 Increasing Probability of exceedance
M-3 magnitude, M
Ground motion

Distance, R Ground motion

Step 3. Develop the ground motion Step 4. By integration of models, produce

attenuation models the seismic hazard curves that express the
probability of exceeding various levels of
ground motion within some time period

where 10a is the mean annual number of earthquakes with zero magnitude, or the number of earthquakes with
the lowest magnitude Mmin that the causative fault is expected to produce, and b is the relative proportion of
strong-to-moderate seismic events. The original Gutenberg–Richter law considered a magnitude range from
–∞ to +∞; however, it is of practical interest to consider earthquakes in the magnitude range that can be

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

realistically produced by the causative fault. lm of earthquakes in the magnitude range Mmin to Mmax is given
by (McGuire and Arabasz, 1990)

exp[− b(M − Mmin )] − exp(Mmax − Mmin )

log lm = v (2:6)
1 − exp (Mmax − Mmin )

where v = exp(a − bMmin), and a and b are parameters used in the Gutenberg–Richter law, given by
a = 2.303a and b = 2.303b.

Knowing the rate of occurrence of all possible earthquakes produced by each seismic source, the computed
seismic scenarios can ﬁnally be expressed in terms of magnitude–distance pairs M-R.

Step 3. The relevant ground motion parameters are then computed using attenuation relationships consid-
ering all possible earthquake scenarios – all pairs of M-R.

Step 4. The ﬁnal step consists of developing the seismic hazard curves that express the mean annual prob-
ability of exceedance at a particular site, taking into account the aggregate risk from different earthquake
scenarios. The annual probability of exceedance for a given value, y*, of the ground motion parameter Y for
all possible pairs M-R is given by
ðð
ly* = P[Y > y*jM, R] fM (M)dM fR (R)dR (2:7)

where fM(M) and fR(R) are the probability density functions of the moment and source-to-site distance,
respectively. For the number of earthquake sources Ns, the probability of exceedance can be computed as

Ns ðð
X
ly* = vi P[Y > y*jM, R] fMi (M)dM fRi (R)dR (2:8)
i=1

The numerical solution of the integrals in Equation 2.8 is not trivial. However, a closed-form solution can be
found by discretising the problem into NM magnitude intervals DM, and NR source-to-site distances intervals
DR, as

Ns X
X NM X
NR
ly* = vi P[Y > y*jMj , Rk ] fMi (Mj )DM fRi (Rk )DR (2:9)
i=1 j=1 k=1

where

(j + 0:5)ðMmax − Mmin Þ
Mj = Mmin + (2:10)
NM

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

(k + 0:5)ðRmax − Rmin Þ
Rk = Rmin + (2:11)
NR

Mmax − Mmin
DM = (2:12)
NM

Rmax − Rmin
DR = (2:13)
NR

In some applications, however, it is required to consider the most likely earthquake magnitude and/or distance
from the source to the site. This can be achieved through the so-called process of deaggregation, whereby the
mean annual probability of exceedance is expressed as a function of the magnitude ly*(M) or the source-to-site
distance ly*(R).

2.7. Case study: damage to the outlet tunnel of the second Kakkonda
hydropower station (This case study is based on the seminal work
of Prof Konagai, see for further details Konagai (2005), Johansson
and Konagai (2007))
The Iwate prefecture Nairiku Hokubu (Mid North Iwate) earthquake of 3 September 1998 shook the
Shizukuishi area at 16:58, injuring ten people. The hypocentre of the earthquake (39.8°N, 140.9°E, focal
depth 10 km, origin time 16:58) was located some 400 km north of Tokyo. The magnitude of the mainshock
was 6.1 on the Japan Meteorological Agency (JMA) scale. Kakkonda hydropower station is located about
20 km north-west of Morioka city in Iwate prefecture, just south of Mount Iwate – a few kilometres to the
south-east of the epicentre.

In Shizukuishi, a town 15 km to the south of the tunnel location, an earthquake with JMA intensity of 7
was registered. Figure 2.6(a) shows the location of the tunnel damage and nearby trench excavations. A
hot spring well intersecting the fault 700 m to the north of the power station was damaged at 180 m depth
by the fault motion (Figure 2.6(b)).

The focal depth of this earthquake was shallow enough for its fault rupture planes to appear on the ground
surface, which is quite unusual for earthquakes of this magnitude. The fault rupture planes were found
roughly lined up at irregular intervals. A concrete outlet tunnel of the second Kakkonda hydroelectric power
station crossed the fault rupture plane (see Figure 2.6). Inevitably, the circular tunnel wall cracked into several
large parts and smaller fragments. The motion of the tunnel relative to the surrounding soil was probably
prevented by the power station; thus, the vertical and left lateral fault motion led to a combination of large
axial and radial stresses in the tunnel. Cracks were observed along 40 m of the tunnel from the power station
to the point where the fault crossed it (Figure 2.7). Some parts of the cracked wall were pushed into the tunnel,
reducing its cross-section. A 2 m high by 2 m wide section completely collapsed onto the invert, allowing soil
with boulders (30–50 cm) to enter the tunnel.

The tunnel was subsequently repaired with steel mesh and concrete, and the ground around the damaged part
was grouted.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

Figure 2.6 (a) Overview of the Kakkonda outlet tunnel, the Shinozaki fault surface rupture, the power station, the
tunnel damage location (splash) and two trench locations. (b) A hot spring well intersecting the fault was damaged
at 180 m depth by the fault motion. Reproduced from Johansson and Konagai (2007). Courtesy NICEE (IIT Kanpur)

Damaged hot
spring well

Trench A

Outlet tunnel

Trench B

Shinozaki fault

Outlet tunnel

Surface rupture
0 100 200 300 m

(a) (b)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 2.7 Cross-section of the tunnel and the power station (dashed-line box), and the tunnel crack map and
cross-sections of the cracked tunnel. Figure taken from Johansson and Konagai (2007). Courtesy NICEE (IIT Kanpur)

TD 0.0 m

TD 10.7 m

TD 35.7 m
TD 38.7 m
TD 40.7 m

2.8. Case study: quantiﬁcation of the PGA for locations in India where
strong-motion records are unavailable
The PGA at a site varies with the magnitude of an earthquake, its depth of focus, and the distance from a given
point to the epicentre or the closest point on the rupturing fault plane. Various empirical relationships exist to
evaluate the PGA: see, for example, Ambraseys et al. (2005a, 2005b). Other sources can be found in Section 3.4.3
in Kramer (1996). However, determining the PGA is a problem in low-seismicity areas and where past earthquake
records are not available. A solution is to obtain the ground shaking characteristics for a particular location from
the Global Seismic Hazard Assessment Program (GSHAP) map, which is described below.

2.8.1 PGA from GSHAP studies

The GSHAP was a major project of the UN/International Decade of Natural Disaster Reduction, conducted in
1992–1998 with the goal of improving global standards in seismic hazard assessment. The GSHAP provided
a global seismic hazard map that was compiled by combining the regional maps produced for different
GSHAP regions and test areas. The map depicts the global seismic hazard as the PGA with a 10% chance of
exceedance in 50 years, corresponding to a return period of 475 years. The basic elements of modern
probabilistic seismic hazard assessment used in the GSHAP were (SED, 2018)

■ Earthquake catalogues and databases: the compilation of a uniform database and a catalogue of
seismicity for the historical (pre-1900), early instrumental (1900–1964) and instrumental periods
(1964 to the present).

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

■ Earthquake source characterisation: the creation of a master seismic source model to describe
the spatial–temporal distribution of earthquakes, using evidence from earthquake catalogues,
seismotectonics, palaeoseismology, geomorphology, mapping of active faults, geodetic estimates
of crustal deformation, remote sensing and geodynamic models.
■ Strong seismic ground motion: the evaluation of ground shaking as a function of earthquake size and
distance, taking into account propagation effects in different tectonic and structural environments and
using direct measures of the damage caused by the earthquake (the seismic intensity) and instrumental
values of ground motion.
■ Computation of seismic hazard: the computation of the probability of occurrence of ground shaking in
a given time period, to produce maps of seismic hazard and related uncertainties at appropriate scales.

Figure 2.8 shows the predicted PGA motion for South Asia obtained from the GSHAP. Figure 2.9 shows the
seismic hazard map of India and adjoining regions for 10% probability of exceedance in 50 years having a
contour interval of 0.05g based on the GSHAP. Details of the studies can be found in Bhatia et al. (1999).

Figure 2.8 Predicted PGA for South Asia (Bhatia et al., 1999)

60°E 90°E 120°E 150°E

45°N 45°N

30°N 30°N

15°N 15°N

0° 0°
60°E 90°E 120°E 150°E

0.0 0.2 0.4 0.8 1.6 2.4 3.2 4.0 4.8

PGA: m/s2

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 2.9 Seismic hazard map of India and adjoining regions for 10% probability of exceedance in 50 years
(contour interval 0.05g) (Bhatia et al., 1999)

65° 70° 75° 80° 85° 90° 95° 100° 105°

40° 40°

35° 35°

30° 30°

25° 25°

20° 20°

PGA: g
0.50
0.45
0.40
0.35
0.30
10° 0.25 10°
0.20
0.15
0.10
5° 0.05 5°
0.00

0° 0°
65° 70° 75° 80° 85° 90° 95° 100° 105°

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

Table 2.3 lists the PGA for some cities in India based on GSHAP studies. The seismic zone factor is also
given in this table for comparison. Figure 2.10 shows the seismic zonation of India based on IS 1893
(BIS, 2002). It should be mentioned that GSHAP maps are based on the tectonic map of India, shown
in Figure 2.11. Details of the study and its basis can be found in Bhatia et al. (1999).

Figure 2.12 shows hazard maps for Europe, Japan, China and the UK.

Figure 2.12(a) is based on European Project The EU-FP7 SHARE Project. Reference www.share-eu.org.

Figure 2.12(b) is based on the work of The Headquarters for Earthquake Research Promotion (HERP), a
governmental organization.

Table 2.3 PGAs for a few cities in India based on the GSHAP hazard map

Location PGA from the Seismic zone in Zone factor Z in

GSHAP map: g IS 1893-1:2002 IS 1893-1:2002

Ahmadabad 0.15–0.20 III 0.16

Bhuj 0.15–0.20 V 0.36
Chennai 0.00–0.05 III 0.16
Dehradun 0.25–0.30 III 0.16
Delhi 0.00–0.05 IV 0.24
Guwahati 0.30–0.35 V 0.36
Hyderabad 0.00–0.05 II 0.10
Kolkata 0.05–0.10 III 0.16
Lucknow 0.00–0.05 III 0.16
Mumbai 0.05–0.10 III 0.16
Patna 0.05–0.10 IV 0.24
Srinagar 0.30–0.35 V 0.36

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 2.10 Seismic zoning of India as laid in IS 1893 (BIS, 2002)

68° 74° 78° 80° 84° 88° 92° 96°

36° 36°

V
IV 32°
32°
V

V
28°
28° IV

V
IV V
II III 24°
24°

V IV

II IV
III 20°
20°
II

16°
IV III
16°

II Zone II (PGA = 0.10g)

Zone III (PGA = 0.16g)
12°
Zone IV (PGA = 0.24g) V
12° Zone V (PGA = 0.36g)

III

8°
8°

72° 76° 80° 84° 88° 92°

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

Figure 2.11 Generalised tectonic map of India and adjoining regions according to the GSHAP. Abbreviations of
tectonic features: ANR, Andaman Nicobar Ridge; CB,Cuddapah Basin; DF, Dharwar Fold; GG, Godavari Graben;
HF, Herat Fault; IBR, Indo Burma Ranges; ITSZ, Indus Suture Zone; KF, Kunlum Fault; KKF, Karakuram Fault; MBT,
Main Boundary Fault; MCT, Main Central Thrust; MG, Mahanadi Graben; SDG, Satpura Damodar Graben; NH,
Naga Hills; NSL, Narmada Son Lineament; PF, Panvel Flexure; RRFZ, Red River Fault Zone; SF, Sagaing Fault; SR,
Sulaiman Range; SP, Shillong Plateau; TPF, Three Pagodas Fault; WAF, West Andaman Fault; WCF: Wang Chao
Fault; XF: Xian Shui He Fault; YZS: Yarlung Zangpo Suture

68° 72° 76° 80° 84° 88° 92° 96° 100° 104°
SH

LT
KU

H FA U
DU

TA G 36°
YN
HIN

36° KK A LT
PAMIR F KF
HF
Peshawar
Srinagar XF 32°
LT

TIBETAN
32° H
U

PLATEAU
FA
CHAMAN

M M ITS
MG

CT A Z YZS
SR

L
Delhi A 28°
28° GA AY
Jaisalmer NG Kathmandu
A
LI
KIRTHA

C BAS
BK IN

ARKAN YOMA NH
AV

CAMBAY SP
AR

RRFZ
GRABEN
R

SDG
24°
IN DIAN 24°
M Calcutta
N SL G

LD G
IBR

20° SHIE IN 20°

LD
Bombay FO
PF

DC AT
GB

H
Koyna G
N
AL

16° ER WCF 16°

ENG

ST
S A G A IN

EA
AB

TGF
DF

OF B
IAN

Madras
BAY

12° 12°
SEA

ANR

WA F

8° 8°
SF

4° 4°
INDIAN OCEAN

0° 0°
68° 72° 76° 80° 84° 88° 92° 96° 100° 104°

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 2.12 (a) European hazard map with 10% probability of exceedance in 50 years. (b) Distribution map for
Japan of the probability of ground motion equal to or larger than seismic intensity 6 or lower occurring within
30 years. (c) Bedrock PGA hazard map for China with a 10% probability of exceedance in 50 years (Li et al.,
2014). (d) Rock outcrop PGA hazard map for the UK for a 475-year return period (Musson and Sargeant, 2007)

Reykjavik

PGA: g
10% exceedance probability in 50 years
0.0 0.1 0.2 0.3 0.4 0.5
Helsinki
Oslo Low Moderate High hazard
Stockholm Tallinn

Edinburgh Riga
Belfast
Dublin Copenhagen
Vilnius

London Amsterdam Berlin Warsaw

Brussels

Prague
Paris

Bratislava Chisinau
Vienna
Bern Budapest
Ljubijana
Zagreb Bucharest
Belgrade
Sarajevo

Lisbon Madrid Podgorica Sofia

Skopje Istanbul
Rome Ankara
Ticana

Athena

(a)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

Figure 2.12 (Continued )

128°E 130°E 144°E 146°E

132°E 134°E 136°E 138°E 140°E 142°E
46°N

146°E 148°E 150°E

46°N

44°N

42°N

40°N
100 km

38°N

36°N

34°N

140°E 142°E
126°E 128°E 28°N
32°N 28°N

124°E 131°E 132°E 26°N

30°N 26°N 26°N

24°N
28°N 24°N 24°N

0% 0.1% 3% 6% 2.6% 100%

(b)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 2.12 (Continued )

PGA: g
≥0.40
0.30
0.20
0.15
0.10
0.05
<0.05
0 250 500 750 km

(c)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

Figure 2.12 (Continued )

9°W 8°W 7°W 6°W 5°W 4°W 3°W 2°W 1°W 0° 1°E 2°E

PGA: g
59°N 59°N
0.00−0.02
0.02−0.04
0.04−0.06
0.06−0.08
58°N 58°N
0.08−0.10
0.10−0.12

57°N 57°N

56°N 56°N

55°N 55°N

54°N 54°N

53°N 53°N

52°N 52°N

51°N 51°N

50°N 50°N

49°N 49°N
9°W 8°W 7°W 6°W 5°W 4°W 3°W 2°W 1°W 0° 1°E 2°E
(d)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

2.9. Quantiﬁcation of fault movement

Fault movement during a seismic event is essentially characterised by the amount of ground displacement.
There are many established empirical relationships for estimating fault displacement (df) according to
earthquake moment magnitude (M). Wells and Coppersmith (1994) proposed the following relationships
based on the pervious and recent earthquake data for various fault types

strike slip fault:

log df = −6:32 + 0:90M (2:14)

normal fault:

log df = −4:45 + 0:63M (2:15)

reverse fault:

log df = −0:74 + 0:08M (2:16)

poorly known fault or blind fault:

log df = −4:80 + 0:69M (2:17)

2.10. Case study: hazard to pipelines due to permanent ground

deformations
Pipelines are often referred to as lifelines since they convey water, gas, oil, sewage and other ﬂuids that play a
pivotal role in our lives. They are located over large geographical regions, and generally buried below ground
for environmental, safety, economic and aesthetic reasons. Therefore, they are subjected to a wide variety of
soil proﬁles and hazards caused by earthquakes. Permanent ground deformations (PGDs) caused by earth-
quakes, faults, landslides, lateral spreading and seismic settlement, severely affect the performance of buried
pipelines. Past earthquakes highlighted the vulnerability of buried pipelines to PGDs. Pipelines experienced
different types of failure modes such as local buckling, beam buckling, tension failure and joint failure. These
failures affect economies, societies and industries. Therefore, the earthquake-resistant design of buried
pipelines is one of the most important aspects in geotechnical and structural engineering. Figure 2.13 shows
various ground rupture patterns – strike-slip, normal and reverse faults – with a pipe crossing the faults.

Figure 2.13 Earthquake-induced ground rupture patterns: (a) strike-slip fault, (b) normal fault and (c) reverse fault

Pipeline
α

–α
βn βt

(b) (c)
α: pipe–fault intersection angle for strike-slip faults
(a) βn: pipe–fault intersection angle for normal faults
β t: pipe–fault intersection angle for reverse faults

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

The hazard parameter in this analysis is the predicted fault displacement. Figure 2.14 shows a numerical
model that can be used to analyse the problem where the pipe–soil interaction is modelled using Winkler
springs. Figure 2.15 shows the effect of a landslide on pipelines, and a model to analyse the problem is shown
in Figure 2.16.

Figure 2.14 Simple ﬁnite-element model of a buried pipeline crossing a strike-slip fault

β
Fault Deflected pipe

Axial soil
δf
springs
δx δx δx δx δx δx δx δx

Initial pipe position δy δy δy δy δy δy δy δy

Lateral soil springs

Figure 2.15 The effect of a landslide on pipelines depends on their orientation: (a) perpendicular crossing,
(b) oblique crossing and (c) parallel crossing (Demirci et al., 2018)

Pipelines subject mainly to bending Pipeline

Pipelines subject to bending and
tension

Pipelines subject to bending and

compression

(a) (b)

Pipelines subject to bending and

tension
Pipelines subject to bending and
compression

(c)

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 2.16 Simple ﬁnite-element model of a buried pipeline crossing a landslide

Deflected pipe
Axial soil
springs

Initial pipe position δy δy δy δy δy

Lateral soil springs

Aside. Two different analysis methods may be used to obtain the pipeline response to PGDs: ﬁnite-element
(FE) models and analytical models. FE software such as ANSYS and ABAQUS are commonly used to
investigate the response of buried pipelines, since the FE method is one of the most powerful numerical tools.
FE models can take into account geometric non-linearity, non-linear soil/pipe behaviour and contact between
the soil and the pipe. FE models can be grouped into two categories, as below.

■ Simple FE models. In this approach, pipelines are considered as beams on Winkler foundations.
Pipelines are modelled as beam elements while soil–pipe interaction is modelled using elasto-plastic
soil springs. The elasto-plastic soil springs are characterised by the ultimate soil force and yield
displacements. The corresponding ultimate soil force and yield displacements are found in guidelines
by Nyman (1984) and the American Lifelines Alliance (2001). Simple FE models for a buried
pipeline crossing a strike-slip fault and a landslide are depicted in Figures 2.14 and 2.16. The pipeline
is modelled as a beam element, and cross-section dimensions such as the pipe diameter (D) and pipe
wall thickness (t) are assigned to beam elements. The non-linear stress–strain behaviour of the pipe
material is assigned to beam elements. Since this is a large deformation problem, the plastic behaviour
of the pipe material must be considered in the model. Soil–pipe interaction is modelled using elasto-
plastic soil springs as suggested in the guidelines above, and further details are provided in Chapter 5.
Axial and lateral soil springs are used for a pipeline crossing a strike-slip fault. In contrast, axial and
vertical soil springs must be used for a pipeline crossing normal and reverse faults. Figures 2.14 and
2.16 show elasto-plastic soil springs in the lateral and axial directions.
■ Advanced FE models. Advanced FE models are three-dimensional (3D) simulations of pipeline crossing
PGD zones. Pipelines are modelled as shell elements, while solid continuum elements are used to
model the soil. An appropriate constitutive model for soil, for example an elastic–perfectly plastic
Mohr–Coulomb model in ABAQUS, is employed to simulate soil behaviour. The parameters needed to
simulate soil behaviour are Young’s modulus E, Poisson’s ratio n, the internal friction angle f, the
dilatation angle y and cohesion c. The pipe diameter D and pipe wall thickness t and elasto-plastic
behaviour of the pipe material are assigned to shell elements. The interaction between the soil and pipeline
is simulated by using contact algorithms. Tangential contact using penalty frictional contact is used to
model friction between the soil and pipe. Normal contact with hard contact is used to allow separation
between the soil and the pipe. The friction coefﬁcient μ between the soil and the pipe is generally selected
as 0.3 in the literature (Vazouras et al., 2010). Appropriate values for the cross-section dimensions and
length in the FE model are required in order to minimise boundary effects. A parametric study should be
carried out to choose appropriate dimensions for the FE model. Figure 2.17 shows an example of an
advanced FE model of a pipeline crossing a strike-slip fault with a 20° fault crossing angle.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Figure 2.17 (a) Cross-section of a 3D FE soil continuum model, (b) plan view of the model showing displacement of the moving block and the ﬁxed block,
(c) the displacement proﬁle of the pipeline and (d) longitudinal pipe strains in the dashed red zone (Demirci et al., 2018)

(a)
z
2.5 m
5m

x 10 m

(b)

Moving block 20°

Fixed block

Equivalent Equivalent
boundary springs (c) boundary springs

(d) LE, LE22

SNEG (fraction = –1.0)
(Avg: 75%)
+1.048e-02
+9.327e-03

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
+8.179e-03
+7.030e-03
+5.882e-03
+4.734e-03
+3.586e-03
+2.437e-03
+1.289e-03
+1.408e-04
–1.007e-03
–2.156e-03
–3.304e-03
Basic concepts of engineering seismology and seismic hazard analysis

59
Seismic Design of Foundations

2.11. Quantiﬁcation of the duration of an earthquake

The duration of strong earthquake ground motion is one of the main parameters characterising natural seismic
hazards, especially for analysis of liquefaction, partial liquefaction and soil–structure interaction, transient
behaviour and so on. It is also important because the amount of cumulative damage that structures may incur
increases with the number of cycles of loading. Seismologists have proposed several deﬁnitions to quantify
the duration of strong motion. The most widely accepted deﬁnitions are

■ significant duration – defined as the time (t) required to build up from 5% to 95% of the integral
acceleration (a) record ∫a2(t) dt (Trifunac and Brady, 1975)
■ bracketed duration – defined as the time elapsed between the first and last occurrence of the
earthquake time record above a certain value such as 0.03g (Ambraseys and Sarma, 1967) or 0.05g
(Bolt, 1973).

An empirical correlation by Dobry et al. (1978) to estimate the signiﬁcant duration (Ts) based on the mag-
nitude (M) of an earthquake is

log (Ts ) = 0:423M − 1:83 (2:18)

It was developed for the western USA, with earthquake magnitude ranging from 4.5 to 7.6.

2.12. Summary
Apart from being spotted, earthquake-induced hazards also need to be quantified. Examples of quantifi-
cation include: peak ground acceleration (PGA), moment magnitude of earthquake, amount of peak
ground displacement (PGD), maximum fault displacement, duration of input motion, height of tsunami etc.
For foundation design in a liquefiable grounds, the important parameters are PGA and moment magnitude.
Examples of the predicted PGA for a few countries are provided in this chapter: Figure 2.8 (for India),
Figure 2.12(a) for Europe, Figure 2.12(b) for Japan, Figure 2.12(c) for China and Figure 2.12(d) for UK.

For designing long linear structures such as pipelines or tunnels, one needs to quantify the amount of fault
movement and some guidelines are provided in Section 2.9. For carrying out transient analysis of a structure
(i.e. how long the ground shaking will continue) one needs to quantify the duration of an earthquake. While
some correlation are provided in this chapter, the readers are referred to latest and specialist literature on
quantiﬁcation of such hazards. For certain type of analysis (time-history analysis), earthquake motion is
required and this aspect is covered in Chapter 3.

REFERENCES
Ambraseys NN and Sarma SK (1967) The response of earth dams to strong earthquakes. Géotechnique 17(3): 181–213.
Ambraseys NN, Douglas J, Sarma SK and Smit PM (2005a) Equations for the estimation of strong ground motions
from shallow crustal earthquakes using data from Europe and the Middle East: horizontal peak ground
acceleration and spectral acceleration. Bulletin of Earthquake Engineering 3(1): 1–53.
Ambraseys NN, Douglas J, Sarma SK and Smit PM (2005b) Equations for the estimation of strong ground motions
from shallow crustal earthquakes using data from Europe and the Middle East: vertical peak ground acceleration
and spectral acceleration. Bulletin of Earthquake Engineering 3(1): 55–73.
American Lifelines Alliance (2001) Seismic Fragility Formulations for Water Systems: Guideline. American
Lifelines Alliance, Washington, DC, USA.
Bhatia SC, Kumar MR and Gupta HK (1999) A probabilistic seismic hazard map of India and adjoining regions.
Annals of Geophysics 42(6), 10.4401/ag-3777.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Basic concepts of engineering seismology and seismic hazard analysis

BIS (Bureau of Indian Standards) (2002) IS 1893-1. Indian standard criteria for earthquake resistant design of
structures. Part 1: General provisions and buildings. BIS, New Delhi, India.
Bolt BA (1973) Duration of strong ground motion. Proceedings of the 5th World Conference on Earthquake
Engineering, Rome, Italy, vol. 1, pp. 1304–1313.
Demirci HE, Bhattacharya S, Karamitros DK, Alexander NA and Singh RM (2018) Finite element model of buried
pipelines crossing strike-slip faults by ABAQUS/Explicit. 16th European Conference on Earthquake
Engineering, Thessaloniki, Greece.
Dobry R, Idriss IM and Ng E (1978) Duration characteristics of horizontal components of strong-motion
earthquake records. Bulletin of the Seismological Society of America 68(5): 1487–1520.
Dziewonski AM, Chou TA and Woodhouse JH (1981) Determination of earthquake source parameters from
waveform data for studies of global and regional seismicity. Journal of Geophysical Research: Solid Earth
86(B4): 2825–2852.
Gutenberg B and Richter CF (1944) Frequency of earthquakes in California. Bulletin of the Seismological Society
of America 34(4): 185–188.
Hanks TC, Hileman JA and Thatcher W (1975) Seismic moments of the larger earthquakes of the southern
California region. Geological Society of America Bulletin 86(8): 1131–1139.
Johansson J and Konagai K (2007) Fault movement related damage examples and fault provisioned design case
histories. Design of Foundations in Seismic Areas: Principles and applications, Bhattacharya S (ed), NICEE,
ISBN:81-904190-3: pp. 17–46.
Joyner WB and Boore DM (1981) Peak horizontal acceleration and velocity from strong-motion records including
records from the 1979 Imperial Valley, California, earthquake. Bulletin of the Seismological Society of America
71(6): 2011–2038.
Kanamori H (1983) Magnitude scale and quantiﬁcation of earthquakes. Tectonophysics 93(3): 185–199.
Kanamori H and Anderson DL (1975) Theoretical basis of some empirical relations in seismology. Bulletin of the
Seismological Society of America 65(5): 1073–1095.
Konagai K (2005) Data archives of seismic fault-induced damage. Soil Dynamics and Earthquake Engineering
25(7–10): 559–570.
Kramer SL (1996) Geotechnical Earthquake Engineering. Prentice-Hall, Upper Saddle River, NJ, USA.
Li J, Wang Y, Chen H and Lin L (2014) Risk assessment study of ﬁre following an earthquake: a case study of
petrochemical enterprises in China. Natural Hazards and Earth System Sciences 14(4): 891.
McGuire RK and Arabasz WJ (1990) An introduction to probabilistic seismic hazard analysis. Geotechnical and
Environmental Geophysics 1: 333–353.
Musson RMW and Sargeant SL (2007) Eurocode 8 Seismic Hazard Zoning Maps for the UK. British Geological
Survey, Keyworth, UK, Technical Report CR/07/125.
SED (Swiss Seismological Service) (2018) Global Seismic Hazard Assessment Program. http://static.seismo.ethz.ch/
gshap/index.html (accessed 06/08/2018).
Trifunac MD and Brady AG (1975) A study on the duration of strong earthquake ground motion. Bulletin of the
Seismological Society of America 65(3): 581–626.
Vazouras P, Karamanos SA and Dakoulas P (2010) Finite element analysis of buried steel pipelines under strike-slip
fault displacements. Soil Dynamics and Earthquake Engineering 30(11): 1361–1376.
Wegener A (1915) Die Entstehung der Kontinente und Ozeane. Vieweg, Braunschweig, Germany.
Wells DL and Coppersmith KJ (1994) New empirical relationships among magnitude, rupture length, rupture width,
rupture area, and surface displacement. Bulletin of the Seismological Society of America 84(4): 974–1002.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

ISBN 978-0-7277-6166-8
https://doi.org/10.1680/sdof.61668.063
ICE Publishing: All rights reserved

Chapter 3
Selection of strong motion for
foundation design
3.1. Introduction
Engineers often need to use site-specific earthquake ground motions for designing foundations. The seismic
design of foundations requires non-linear time history analysis to evaluate soil–structure interaction effects.
This requires the selection of an appropriate time history as the input. For countries or sites (e.g. Japan) where
earthquake recorded data are available (in other words, where fault characteristics to some extent have been
identified), manipulation of these records can generate the required site-specific motion to the satisfaction of
most stakeholders. However, where these records are unavailable (for countries such as India or China),
artificial ground motions are generated for analysis and design purposes. Several methods are used in practice
for the generation of these artificial ground motions. These approaches include

■ direct use of strong motion recorded elsewhere having a similar expected earthquake magnitude at the
site under consideration
■ scaling of strong motion records to the expected peak bed rock acceleration (known as PGA) expected at the
location
■ code-speciﬁed spectrum-compatible ground motion.

This chapter discusses all these approaches and provides a comparative study on the performance of these
three approaches for a typical site where response spectra and the expected PGA are given.

3.2. Background to the problem

Engineers need to select or develop site-speciﬁc ground motions to design foundations for important
structures that require non-linear time history analysis. Numerous design codes and research publications
provide guidance for the selection of seismic ground motions for use in this kind of analysis. These docu-
ments are mostly based on the use of a number of real recorded ground motions, scaled by a constant factor
over the full duration, to be approximately consistent with the expected seismic hazard at the site. Seismo-
logical characteristics of the records, such as the earthquake magnitude and epicentral distance, are usually
considered in the selection of real records, as they are likely to inﬂuence the shape of the response spectrum,
the energy content and the duration of ground shaking, and therefore the expected demand on structures.

A typical foundation design process in seismic areas is shown in Figure 3.1. After identifying the expected
seismic hazard at the site, the selection of appropriate earthquake records forms an important step in the
process. The parts of the design methodology where these records are used are highlighted in the design chart,
along with some typical design standard software (examples include SIREN, DEEPSOIL SHAKE and
DYNA). In selecting these records, the only data available to the engineers through their seismologist col-
leagues comprise the earthquake source mechanisms in the vicinity of the region of interest (location of the
fault and the likely rupture characteristics) and, in some cases, some recorded data of past earthquakes in that
region. For cases where recorded data are available, the manipulation of these records can generate the
required site-speciﬁc motion to the satisfaction of the stakeholders.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 3.1 Flowchart showing a seismic design methodology for foundations in seismic areas. (Adapted from
Ghosh and Bhattacharya, 2008)

Identification of geohazards
(at the site)

Probabilistic seismic hazard

study determine bedrock
spectra

Bedrock time histories

(seed selection, RspMatch,
SEISMOMATCH, WAVGEN)

However, where these records are unavailable, artiﬁcial ground motions have to be generated for analysis and
design purposes. In these cases, engineers are left with two pointers to generate an input motion

■ country- or region-speciﬁc codes of practice that will provide a response spectrum for different type of
soils (typically soft soil, stiff soil and rock)
■ the expected PGA at the location.

This is often the situation for most developing countries that are prone to seismic hazard, as a sufﬁcient
number of recorded motion is usually unavailable.

Several methods are used in practice for the generation of these artiﬁcial ground motions. The approaches
include

■ direct use of strong motion recorded elsewhere having a similar expected earthquake magnitude to the
site under consideration
■ scaling of strong motion records to the expected PGA expected at the location
■ code-speciﬁed spectrum-compatible ground motion.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Selection of strong motion for foundation design

3.3. Criteria for the selection of earthquake records

Ground motion records to be used in geotechnical analysis should represent the potential earthquake hazards
at the site; that is, consideration of the magnitude, distance, site conditions, source mechanism, directivity and
other effects obtained from seismic hazard analysis. In this regard, the appendix of ISO 19901-2 (ISO, 2017)
states that a minimum of four records should be used for analysis, and that each record should be selected
according to the following criteria

Given the magnitude and distance of events dominating Extreme Level Earthquake (ELE) ground
motions, the earthquake records for time history analysis can be selected from a catalogue of historical
events. Each earthquake record consists of three sets of tri-axial time histories representing two
orthogonal horizontal components and one vertical component of motion. In selecting earthquake
records, the tectonic setting (e.g. faulting style) and the site conditions (e.g. hardness of underlying
rock) of the historical records should be matched with those of the structure’s site.

A similar approach is also suggested in EN 1998-1 (CEN, 2004). The Eurocode allows the use of any form of
accelerogram for analysis, such as real (unscaled), artiﬁcial or obtained by simulation of the seismic source,
propagation and site effects.

Three approaches are often used, and each method has its advantages and disadvantages – discussed later in
this chapter through an example.

Method 1: direct use of the strong motion record

In this method, a strong motion is chosen to match the seismicity of the location, either in terms of the moment
magnitude (Mw) or the expected PBRA. It may be difﬁcult to ﬁnd a recorded motion satisfying both the
requirements. For example, a site may expect an Mw 6.0 earthquake and a PBRA of 0.2g. Therefore, a
designer has to look for the recorded data matching these characteristics. This method is rarely used in
practice, as it is quite unlikely that the shape of the spectra will match the expected bedrock spectra at a
particular site. This will be illustrated later using an example.

Method 2: scaling of the strong motion record to the expected PBRA

In this method, an earthquake record is chosen from a database to match Mw, and then the maximum
amplitude is scaled to match the expected PBRA.

Method 3: intelligent scaling or code-speciﬁed spectrum-compatible motion

Traditionally, seismic hazard at a site for design purposes has been represented by design spectra. Thus, all
seismic design codes and guidelines require scaling of the selected ground motion time histories so that they
match or exceed the controlling design spectrum within a period range of interest. Figure 3.2 shows an
example to illustrate the outcome of the methodology by considering the 1999 Ch-Chi earthquake record as
an example. The unaltered motion is plotted and compared with the response spectra in IS 1893 (BIS, 2002)
for hard soil or rock. A spectrum-compatible motion is also plotted, and it may be noted that the plot faithfully
matches the spectrum – hence the name of this method. Note: spectrum here refers to response spectrum and
details can be found in Chapter 5 (Section 5.4).

Several methods of scaling time histories are available. Essentially, in all these methods an input motion is
intelligently selected from a strong motion database and the motion is manipulated in such a way so as to
obtain a motion that matches the target design spectra. The manipulation can be carried out in the frequency
domain where the time-frequency content of the recorded ground motions is manipulated in order to obtain a
good match. Mathematically, following Alexander et al. (2014), this can be expressed as follows: a process

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 3.2 An example of earthquake spectra of selected motion matched to code-speciﬁed spectra

6
Target spectra (IS-1893:2002)
Rock or hard soil
5 Chi Chi Taiwan original spectra
Scaled
RSPMatch spectra
Spectral acceleration: m/s2

0
0.01 0.1 1 10
Period: s

transforming a known signal or input motion x(t) into a similar input motion y(t) that matches the spectrum.
The additional aim is that the transformed signal y(t) should maintain very similar non-stationary charac-
teristics such as the general envelope, the time location of large pulses and the variation of the frequency
content with time. In other words, the target signal y(t) should look for all intents and purposes like a real
earthquake.

Aside: what is the attraction of scaling input motion in the way as described above?

■ One of the main sources of uncertainty in earthquake engineering is the earthquake itself; that is,
its spatiotemporal incidence and time-series. Bounding this uncertainty in the predicated ground
motion time series is a complex question. As a result, design codes around the world concentrate on
satisfying a response spectrum of some kind. However, for major or critical infrastructure projects
(e.g. a nuclear reactor or important lifeline structures such as bridges or hospitals), ground motion
time series are needed for time history analyses. These are required for serviceability limit state
considerations rather than for the ultimate limit state. Now the question is, what input motions to
choose and how many of them are required? The brute-force method is to randomly sample all
accelerograms in order to obtain a reasonably good estimate of the population statistics. This, of
course, would require thousands or even hundreds of thousands of records to be effective. Therefore,
a selective or stratiﬁed sampling of a small set of real accelerograms is selected that satisﬁes certain
geophysical and structural criteria.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Selection of strong motion for foundation design

■ Spectrally matched records are therefore often used in practice. This approach can be easily criticised,
as it modiﬁes the recorded time series by adjusting its time–frequency content. However, it is also
considered overly conservative as it matches a ﬂat and smooth broadband spectrum, unlike most real
earthquakes.
■ The perceived attractiveness of spectrally matched records is that fewer accelerograms need be used in
time history analyses.

3.4. Methods for spectral matching

3.4.1 Perturbation approach
Most of these methods are based on frequency domain techniques that alter the Fourier amplitude (but not
phase) spectrum of the original ground motion so as to match a target response spectrum. A ﬂaw of this
approach is that it changes the non-stationary nature of the records. Lilhanand and Tseng (1987) introduced a
perturbation technique based on Duhamel’s integral that modiﬁes the original signal into a spectrum-
matching one. The perturbation added was localised in time so that it attempted to maintain similar non-
stationary statistics to the original. Abrahamson (1992) developed the approach discussed in Lilhanand and
Tseng (1987), and produced a software called RspMatch, which was further developed by Hancock et al.
(2006) in RspMatch2005. This algorithm is embedded in the SeismoMatch software.

3.4.2 Wavelet-based approach

Mukherjee and Gupta (2002) suggested a ‘wavelet’-based modiﬁcation to the original time series, and
developed the computer code WAVGEN. A wavelet-based procedure uses the decomposition of a recorded
accelerogram into the desired number of time histories with non-overlapping frequency content, and then
each time history is suitably scaled to match the response spectrum of the revised accelerogram with the
speciﬁed design spectrum suggested by a code of practice. An example is presented later in this chapter
showing the output from such an analysis.

3.4.3 Volterra series approach

This is a novel algorithm developed by Alexander et al. (2014) whereby any accelerogram can be morphed
into a spectrum-matching one. The seed time series is ﬁrst decomposed into a set of wavelet levels, and these
frequency-banded components are subsequently polynomially mixed to form a discrete and truncated
Volterra series expansion. Thus, the original signal is re-expressed as a linear combination of Volterra series
terms. The non-linear process that transforms this original time series into a spectrum-matching one requires
an optimal set of coefﬁcients for this Volterra series. This minimisation problem is solved using the
Levenberg–Marquardt algorithm.

The Volterra series originates as a generalisation of the Taylor series expansion of a function. However, it can
also be considered as a generalisation of the convolution integral – a transformation of an input function in
time by some system. A reliable method of selecting an optimal Volterra model is not generally available.
Therefore, any problem, whether linear or non-linear, is unique and particular to its circumstances. So, the
choice of a Volterra model is, in a sense, arbitrary. In the method of Alexander et al. (2014), the wavelet
transform is employed in order to estimate the Volterra kernels of the series expansion (of x(t)) that maps x(t)
into an inﬁnite set of y(t). Given that y(t) is some Volterra series expansion of x(t), it inherits, to some degree,
the shape and form of x(t). Consequently, from this inﬁnite set of y(t) we seek a particular member that has a
response spectrum that matches some target response spectrum.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Consider a non-linear process that transforms a frequency domain signal X(w) into Y(w). The general form of
this process is given by a Volterra series expansion (Boyd, 1984; Chanerley, 2007; Fa-Long, 2011; Schetzen,
1980), and therefore

X∞ ð ð
1
Y ðw Þ = … Hn ðw1 , ⋯, wn ÞX ðw − w1 Þ ⋯ X ðw − wn Þdw1 ⋯dwn (3:1)
n=1
n!

where Hn is known as the nth Volterra kernel. This can be inverse Fourier transformed into the time domain to
obtain
X∞
1
yðt Þ = hn ðt Þxðt Þn (3:2)
n=1
n !

where y(t) and x(t) are the inverse Fourier transforms of signals Y(w) into X(w). The time domain version of
the nth Volterra kernel Hn is hn. This is similar to a Taylor series but the coefﬁcients are themselves time-
varying functions.

The original signal x(t) is re-expressed as the linear combination of narrow-band frequency components
fi(t) as

X
m
xðt Þ ≃ ai fi ðt Þ (3:3)
i=1

where ai are amplitude coefﬁcients. This decomposition could be a Fourier series, but in the method of
Alexander et al. (2014) a wavelet decomposition technique is employed using the stationary wavelet
transform. In Equation 3.3 the total number of terms is m; that is, m – 1 wavelet detail levels plus one
wavelet approximation level. Hence, technically, the number of levels of wavelet decomposition should
be considered to be m – 1 rather than m.

The more common transform used is the discrete wavelet transform (DWT), which operates using digital
low-pass and high-pass filters and then decimalises the data; however, decimalisation leads to shift variance
and aliasing. In effect the DWT is an octave digital filter. However, the filter banks employed in the DWT will
experience phase and amplitude distortion, and aliasing. The filters are usually overlapping; therefore,
aliasing errors will inevitably occur. It has been shown that the reconstructed signal will be aliased as a
consequence of decimalisation and imaging. Moreover, any processing is performed only on the data
available after down-sampling; therefore, missing data samples are unprocessed. Furthermore, during
reconstruction, the interpolation will proceed on the available processed samples, resulting in some distortion.

The stationary wavelet transform (SWT), also known as the translation invariant transform, operates in a
different manner to the standard DWT. The SWT is shift invariant and does not alias. In this approach the
number of data samples is kept the same because down-sampling is avoided at each level of decomposition.
Instead, the filter impulse response is interpolated to match the filter bandwidths to those of the sub-bands.
The interpolation is performed by inserting zeros in between the filter coefficients – procedure referred to as
the à trous algorithm (‘with holes’ algorithm). It is thus not necessary during reconstruction to up-sample the
data, as when applying the DWT. However, as for the DWT, it is necessary to apply synthesising filters,
which are flipped versions of the analysis filters at each decomposition level.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Selection of strong motion for foundation design

In order to maintain a linear phase characteristic, the biorthogonal wavelet family is recommended, and the
higher-order ‘bior3.9’ wavelet (from the MATLAB Wavelet Toolbox) can be used to ensure the appro-
priate ﬁlter characteristics for wavelet approximation and detail estimation. Substituting Equation 3.3 into
Equation 3.2, we obtain

X
m
1 XX
m m
yðt Þ = ½h1 ðt Þai fi ðt Þ + h2 ðt Þai aj fi ðt Þfj ðt Þ + ⋯ (3:4)
2
i=1 i=1 j=1

or in the frequency domain

m ð
X
Y ðw Þ = ½H1 ðw1 Þai Fi ðw − w1 Þdw1 +
i=1
m ðð
(3:5)
1X m X
H2 ðw1 , w2 Þai aj Fi ðw − w1 ÞFj ðw − w2 Þdw1 dw2 + ⋯
2 i=1 j=1

The above is coded in MATLAB, and is available to download. Alexander et al. (2014) noted that this is an
especially sophisticated method for generating synthetic ground motion.

Aside. While the codes specify the broad aspects of selection (e.g. choosing a certain number of records and
the importance of site conditions), the following are important considerations when choosing the seed
motion.

■ Quality of the recorded motion; that is, the sampling frequency and whether or not all the frequencies
of the strong motion are captured.
■ Seismo-tectonic features (subduction or stable/active continental region) and the style of faulting
(shallow, deep, strike-slip or blind fault).
■ Ground condition; that is, the wave propagation path from the source to the site.
■ In using the spectral matching method, expert judgement is necessary for selection of the duration of
the required input records and requirements for scaling of separate components (two orthogonal and
one vertical) of the ground motion.

Examples of websites that provide archives of strong-motion data are

■ K-Net (Japan): http://www.kyoshin.bosai.go.jp

■ COSMOS (worldwide): https://strongmotioncenter.org
■ PEER (worldwide shallow crustal events): https://ngawest2.berkeley.edu
■ European and Middle East Strong-Motion Data (ISESD): http://isesd.hi.is.

3.5. Case study: application of spectrum-compatible input motion

3.5.1 Kolkata, India
A study of the city of Kolkata (earlier known as Calcutta) in eastern India is considered, to demonstrate the
three approaches discussed in this chapter. The lack of recorded earthquake data and the high seismic risk of
this city are notable features.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Kolkata lies over the Bengal Basin at the boundary of seismic zones III and IV of the seismic zonation map of
India in IS 1893-1 (BIS, 2002). In the past, the city has seen far- and near-source earthquakes due to asso-
ciated Himalayan tectonics. The most damaging earthquake so far has been the 1897 earthquake, which
caused widespread damage in the city. Several faults have been identiﬁed in this region – many of which
show evidence of movement. There are earthquake fault lines barely 100 km from Kolkata. However, in
recent years the city has seen a rapid growth in population as well as construction of major infrastructure
projects. Seismic microzonation of this fast-expanding mega city into seismic zones IV and V is essential for
retroﬁtting old and vulnerable buildings and for the earthquake-resistant design of new constructions and
installations. In trying to understand the seismic hazard of the city, it is worth remembering that the earth-
quake database for India is still incomplete, especially with regard to earthquakes prior to the historical period
(before 1800 AD). According to Global Seismic Hazard Assessment Program (GSHAP) data, the state of West
Bengal falls in a region of low seismic hazard in the south-west that rises steadily towards the east and the
north of the state. The next section lists the parameters for selecting the earthquake ground input motion. The
methodology adopted in the present study involves the following

■ selection of the peak ground acceleration (PGA) at the bed rock level, to represent the seismic hazard
■ development of synthetic ground motions, so as to match the target code-speciﬁed response spectrum
of the site with the minimum number of undershoots across the whole range of the spectrum
■ ground response analysis for different sites in the study area using the synthetically developed ground
motion.

The ground response studies enable the computation of acceleration time histories at the top of each of the soil
layers as well as on the ground surface, ampliﬁcation rating and response spectra at each site. This is covered
in detail in the Chapter 4.

3.5.2 Selection of PGA (Peak Ground Acceleration)

According to the seismic zoning map of India in IS 1893-1 (BIS, 2002), the city of Kolkata lies in moderate
seismic zone (zone III), with a zone factor of 0.16 and an expected Mw of 6–6.5. The GSHAP map, which is
based on the 10% probability of exceedance in 50 years, speciﬁes a maximum PGA of 1.6 m/s2 (0.163g) for
Kolkata. It is of interest to note that the PGA of 0.163g suggested by the GSHAP is the mean of the limits 0.1g
and 0.3g provided by Mohanty and Walling (2008). Hence, in the present study the maximum PGA of 0.163g
has been selected for Kolkata to study the effect of local soil conditions due to seismic excitation.

3.5.2.1 Target bedrock spectrum

The target bedrock spectrum is obtained from IS 1893-I (BIS, 2002), based on the desired PGA (0.163g) at
5% damping for rock or hard soil. This spectrum assumes that the bedrock velocity is at least 760 m/s. In the
absence of a detailed uniﬁed hazard response spectrum calculation, this is a reasonable assumption. This
spectrum is shown in Figure 3.2.

3.5.3 Development of synthetic accelerograms (time histories)

WAVGEN, a wavelet-based method developed by Mukherjee and Gupta (2002), is used in this study to
develop synthetic earthquakes for the Kolkata region. While selecting a suitable strong motion, several
important factors are considered. These include similar magnitude, a peak acceleration close to the target
value, a similar source–site (i.e. fault to site) distance and similar site conditions for wave propagation.
Mohanty and Walling (2008) explored the seismicity of the area by studying the 1897 Shillong earthquake,
the 1934 Bihar earthquake and the 1950 Assam earthquake. They suggest that the possible earthquake

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Selection of strong motion for foundation design

sources that have rocked Kolkata and caused signiﬁcant damage are from the Himalayan frontier and the
Arakan–Yoma ranges. Therefore, based on the tectonic map of India (Figure 3.3), the occurrence of earth-
quakes in Kolkata (Bengal basin) may be considered to be due to thrust fault mechanisms.

Based on these factors, five earthquakes in the range Mw 6.0–6.9 recorded at rock sites (site class A in the US
Geological Survey classification for which the shear wave velocity Vs > 750 m/s or the response spectra for
rock/hard soil were in accordance with IS 1893-I (BIS, 2002)) were selected from the earthquake database
compiled by the Pacific Earthquake Engineering Research Center (PEER), University of California at
Berkeley. The selected earthquakes are representive of earthquakes of similar magnitude for Kolkata and its
surroundings. Table 3.1 lists the five earthquake records and their characteristics. Figure 3.4 shows the time
history of the selected strong motions.

Figure 3.3 Seismo-tectonic features around Kolkata

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Table 3.1 Characteristics of the selected earthquake records

Earthquake Northridge San Fernando Whittier Narrows Loma Prieta Northridge

Date 17/01/1994 09/02/1971 01/10/1987 18/10/1989 17/01/1994

Recording station 127 Lake 127 Lake 24399 Mt Wilson- 47379 Gilroy 90017 LA –
(ST) Hughes #9 Hughes #9 CIT Seista Array #1 Wonderland Avenue
Magnitude Mw = 6.7 Mw = 6.6 Mw = 6.0 Mw = 6.9 Mw = 6.7
ML = 6.6 Ms = 7.6 ML = 5.9 Ms = 7.1 ML = 6.6
Ms = 6.7 Ms = 5.7 Ms = 6.7
PGA, amax: g 0.165 0.157 0.186 0.209 0.172
Closest to fault 26.8 23.5 21.2 11.2 22.7
rupture: km
Style of faulting Thrust fault Reverse fault Thrust fault Thrust fault Thrust fault

ML, local magnitude; Ms, surface wave magnitude

Figure 3.4 Time history of selected strong motion records

0.2
Northridge (ST: 127 Lake Hughes #9)
0.1 amax = 0.165g
Acceleration: g

0.0

–0.1

–0.2
0 5 10 15 20 25 30 35 40
Time: s

0.2
San Fernando
0.1 amax = 0.157g
Acceleration: g

0.0

–0.1

–0.2
0 5 10 15 20 25 30 35 40
Time: s

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Selection of strong motion for foundation design

Figure 3.4 (Continued )

0.2
Whittier narrows
0.1 amax = 0.186g

Acceleration: g
0.0

−0.1

−0.2
0 5 10 15 20 25 30 35 40
Time: s

0.3
Loma Prieta
0.2
amax = 0.209g
Acceleration: g

0.1

0.0

–0.1

–0.2

–0.3
0 5 10 15 20 25 30 35 40
Time: s

0.2
Northridge (ST: 90017 LA − Wonderland Avenue)
0.1 amax = 0.172g
Acceleration: g

0.0

–0.1

–0.2
0 5 10 15 20 25 30 35 40
Time: s

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 3.5 shows the modiﬁed and spectrally matched input motion. The response spectra of the selected
earthquake motions before and after matching the target spectra are shown in Figure 3.6. The peak
accelerations (amax) of the spectrally matched records are indicated in Figure 3.5 for the respective strong
motions.

Figure 3.5 Target-spectra-matched time history of acceleration

0.2
Northridge (ST: 127 Lake Hughes #9)

0.1 amax = 0.154g

Acceleration: g

0.0

–0.1

–0.2
0 5 10 15 20 25 30 35 40
Time: s

0.2
San Fernando
0.1 amax = 0.172g
Acceleration: g

0.0

–0.1

–0.2
0 5 10 15 20 25 30 35 40
Time: s

0.2
Whittier Narrows
0.1 amax = 0.167g
Acceleration: g

0.0

–0.1

–0.2
0 5 10 15 20 25 30 35 40
Time: s

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Selection of strong motion for foundation design

Figure 3.5 (Continued )

0.2
Loma Prieta
0.1 amax = 0.158g

Acceleration: g
0.0

–0.1

–0.2
0 5 10 15 20 25 30 35 40
Time: s

0.2 Northridge (ST: 90017 LA – Wonderland Avenue)

amax = 0.21g
Acceleration: g

0.1

0.0

–0.1

–0.2

0 5 10 15 20 25 30 35 40
Time: s

Figure 3.6 Comparison of the target spectra and the response spectra of the earthquake time history

0.8 Target spectrum

(IS: 1893–2002, rock or hard soil)
0.7
Spectrum of original motion
Spectral acceleration: g

0.6 (Northridge, ST: 127 Lake Hughes #9)

0.5 Spectrum of modified motion

0.4

0.3

0.2

0.1

0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Period: s

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 3.6 (Continued )

0.5
Target spectrum
(IS: 1893–2002, rock or hard soil)
0.4
Spectral acceleration: g Spectrum of original motion (San Fernando)

Spectrum of modified motion

0.3

0.2

0.1

0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Period: s
0.8
Target spectrum
0.7 (IS: 1893–2002, rock or hard soil)

0.6 Spectrum of original motion (Whittier Narrow)

Spectral acceleration: g

Spectrum of modified motion

0.5

0.4

0.3

0.2

0.1

0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Period: s
0.8
Target spectrum
0.7 (IS: 1893–2002, Rock or hard soil)

0.6 Spectrum of original motion (Loma Prieta)

Spectral acceleration: g

0.5 Spectrum of modified motion

0.4

0.3

0.2

0.1

0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Period: s

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Selection of strong motion for foundation design

Figure 3.6 (Continued )

0.5
Target spectrum
(IS: 1893–2002, rock or hard soil)
0.4
Spectrum of original motion

Spectral acceleration: g
(Northridge, ST: 90017 LA − Wonderland Avenue)

0.3 Spectrum of modified motion

0.2

0.1

0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Period: s

3.6. Summary
Engineers often need strong motion for carrying out different types of time history analysis including ground
response analysis (which is covered in Chapter 4). One of the greatest uncertainties in earthquake engineering
are the earthquake themselves! Now the question is what strong motion to choose for time history or ground
response analysis and how many of them are required to arrive at a defensible and meaningful conclusion?
The brute-force method is to randomly sample all accelerograms in order to obtain a reasonably good estimate
of the population statistics. This, of course would require thousands or even hundreds of thousands of records
in order to be effective. Therefore, a selective or stratiﬁed sampling of a small set of real accelerograms can be
selected that satisfy certain criteria. This chapter provides the background behind the selection method, which
is based on response spectrum compatibility.

REFERENCES
Abrahamson NA (1992) Non-stationary spectral matching. Seismological Research Letters 63(1): 30.
Alexander NA, Chanerley AA, Crewe AJ and Bhattacharya S (2014) Obtaining spectrum matching time series
using a reweighted volterra series algorithm (RVSA). Bulletin of the Seismological Society of America 104(4):
1663–1673.
BIS (Bureau of Indian Standards) (2002) IS 1893-1. Indian standard criteria for earthquake resistant design of
structures. Part 1: General provisions and buildings. BIS, New Delhi, India.
Boyd S, Chua LO and Desoer CA (1984) Analytical foundations of Volterra series. IMA Journal of Mathematical
Control and Information 1(3): 243–282.
CEN (European Committee for Standardisation) (2004) EN 1998-1. Design of structures for earthquake resistance.
Part 1: General rules, seismic actions and rules for buildings. CEN, Brussels, Belgium.
Chanerley AA and Alexander NA (2007) Correcting data from an unknown accelerometer using recursive least
squares and wavelet de-noising. Computers and Structures 85(21–22): 1679–1692.
Fa-Long L (2011) Digital Front-End in Wireless Communications and Broadcasting. Cambridge University Press,
Cambridge, UK.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Hancock J, Watson-Lamprey JA, Abrahamson NA et al. (2006). An improved method of matching response
spectra of recorded earthquake ground motion using wavelets. Journal of Earthquake Engineering 10: 67–89
(special issue).
ISO (International Organization for Standardization) (2017) ISO 19901-2. Petroleum and natural gas industries. Speciﬁc
requirements for offshore structures. Part 2: Seismic design procedures and criteria. ISO, Geneva, Switzerland.
Lilhanand K and Tseng WS (1987) Generation of synthetic time histories compatible with multiple design response
spectra. Transactions of the 9th International Conference on Structural Mechanics in Reactor Technology,
Lausanne, Switzerland, vol. K1, pp. 105–110.
Mohanty WK and Walling MY (2008) Seismic hazard in Mega City Kolkata India. Journal of the International
Society for the Prevention and Mitigation of Natural Hazards 46, doi: 10.1007/s11069-007-9195-1.
Mukherjee S and Gupta VK (2002) Wavelet-based generation of spectrum-compatible time-histories. Soil
Dynamics and Earthquake Engineering 22: 799–804.
Schetzen M (1980) The Volterra and Wiener Theories of Nonlinear Systems. Wiley, New York, NY, USA.

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

ISBN 978-0-7277-6166-8
https://doi.org/10.1680/sdof.61668.079
ICE Publishing: All rights reserved

Chapter 4
Ground response analysis
4.1. Introduction
Damage caused by earthquakes varies greatly from location to location, and is governed by local site con-
ditions and the characteristics of the input motion. The issues related to ground motion are discussed in
Chapter 3. Loosely speaking, from the point of view of foundation analysis and design, ground behaviour can
be of two types

■ Liquefaction, whereby the soil momentarily transforms from being a solid-like material to a fluid-like
material. This is typical of loose to medium-dense sand and fill materials. Often, some other intermediate
soils (neither pure sand nor pure clay) may also liquefy. The aspect of liquefaction engineering is dealt
with in detail in Chapters 6 and 7. Methods for liquefaction remediation are discussed in Chapter 12.
■ Amplification of ground motion, which is typical of soft soil sites such as a clay-dominated ground profile.

However, a realistic ground profile can be a mix of different types of soils. These often result from land
reclamation during urbanisation. For example, Port Island area suffered extensive damage during the 1995
Kobe earthquake due to the type of fill material, which liquefied. The aim of this chapter is to provide
concepts to allow the behaviour of ground during earthquakes to be assessed. In other words, how a ground
supporting the foundations will respond to earthquakes.

The behaviour of ground is usually predicted by ground response analysis, which requires a thorough
understanding of the wave propagation pattern from the seismic source (i.e. where the fault ruptured) to the
site of interest. To carry out wave propagation analysis, it is necessary to understand the soil and ground
conditions (e.g. borehole data and geological and geomorphological information). As the extent of seismic
damage or the seismic hazard depends directly on the ground motion, zoning for ground motions is essential
in seismic hazard assessment. The analysis also requires the soil characteristics and, in particular, modulus
reduction curves (i.e. the variation of the shear modulus of the soil G for a range of strains). Damping values
of soils are also necessary. Figure 4.1 shows examples of ground response analysis whereby the expected
ground velocity of an entire country (Japan in this example) can be carried out. Other examples are given in
Chapter 2 (Figure 2.12). This chapter provides the background knowledge necessary for ground response
analysis.

Aside case study: the 2016 Kumamoto earthquake

An active sequence of earthquakes (foreshock, mainshock and aftershocks) hit the Kumamoto area of Japan
in April 2016, resulting in 69 deaths and considerable economic loss. The earthquakes induced numerous
ground failures and cascading geohazards, causing major damage to important infrastructure. Of interest is
the extensive liquefaction that occurred only in some areas of Kumamoto City (Figure 4.2). A rectangular
area of 2 km by 20 m between the Shirakawa River and the Midorikawa River in Kumamoto City liqueﬁed.
Based on the investigation of old city maps, a possible explanation is that the area was an old natural river
dike that had been reclaimed. Figure 4.2 also shows selected borehole proﬁles. Bhattacharya et al. (2018)
reported that the ground consists of alluvium – a relatively new deposit overlying diluvium. The top soil is
black-coloured and organic, possibly of volcanic nature, known locally as Kuroboku.

Seismic Design of Foundations

Figure 4.1 Expected peak ground velocity with 5% probability of exceedance in 50 years. Courtesy NIED Professor
Midorikawa

Seismic Design of Foundations

Ground response analysis

Figure 4.2 Observed liquefaction in an area of Kumamoto following the earthquake, together with sample
borehole data

Borehole 1 N value
0 10 20 30 40 50
0
Borehole 1
Sand
Shirakawa river Silty clay 5

10
Sand

Depth: m
Borehole 2
15

Silt 20

Sand 25

Gravel 30

Borehole 2 N value
0 10 20 30 40 50
0
Sand
Silty clay 5

10
Sand

Depth: m
15

Silt 20
Midorikawa river
Sand 25
Silt
Gravel 30

4.2. Seismic wave propagation in layered media

Past earthquakes – for example the 1985 Mexico earthquake – have shown that the damage caused by
earthquakes is more severe in soft ground than in hard deposits. This is explained by the fact that earthquake
motions amplify more in soft ground. However, it is well known to seismologists that earthquake motion
attenuates with distance from the fault. In this sense, amplification is a curious phenomenon, and was
identified as an important issue following the 1985 Mexico earthquake. Amplification of earthquake motion
occurs at the ground surface where almost all structures are found, and therefore it is important to predict the
amplification. This chapter starts by explaining the mechanism of site amplification.

4.2.1 Types of waves

Figure 4.3 is a schematic showing paths of earthquake wave propagation from a fault to a site of interest
where the structure will be constructed. Two types of wave arrive at the site, as described below.

■ Body waves. There are two types of body wave: the P wave and the S wave. P waves arrive at the
site ﬁrst (i.e. primary), followed by S waves (i.e. secondary). The direction of propagation of P waves
is the same as the direction of vibration, and the mechanism is alternate compression and rarefaction
(tension), which also causes an instantaneous change in density. Therefore, P waves are often

Seismic Design of Foundations

termed ‘longitudinal’ or ‘compression’ waves. In contrast, the direction of wave propagation of the
later-arriving S waves is perpendicular to the direction of vibration, and as a result they are often
termed ‘transverse’ or ‘shear’ waves (i.e. causing shear deformation of the soil). Due to refraction
phenomenon, seismic wave propagation is vertical near a site, and as a result P waves cause vertical
vibration, while S waves cause horizontal vibration. Since the amplitude of S waves is usually much
larger than that of P waves and their predominant period is closer to that of structures, S waves are the
most dangerous from the viewpoint of earthquake-resistant design. When soil particles vibrate in the
plane of S-wave propagation, the S wave is called an SV wave; when the soil particles vibrate out of
plane, the S wave is called an SH wave.
■ Surface waves. Surface waves are generated by the interference of body waves at an irregular geometric
conﬁguration, such as the edge of a basin between the fault and the site, and travel horizontally along the
ground surface. The predominant period of a surface wave is long and its amplitude is small, and it is
therefore less important in earthquake-resistant design. However, displacement of surface waves can
become large, and can affect underground linear structures such as pipelines. Another characteristic of
surface waves is that their duration is long, so they may cause resonance of long-period structures such as
high-rise buildings.

4.2.2 Idealisation of the wave problem

Paths from the fault to the site of interest are usually divided into three parts, as follows.

■ The path from the fault to the seismic base layer where the ground motion attenuates depending on
the distance of travel. The seismic base layer is usually deﬁned where the shear wave velocity is about
3000 m/s and where non-linear behaviour is not expected.

Figure 4.3 Paths of earthquake waves from the fault to a site

Rock outcrop

Surface wave
Epicentre

Soil deposit
Body wave
Fault rupture

Engineering seismic
base layer

Seismic bedrock
Focus of earthquake

Geological strata

Seismic Design of Foundations

Ground response analysis

■ The path from the seismic base layer to the engineering seismic base layer. The definition of the
engineering base layer depends on the design specification, but usually varies between 300 and 700 m/s.
In Japan, 700 m/s is prescribed in the design specification of nuclear power plants, and 300 m/s appears
in the design specification of port facilities. However, 350–400 m/s is used for highway structures and
buildings in Japan. The shear wave velocity of 300–400 m/s corresponds to a standard penetration test
(SPT) N value of 50. The physical meaning of the engineering seismic base layer is difficult to explain
but is engrained in practice possibly due to practical reasons. For practical design, a borehole for site
investigation extends up to the foundation level. However, if the shear wave velocity of the engineering
seismic base layer is greater than the shear wave velocity of the foundation level, the borehole depth is
usually extended. In such cases, knowledge of the geology and geomorphology of the site can be helpful.
■ The path from the engineering seismic base layer to the ground surface that affects the built
environment. This part is the most complex, and in this phase the earthquake motion can either be
amplified or, in some cases, attenuated. Furthermore, non-linear behaviour including soil liquefaction
makes the overall behaviour even more complicated.

Table A.16 in Appendix A provides guidance on engineering base layer selection.

Readers are referred to the excellent book on ground response analysis by Yoshida (2015) for further reading.

4.3. Methods of ground response analysis

A number of techniques are available for ground response analysis. The methods mainly differ in the simplifying
assumptions that are made in the representation of the stress–strain relations of the soil and the methods adopted to
integrate the equations of motion. It may further be noted that stress–strain relations for soils are not simple due to
the fact that the problem is cyclic/dynamic and may involve the development of pore water pressure, which often
leads to a so-called ‘phase change’ in the soil. The current methods of dynamic response analysis have evolved
over the years due to the changing needs of practice, cheaper computing power and better understanding of soil
behaviour under cyclic loading. The methods of analysis can be broadly grouped into three categories

■ linear analysis
■ equivalent linear analysis
■ non-linear analysis.

Linear analysis
Owing to its simplicity, linear analysis is extensively used to study analytically the dynamic response of soil
deposits. Closed-form analytical solutions can be derived for idealised geometries and soil properties. Examples
include modelling a deposit by considering a uniform layer with the soil stiffness either constant or varying with
depth in a way that can be expressed by simple mathematical functions. However, soil does not behave elastically
and its material properties can change in space, and in such situations closed-form analytical solutions are not
possible, and numerical techniques such as ﬁnite-element or ﬁnite-difference methods are plausible options.

Equivalent linear analysis

This method is based on the lumped-mass model of soil deposits resting on a rigid base to which the seismic
motions are applied. This method can be generalised to incorporate wave propagation with an energy-
transmitting boundary. The seismic excitation can be applied at any level in this model.

Seismic Design of Foundations

Non-linear analysis
A non-linear analysis is usually performed by using a discrete model such as ﬁnite-element and lumped-mass
models where time domain step-by-step integration of the equations of motion are carried out. To have
meaningful results, the stress–strain characteristics of the soil must be realistically modelled, which calls for
an advanced constitutive relationship.

4.4. Linear ground response analysis procedure

The linear analysis procedure for ground response analysis is described in this section. This section also
provides the information of ground properties that are necessary.

4.4.1 Assumptions in the analysis

■ Soil and rock layers are considered linear as well as one dimensional; that is, the soil and bedrock
surface are assumed to extend inﬁnitely in the horizontal direction.
■ Since the wave propagation velocities at shallower depths are generally lower than in the materials
beneath them, inclined rays that strike horizontal layer boundaries are usually reﬂected to a more
vertical direction (i.e. Snell’s law). In other words, the waves deviate towards the normal. By the time
the rays reach the ground surface, multiple refractions have often bent them to a nearly vertical
direction. This analysis is based on the fact that the response of a soil deposit is predominantly caused
by SH waves propagating vertically from the underlying bed rock.

The shear wave velocity and damping of waves are denoted by cs and z, respectively, with the layer number
included as an index. The input bedrock displacement amplitude is v0 (the input motion may be harmonic),
and the target is to obtain the amplitude of the ground motion v1 at the surface layer. The ratio v1/v0 is the
ampliﬁcation function for a particular soil layer.

The origin of the local coordinate system with the z axis pointing downwards is located at the top of the layers.
For SH waves, the out-of-plane displacement with the amplitude v is speciﬁed by

v(z, x) = v(z) exp ( − ikx) (4:1)

where

v(z) = ASH expðiktz) + BSH expð− iktzÞ (4:2)

ASH and BSH are the amplitudes of the waves travelling in the negative and positive z directions, respectively,
where the wave number for excitation frequency w is

w
k= (4:3)
c
where

c*s
c= (4:4)
mx
The complex shear wave velocity for damping z is

c*s = cs (1 + ix) (4:5)

Seismic Design of Foundations

Ground response analysis

For the angle of incidence measured from the horizontal ySH

mx = cos ySH (4:6)

sffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
t= − 1 = tan ySH (4:7)
m2x

w
kt = sin ySH (4:8)
cs *
The dynamic stiffness matrix SLSH of a soil layer of depth d is given by
" #
L ktG* cos ktd −1
SSH = (4:9)
sin ktd −1 cos ktd

where the complex shear modulus G* is

G* = G(1 + 2xi) (4:10)
The equation of motion for soil layers is
( ) " #( )
Q1 ktG* cos ktd −1 v1
= (4:11)
Q2 sin ktd −1 cos ktd v2

where Q is the external load amplitudes and v is the displacement amplitude.

As the half-space is mostly used to represent the (bed)rock, a superscript R is introduced, and the dynamic-
stiffness matrix SRSH is given by

½SRSH = iktG* (4:12)

The equation of motion for rock, with the subscript 0 denoting the free surface of the half-space, is

Q0 = iktG*v0 (4:13)

Now, combining stiffness matrices for each individual layer we get the global stiffness matrix [K*], while the
equation of motion is
8 9 8 9 8 9
>
> Q1 > > >
> 0 >> >
> v1 >>
>
> >
> >
> >
> >
> >
>
>
> >
> >
> >
> >
> >
>
>
> Q >
> > 0 > > v >
< 2
= > < > = >
< 2 >
=
: = : = ½K* : (4:14)
>
> >
> > > > >
>
> > > > >
> >
> >
>
>
> : > >
>
>
>
> : >
>
>
>
>
> : >>
>
>
> >
> > > > >
: ; > : > ; >
: >
;
Qi+1 Q0 vi+1

Solving Equation 4.14 gives the displacement vector, and the ampliﬁcation caused by soil layers is obtained
by the ratio v1/v0. An algorithm for linear ground response analysis is presented in Figure 4.4, and it can be
easily programmed in software such as MATLAB.

Seismic Design of Foundations

Figure 4.4 Algorithm for linear site response analysis

START

Harmonic or
Harmonic transient? Transient

Enter displacement amplitude Enter the input

of harmonic excitation, acceleration time history
time interval and duration

Obtain velocity and

displacement time
Input rock details (damping, history
density and shear wave
velocity) and layer details FFT (Fast Fourier Transform)
(layer thickness, density,
damping shear wave Displacement in
velocity) frequency domain

Compute global stiffness

matrix for each frequency Input rock details (damping,
and load vector density and shear wave
velocity) and layer details
(layer thickness, density,
damping shear wave
Get displacement vector
velocity)
for each frequency

Compute global stiffness

Divide displacement at first matrix for each frequency
layer by input dispacement and load vector
amplitude (i.e. amplification
function)

Get displacement vector

for each frequency
Amplification function
for each frequency

Divide displacement at first

layer by input dispacement
amplitude (i.e. amplification
Plot of amplification
function)
function vs. frequency

Amplification function
for each frequency
END

Surface acceleration
with frequency

FFT

END Surface acceleration in

time domain

Seismic Design of Foundations

Ground response analysis

4.4.2 Dynamic soil properties

One of the most important aspects of ground response analysis is the dynamic soil properties. Most soils
under moderate cyclic loads yield progressively, compact, generate positive pore water pressure and soften.
Collapsible soils such as loose sands can liquefy in the absence of drainage. The frequency of cyclic loading
conditions is therefore important, since the pore pressure generated by one full load cycle may only partially
dissipate before the next load cycle. For clays, not only strength but also stiffness (G) degrades. The amount
of stiffness degradation is a function of the level of strain and the number of cycles of loading. Extensive
laboratory and ﬁeld studies have clariﬁed many aspects of the behaviour of clays under such loading. Stiffness
degradation has been correlated with the plasticity index (PI) of the soil. For a particular cyclic strain in the
soil, more-plastic clays exhibit less degradation. Figure 4.5 shows a graph based on 16 sets of published
experimental data from strain-controlled tests on clays having different PIs and overconsolidation ratios
(OCRs). In the graph, a PI of zero indicates sand and a PI of 200 represents highly plastic Mexico City clays,
which show little non-linearity up to 0.1% cyclic shear strain. Gmax is the secant shear modulus at small strain.
Figure 4.6 shows the effect of degradation on G/Gmax for different numbers of cycles (N = 1, 10, 100 and
1000) for normally consolidated clays (OCR of 1).

When analysing a soil system, three important features of soil behaviour must be considered

■ Soil is a non-linear material, in which the stiffness progressively decreases with increasing shear
stress, until at a sufﬁciently high shear stress level it deforms plastically.
■ Soil is an inelastic material. Therefore, when subject to cyclic loading, it exhibits hysteretic damping
(energy loss), which increases with increasing shear strain. Typical damping curves are shown in
Figure 4.7.
■ The soil properties, including strength, stiffness and pore water pressure, may be affected by repeated
cycles of load. This is particularly relevant to saturated sands and silts that may suffer a build-up in
pore water pressure with continued load application possibly resulting in liquefaction.

Figure 4.5 Relations between G/Gmax versus cyclic shear strain for normally consolidated and overconsolidated
clays (Vucetic and Dobry, 1991)

1.0

0.8

0.6
G/Gmax

OCR = 1–15
0.4
PI =
200
100
0.2 50
30
15
0
0
0.0001 0.001 0.01 0.1 1 10
Cyclic shear strain, γc: %

Seismic Design of Foundations

Figure 4.6 Effect of cyclic stiffness degradation on G/Gmax versus cyclic shear strain for soils having different PIs and
loaded to different numbers of cycles

1.0

0.8 N=1
N = 10
N = 100
N = 1000
0.6
G/Gmax

OCR = 1
0.4
PI =
N=1 200
N = 10 100
0.2
N = 100 50
N = 1000 30
15
0
0
0.0001 0.001 0.01 0.1 1 10
Cyclic shear strain, γc: %

Figure 4.7 Damping curves for real soils (north-west India from the foothills of the Himalayan seismic belt)

YF 45% Rd 50 kPa
YF 45% Rd 100 kPa Upper bound
YF 45% Rd 200 kPa
YF 45% Rd 300 kPa (this study)
YF 65% Rd 50 kPa
YF 65% Rd 100 kPa Upper bound
30 YF 65% Rd 200 kPa
YF 65% Rd 300 kPa (Seed and Idriss, 1970)
YF 85% Rd 50 kPa
YF 85% Rd 100 kPa
YF 85% Rd 200 kPa
YF 85% Rd 300 kPa
Lower bound (this study)
Upper bound (this study)
Damping: %

20 Upper bound (Seed and Idriss, 1970)

Lower bound (Seed and Idriss, 1970)
Upper bound (Darendeli, 2001)
Lower bound (Darendeli, 2001)

10
Upper band
(Darendeli, 2001)

0
1E-4 1E-3 0.01 0.1 1 10

Shear strain: %

Seismic Design of Foundations

Ground response analysis

Other important aspects of ground response analysis is the normalised shear modulus (G/Gmax) and damping
curves. The G/Gmax–g and D–g curves proposed by Seed and Idriss (1970) and Darendeli (2001) are
widely used in ground response studies and are incorporated in many software packages. Seed and Idriss
(1970) developed the modulus and damping curves by collating experimental data and fitting these with a
single parameter (reference shear strain) in the hyperbolic model, while Darendeli (2001) utilised two
parameters (reference shear strain and curvature coefficient) to properly fit the experimental data. Seed and
Idriss curves provided only the ranges (upper, lower and mean range) without distinguishing them for various
confining pressures. However, Darendeli (2001) provided curves for different confining pressures (25, 100,
400 and 1600 kPa) for soils of varying plasticity and OCRs. Figure 4.7 shows experimentally obtained
damping curves for real soils.

4.4.2.1 Mathematical forms

The non-linear behaviour of the shear modulus of soil along with the shear strain g was first studied by Hardin
and Drnevich (1972). They proposed the well-known hyperbolic model relating shear stress to Gmax and g,
which was further modified by Darendeli (2001). The modified hyperbolic model is very simple and based on
large-scale laboratory testing on the stress–strain behaviour of soil. Darendeli (2001) introduced a reference
strain into the model, which is basically the strain value at 50% degradation of the shear modulus. The
modified hyperbolic model forms the basis for the various state-of-the-art ground response analysis software
packages, and is expressed as

G 1
= (4:15)
Gmax 1 + (g =gref )a

where g is the shear strain, gref is the reference shear strain (strain at G/Gmax = 0.5) and a is the cur-
vature coefﬁcient. The gref and a values determine the shape of the G/Gmax–g curve.

Aside: the use of empirical backbone curves. The value of gref can be obtained using the resonant-
column test on soils. The test procedure is explained in the next section. In the absence of tests, this can
be determined empirically using an equation based on the uniformity coefﬁcient of soil (Cu) proposed by
Menq (2003) as
ng
sm0
gref = Ag (4:16)
Pa

where Ag = 0.12Cu–0.6, ng = 0.5Cu–0.15 and Pa is the atmospheric pressure. This equation is valid for
sandy and gravelly soils only.

Dammala et al. (2017) and Puri et al. (2018) carried out extensive resonant-column tests on different types of
real soils from the Himalayan seismic belt to obtain the modulus reduction curves see Figure 4.7. The
obtained results were compared with standard curves, and signiﬁcant deviations noted. It is therefore
important to obtain realistic modulus reduction curves, which can be obtained by carrying out Resonant
Column tests and these are advisable for use in Ground Response Analysis.

4.4.3 Resonant-column test

The resonant-column test is used to obtain the modulus reduction curves, and the experimental test set-up is
shown in Figure 4.8. The basic principle involved in this test is the theory of wave propagation in prismatic
rods where a cylindrical soil specimen is harmonically excited until it reaches the state of resonance (i.e. the

Seismic Design of Foundations

Figure 4.8 Experimental set-up for the resonant-column test

Cyclic torsional load

Top cap
(free end)

Specimen

All around pressure

Membrane

Base pedestal
(fixed end)

peak response). The frequency at which the peak response is achieved is noted, and is essentially the resonant
frequency from where the shear wave velocity (Vs) and corresponding shear modulus (G) of the sample are
determined using wave propagation theory. Once the resonant frequency is achieved at a particular input
voltage (i.e. a particular strain level in the resonating soil), the input current to the coils is halted, to allow a
free vibration test. The peak amplitude of each cycle is determined, and the corresponding damping ratio (D)
is evaluated based on the log-decrement method. The input voltage to the system is further increased, which
in turn increases the strain in the soil specimen, and the corresponding shear modulus and damping ratio are
determined. Readers are referred to specialist laboratory textbooks for further details on resonant-column
tests.

4.4.4 Example of linear ground response analysis

A generalised soil proﬁle and the soil parameters recommended by a geotechnical investigation are provided
in Table 4.1. The shear wave velocity for the soil layers has been determined from SPT N values based on the
correlation of the Japanese highway code, and is

ϑs = 80N 1=3 m=s for sands (4:17)

ϑs = 100N 1=3 m=s for clays (4:18)

Seismic Design of Foundations

Ground response analysis

Table 4.1 Generalised soil proﬁle with soil parameters

Strata No. Basic description Thickness: m SPT N value Unit weight, Shear wave velocity,
kN/m3 ϑs m/s

I Soft silty clay 2.0 3 19.1 144

II Soft clayey silt 5.0 2 18.2 126
IIIA Loose sandy silt 12.5 8 18.0 160
IIIB Medium-dense silty sand 3.0 22 19.0 224
IV Stiff clayey silt 2.5 8 18.4 200
V Medium-dense silty sand 3.0 28 19.0 243
VI Medium-stiff silty clay 15.5 11 18.0 222
VII Stiff clayey silt 9.5 32 18.5 317
VIII Very dense silty sand 12 >100 20.0 500

Layer VIII is assumed to be the engineering base layer for the site (see Figure 4.3). A spectrum-compatible
time history as discussed in Chapter 3 and shown in Figure 4.9 was adopted for the input bedrock motion.
Figure 4.10 shows the ampliﬁcation response obtained from the analysis.

The outputs of the ground response analysis are given below.

Figure 4.10 shows the amplification function computed for the site generated using algorithm presented in
Figure 4.4). The shape of the amplification function indicates that significant amplification will occur at
several natural frequencies and that at higher frequencies (> 6 Hz) the amplification will be suppressed.

Figure 4.9 Spectrum-compatible acceleration time history used in the study (1999 Chi-Chi earthquake)

0.30

0.20

0.10
Acceleration: g

0.00

−0.10

−0.20

−0.30
0 30 60 90 120 150
Time: s

Seismic Design of Foundations

Figure 4.10 Ampliﬁcation response obtained from the linear harmonic analysis

2.8
2.4
2.2
Amplification function

2.0
1.8

1.6

1.4
1.2

1.0

0 1 2 3 4 5 6 7 8 9 10
Frequency: Hz

4.4.5 Example of non-linear ground response analysis

This section presents an example of nonlinear ground response analysis carried out to investigate the effect of
soil liquefaction on the ground surface motion. The chosen site is the port area of Kumamoto city (Japan),
which is located on a reclaimed island in the Ariake Sea. Field investigations carried out in the area following
the 2016 Kumamoto earthquake reported widespread liquefaction (Bhattacharya et al. 2018). The recorded
ground motions are available from the KiK-net strong-motion seismograph network, which is equipped with
seismographs installed inside the borehole and on the ground surface. As the ground motion recorded inside
the borehole is representative of bedrock condition, the comparison between the two records allows the
evaluation of possible site ampliﬁcation effect. The maximum ground acceleration was recorded at the station
KMMH16 near Mashiki town, close to the southern end of the causative Futagawa fault and about 20 km
from the chosen site. The ampliﬁcation of the ground surface motion due to the presence of shallow soft
layers is visible from the comparison between acceleration time-histories recorded inside the borehole and
outside the borehole – see Figure 4.11.

The soil profile in the port area of Kumamoto City consists of loose-to-medium dense sand, with relative
density ranging from 45 to 75% in the first 12 m, reducing to less than 30% at depths below 15 m (see Figure
4.12a). As there were no recording stations at the site, the ground response needs to be estimated by means of
a site response analysis. As the soil is known to have liquefied, the equivalent linear static approach would not
be suitable due to the highly nonlinear behaviour expected from the liquefied soil, therefore a nonlinear
approach is used instead. In this example, the Open System for Earthquake Engineering Simulation software
framework – OpenSees (Mazzoni et al. 2006) is used to carry out a one-dimensional nonlinear response
analysis, adopting the ground motion recorded in the borehole at the KMMH16 station as input motion for the
analysis. The soil behaviour is modelled using the Manzari-Dafalias constitutive model (Dafalias and
Manzari, 2004); this enables the evaluation of the excess pore water pressure developed during the shaking.
The results from the ground response analysis are illustrated in Figure 4.12. From the normalised excess pore
water pressure profile, ru, it can be seen that liquefaction occurred at a depth between 10 to 20 m. The ground

Seismic Design of Foundations

Ground response analysis

motion in the liquefied layer shows a significant reduction in acceleration amplitudes, whereas the surface
ground motion is comparable with that recorded at the bedrock. This example shows that the occurrence of
liquefaction may de-amplify the ground motion in the liquefied layer. Although this effect may be seen as
beneficial, the occurrence of liquefaction is responsible for significant loss of strength and stiffness of the
foundation soil, which may have catastrophic consequences on the overall seismic performance of founda-
tions and structures (see Chapter 6).

Figure 4.11 Mainshock at the KMMH16 recording station. The black solid line denotes ground motion recorded
inside the borehole, which is representative of bedrock condition. The dotted line indicates the surface ground
motion. The higher acceleration ordinates recorded at ground surface are caused site ampliﬁcation effect due to the
presence of soft shallow deposits

1.2
Ground surface
1 Bedrock
0.8

0.6
Acceleration: g

0.4

0.2

–0.2

–0.4
EW component
–0.6
KMMH16 recording station
–0.0
0 10 20 30 40 50 60
Time: s

Figure 4.12 (a) soil and SPT-count proﬁle in the port area of Kumamoto City, and excess pore water pressure ratio,
ru, computed from site response analysis. (b) Comparison between recorded input ground motion at bedrock and
computed acceleration responses within liqueﬁed layer and at ground surface

SPT-count Relatively density [%] 0.25

0 10 25 30 60 Bedrock
0 Ground surface
W.T. 0.2
Within liquefied layer
5 0.15

10 0.1
Acceleration: g

Sand 0.05
15
Depth: m

0
20
–0.05
25 –0.1
Liquefaction
30 1.0 –0.15
ru 0.5 –0.2
35
–0.25
0.0 0 10 20 30 40 50 60
40
No liquefaction Time: s
(a) (b)

Seismic Design of Foundations

4.5. Equivalent linear and non-linear programs for ground response

analysis
Many programs are available for ground response analysis. A summary of some of these software packages is
provided in Table 4.2.

Details of the codes can be found using standard search engine such as Google.

4.5.1 Equivalent linear/non-linear ground response analysis procedure

The non-linear hysteretic stress–strain behaviour of cyclically loaded soils can be approximated by
equivalent linear soil properties. In this method, the equivalent linear shear modulus is generally taken as
a secant shear modulus, and the equivalent linear damping ratio as the damping ratio that produces
the same energy loss in a single cycle as the actual hysteresis loop. Since the linear approach requires the
shear modulus and damping to be constant for each layer, the problem becomes one of determining
the values that are consistent with the level of strain induced in each layer. The modulus reduction curve
and the damping ratio curves developed in laboratory tests using harmonic loading are characterised by
the peak shear strain amplitude.

The shear strain time history for real earthquake motion is highly irregular, and it is common to characterise
the transient record in terms of the effective shear strain, which is often taken as 65% of the peak shear strain.
Since the computed strain level depends on the values of the equivalent linear properties, an iterative pro-
cedure is adopted to ensure that the properties used in the analysis are compatible with the computed strain
levels in all layers. The equivalent linear approach to the one-dimensional ground response analysis of
layered sites has been coded into a widely used program called SHAKE (Schnabel et al., 1972). The linear
one-dimensional ground response analysis employed here is based on the assumption that the ground surface
and all material boundaries below the ground surface are horizontal and extend infinitely in all lateral
directions. Although these assumptions can never be strictly satisfied, they are satisfied sufficiently for
engineering purposes at many sites.

4.5.2 Case study: ground response analysis for Harayana, India

The northern Indian state of Haryana is in the Himalayan seismic belt and is experiencing rapid infrastructural
development. Haryana has a large network of roadways, railways and irrigation canals with a vast number of
bridges, and a nuclear power plant is being planned in Fatehabad District. The important structures in the state
are shown in Figure 4.13(a). The proximity of Haryana to the Himalayan thrust system and Delhi–Aravalli
fold belt makes it susceptible to both moderate and severe earthquakes, and the fault locations are shown in

Table 4.2 Example geotechnical software available for site response analysis

Dimensions Equivalent linear Nonlinear

1D Dyneq, Shake91, ShakeEdit, ProShake, AMPLE, DESRA, DMOD, FLIP, SUMDES, TESS,
Shake2000, EERA, SIREN CyberQuake, DeepSoil, NERA, FLAC, ShearBeam
2D/3D FLUSH, QUAD4/QUAD4M, TLUSH, DYNAFLOW, TARA-3, FLIP,VERSAT, DYSAC2,
QUAKE/W, SASSI2000 LIQCA, FLAC, PLAXIS, SWANDYNE

Seismic Design of Foundations

Ground response analysis

Figure 4.13(b). Structures such as nuclear power plants and bridges are lifeline structures for the state, and
damage or failure of these important facilities during an earthquake can further magnify the disaster. Hence,
safety against earthquake and earthquake-induced hazards is extremely important.

In order to demonstrate the application of the ground response analysis based on soil testing data, a non-linear
ground response analysis was undertaken for ﬁve sites along the Yamuna River in Haryana) (locations shown in
Figure 4.13a) using the DEEPSOIL program (version 6.1). Borehole data based on SPT data were obtained from
the recent geotechnical database developed for Haryana. All the sites were explored up to refusal, and these fall
under site class D, following the recommendations from the National Earthquake Hazards Reduction Program.

Figure 4.13 (a) Location of important structures, soil procurement sites and boreholes along the Yamuna River in
Haryana (YC, Yamuna coarse sand; YF, Yamuna ﬁne sand). (b) Major tectonic features in and around Haryana

31°
N Hi
ma
ch
al
E Pr
W ad
b es
nja h
Pu
S

30°

Site 1

Sites 2 and 3
Latitude (N)

Uttar Pradesh
State of Haryana
29°
Site 4
Yamuna River
lhi
Ra

De
ja
st
ha
n

Site 5

28° River
Sites
Nuclear power plant
Important bridges
Skyscraper
YC procurement site
YF procurement site

27°
74° 75° 76° 77° 78°
Longitude (E)
(a)

Seismic Design of Foundations

Figure 4.13 (Continued )

33°
1 2 N
17 18
Chamba 20
3 Sialkot 16
Kathua W E
Manali
32°
Gurdaspur
19 S
4 Chiniot Mandi
Amritsar
Lahore Hoshiarpur
Jalandhar
31°
Ferozpur 15
Chandigarh Uttarkashi
Ambala Tehri
21
Patiala Dehradun
Bathinda
30° 22
Saharanpur
23 24
14 Pithorgarh
Latitude (N)

13
Bijnor Nainital
29° Hisar
5
Rohtak
Bhiwani 11 Moradabad
12
Bulandshehar Bareily
28° Rewari
Bikaner Ratangarh 10
6 Aligarh
9
8
7
Mathura
Alwar
27° Agra Firozabad

Jodhpur
26°

25
Shiv Puri

25°
72° 73° 74° 75° 76° 77° 78° 79° 80° 81°
Longitude (E)
(b)

The bedrock was assumed at refusal (i.e. for N > 50 for 15 cm penetration or N > 100 for 30 cm penetration of the
SPT split-spoon sampler) and modelled as an elastic half-space with 2% damping, a density of 25 kN/m3 and a
shear wave velocity of 760 m/s. The soils from each of the geological proﬁles (YC and YF) were tested at the
SAGE (Surrey Advanced Geotechnical Engineering) laboratory. Figure 4.14 shows the modulus reduction curves
from the tests, and these data were used to carry out the ground response analysis.

Based on the peak ground acceleration (PGA) (g) values for rock sites obtained from the deterministic seismic
hazard analysis of Haryana, input motions for each site were selected. Suitable recorded acceleration time
histories of the Chamoli and Uttarkashi earthquakes recorded at rock sites were selected to carry out the
ground response analysis. Acceleration time histories used in the study (as shown in Figure 4.15) were
collected from the Consortium of Organizations for Strong-Motion Observation Systems (COSMOS). For
the ground response model, each layer must have its own modulus degradation and damping ratio curves

Seismic Design of Foundations

Ground response analysis

Figure 4.14 Shear modulus degradation and damping curves for ﬁne and coarse sands

1.0 35 1.0
32
30
0.8 28
0.8
25 24
0.6

Damping: %

Damping: %
20 0.6 20
Fitted G/Gmax curve Yamuna fine sand
G/Gmax

G/Gmax
Experimental G/Gmax values Yamuna fine sand
at 100 kPa 16
at 50 kPa 15
0.4 Fitted damping (%) curve 0.4
Experimental damping (%) 12
values 10
8
0.2 0.2
5 4

0 0 0 0
1E-4 1E-3 0.01 0.1 1 10 1E-4 1E-3 0.01 0.1 1 10
Strain: % Strain: %
1.0 32 1.0 30

28
25
0.8 0.8
24
20
20
Damping: %
0.6

Damping: %
0.6
G/Gmax

G/Gmax
Yamuna fine sand 16 15
at 200 kPa Yamuna fine sand
0.4 12 0.4 at 300 kPa
10
8
0.2 0.2
5
4

0 0 0 0
1E-4 1E-3 0.01 0.1 1 10 1E-4 1E-3 0.01 0.1 1 10
Strain: % Strain: %
1.0 1.0 28
28
24
0.8 24 0.8
20
20
Damping: %

Damping: %
0.6 0.6 16
16
G/Gmax

G/Gmax

Yamuna coarse sand Yamuna coarse sand

at 50 kPa at 100 kPa 12
0.4 12 0.4

8 8
0.2 0.2
4 4

0 0 0 0
1E-4 1E-3 0.01 0.1 1 10 1E-4 1E-3 0.01 0.1 1 10
Strain: % Strain: %
1.0 1.0
24
20

0.8 20 0.8

15
16
Damping: %

Damping: %

0.6 0.6
G/Gmax
G/Gmax

Yamuna coarse sand 12 Yamuna coarse sand

10
at 200 kPa at 300 kPa
0.4 0.4
8
5
0.2 0.2
4

0 0 0 0
1E-4 1E-3 0.01 0.1 1 10 1E-4 1E-3 0.01 0.1 1 10
Strain: % Strain: %

Seismic Design of Foundations

Figure 4.15 Selected input acceleration time histories

0.25 0.3
Chamoli earthquake longitudenal component Chamoli earthquake transverse component
0.20 PGA = 0.199g PGA = 0.36g
0.2
Predominant frequency = 1.53 Hz Predominant frequency = 1.67 Hz
0.15
0.1
Acceleration: g

Acceleration: g
0.10
0.0
0.05

0.00 −0.1

−0.05 −0.2

−0.10 −0.3
−0.15
−0.4
0 5 10 15 20 25 0 5 10 15 20 25
Period: s Period: s
0.3 0.4
Uttarkashi earthquake longitudenal component Uttarkashi earthquake transverse component
PGA = 0.242g 0.3 PGA = 0.310g
0.2
Predominant frequency = 4.08 Hz Predominant frequency = 2.91 Hz
0.2
Acceleration: g

Acceleration: g

0.1
0.1
0.0
0.0
−0.1
−0.1

−0.2 −0.2

−0.3 −0.3
0 10 20 30 40 0 10 20 30 40
Period: s Period: s

based on its mean effective confining pressure, and, as a result, tests were carried out for different con-
fining pressures (50, 100, 200 and 300 kPa). The results of the ground response analysis in terms of
amplifications in the PGA are reported in Figure 4.16, together with the variation of the PGA with depth
for all five sites.

The results for the sites were compared with the results obtained using the standardised modulus reduction
curves proposed by Seed and Idriss (1970) and Darendeli (2001). A difference of 11–34% was observed in the
amplification factors for these sites. The variation can be attributed to the over-prediction of strains at these
sites by the standard curves, as shown in Figure 4.17. The impact of using region-specific curves was further
investigated by plotting the response spectra for all the cases, as shown in Figure 4.18. It was observed that the
spectral acceleration followed the same trend as for the PGA. For all the sites, the Seed and Idriss (1970) and
Darendeli (2001) curves underestimated the response for almost all periods. According to the available
geotechnical database, the state of Haryana has a large area with a medium-dense soil profile, so the use of
standard curves will lead to underestimation of the earthquake hazard. Therefore, it is recommended that for
high-seismicity region such as Haryana the design of lifeline structures such as bridges, nuclear power plants
and pipelines must be based on site-specific, or at least region-specific, modulus degradation and damping
ratio curves.

Seismic Design of Foundations

Ground response analysis

Figure 4.16 Variation of the PGA with depth using different soil curves at different sites

PGA: g
0.18 0.24 0.30 0.2 0.4 0.6 0.2 0.4 0.6 0.3 0.4 0.5 0.6 0.2 0.3 0.4 0.5
0 0 0 0 0

5 5 5 5 5

10 10 10 10 10
Depth: m

15 15 15 15 15

20 20 20 20 20

25 25 25 25 25

Site 1 Site 2 Site 3 Site 4 Site 5

30 30 30 30 30
This study
Seed and Idriss (1970)
Darendeli (2001)

Figure 4.17 Variation of the maximum strain with depth using different soil curves at different sites

Shear strain: %
0 0.15 0.30 0 0.2 0.4 0.6 0 0.2 0.4 0.6 0 0.6 1.2 1.8 0 0.2 0.4 0.6
0 0 0 0 0
Site 1 Site 2 Site 3 Site 4 Site 5

5 5 5 5 5

10 10 10 10 10
Depth: m

15 15 15 15 15

20 20 20 20 20

25 25 25 25 25

30 30 30 30 30
This study
Seed and Idriss (1970)
Darendeli (2001)

Seismic Design of Foundations

100

Seismic Design of Foundations

Figure 4.18 Comparison of the surface acceleration spectra obtained using different soil curves with bedrock spectra at different sites

2.8
Top of rock Site 1 2.6 Top of rock Site 2 1.8 Top of rock Site 3
1.0
This study This study
Seismic Design of Foundations

This study
Seed and Idriss (1970)
2.4 Seed and Idriss (1970) Seed and Idriss (1970)
1.6
Darendeli (2001) 2.2 Darendeli (2001) Darendeli (2001)
0.8 1.4
2.0
1.8 1.2
0.6 1.6
1.4 1.0
0.4 1.2 0.8
1.0
0.8 0.6
0.2
0.6 0.4
0.4

Peak spectral acceleration: g

0.2

Peak spectral acceleration: g

Peak spectral acceleration: g
0 0.2
0 0
0.01 0.1 1 10 0.01 0.1 1 10 0.01 0.1 1 10
Period: s Period: s Period: s
2.5 1.8
Top of rock Site 4 Top of rock Site 5
This study
This study 1.6 Seed and Idriss (1970)
Seed and Idriss (1970)
2.0 Darendeli (2001)
Darendeli (2001)
1.4
1.2
1.5
1.0
0.8
1.0
0.6

0.5 0.4

0.2

Peak spectral acceleration: g

0 0
0.01 0.1 1 10 0.01 0.1 1 10
Period: s Period: s
Ground response analysis

4.6. Simpliﬁed method to obtain the natural period of the ground

The time period for ground of thickness h and for a shear wave velocity of vs is given by

4h
TG = (4:19)
vs

A commentary on this formula and a way to remember it follows.

From the deﬁnition of wave propagation in physics we know that velocity v is related to frequency f and
wavelength l by

v = fl (4:20)

The ground has a quarter-wavelength resonance. Therefore, one can easily see that l = 4h. So

v
v = f 4h or f = (4:21)
4h

The time period of the ground is thus

4h
TG = (4:22)
v

4.6.1 Case study: the 1985 Mexico City earthquake

It is reported that 6000 buildings were damaged due to the earthquake, and, later, it was found that the damage
was concentrated in a particular zone. Much of the city is built on the edge of a lake that has been dry since the
lake was drained. Sediments deposited by the lake produced a subsurface soil proﬁle.

For a site having a shear wave velocity vs of 75 m/s and a ground depth of 37 m, the time period ≈ 2 s.

4.7. Natural period of the ground for a layered soil

For a layered soil, Equation 4.19 can be extended as

X soil thickness
TG = 4 (4:23)
vs

such that

H1 H2
TG = 4 + + ::: (4:24)
vs1 vs2
where vs is the shear wave velocity of each layer and Hi is the thickness of each layer.

Standard correlations can be used to obtain the shear wave velocity: for example, vs = 100N1/3 for clay and
vs = 80N1/3 for sand where N is the SPT value.

The readers are referred to Appendix A for other correlations.

101

Seismic Design of Foundations

4.8. Summary
This chapter provides details of ground response analysis: the prediction of the behavior of the ground due to
the strong motion – wave propagation patterns from seismic source (that is, where the fault ruptured) to the
site (that is, the location we are interested in). To carry out the analysis, one needs ground conditions (i.e.
borehole data, geological and geomorphological information i.e. layering of the ground), soil properties and
the expected strong motion at the bedrock level or at the engineering base layer. One of the most important
ingredients of the analysis is modulus reduction curve, that is, the variation of Shear Modulus of the soil (G)
and damping for a range of strain. These types of analysis are usually carried out using software programs.
Examples of ground response analysis are presented in this chapter.

REFERENCES
Bhattacharya S, Hyodo M, Nikitas G et al. (2018) Geotechnical and infrastructural damage due to the 2016
Kumamoto earthquake sequence. Soil Dynamics and Earthquake Engineering 104: 390–394.
Dafalias YF, Manzari MT. Simple plasticity sand model accounting for fabric change effects. Journal of
Engineering Mechanics 2004.
Dammala PK, Krishna AM, Bhattacharya S, Nikitas G and Rouholamin M (2017) Dynamic soil properties
for seismic ground response studies in Northeastern India. Soil Dynamics and Earthquake Engineering
100: 357–370.
Darendeli MB (2001) Development of a New Family of Normalized Modulus Reduction and Material Damping
Curves. PhD thesis, University of Texas, Austin, TX, USA.
Hardin BO and Drnevich VP (1972) Shear modulus and damping in soils: design equations and curves. Journal of
the Soil Mechanics and Foundations Division, ASCE 98(7): 667–692.
Mazzoni S, McKenna F, Scott MH, Fenves GL (2006) OpenSees command language manual. Paciﬁc Earthquake
Engineering Research (PEER) Center, 264.
Menq FY (2003) Dynamic Properties of Sandy and Gravelly Soils. PhD thesis, University of Texas, Austin, TX,
USA.
Puri N, Jain A, Mohanty P and Bhattacharya S (2018) Earthquake response analysis of sites in state of Haryana
using DEEPSOIL software. Procedia Computer Science 125: 357–366.
Schnabel PB, Lysmer J and Seed HB (1972) SHAKE: A Computer Program for Earthquake Response Analysis of
Horizontally Layered Sites. Earthquake Engineering Research Center, University of California, Berkeley, CA,
USA, Report EERC 72-12.
Seed HB and Idriss IM (1970) Analyses of ground motions at Union Bay, Seattle during earthquakes and distant
nuclear blasts. Bulletin of the Seismological Society of America 60(1): 125–136.
Vucetic M and Dobry R (1991) Effect of soil plasticity on cyclic response. Journal of Geotechnical Engineering
117(1): 89–107.
Yoshida N (2015) Seismic Ground Response Analysis. Springer, Dordrecht, the Netherlands.

102

Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

ISBN 978-0-7277-6166-8
https://doi.org/10.1680/sdof.61668.103
ICE Publishing: All rights reserved

Chapter 5
Seismic analysis methods related to
foundation design
5.1. Introduction
Different structural concepts are necessary when designing foundations. The aim of this chapter is to provide
the essential structural engineering concepts that are necessary to carry out analysis of foundations. Readers
are referred to standard structural dynamic textbooks for further details: two classics are by Chopra (2013)
and Housner and Jennings (1982).

5.2. Fundamental period of structures

For any dynamic calculation, an important aspect is the determination of the time period of the structure (or its
natural frequency), in order to avoid resonance or to mitigate against resonance-induced effects. In the context
of earthquake engineering, resonance is the phenomenon that occurs when the natural frequency of the
applied loading comes close to the natural frequency of the structure, and the amplitude of forced vibrations of
the structure is magniﬁed. The magniﬁcation depends on the degree of energy dissipation from the system
(i.e. damping).

In practice and in simpliﬁed methods of calculations, the fundamental vibration period of a building is most
often estimated through empirical formulae based on the geometry and type of superstructure. For an ordinary
structure, the fundamental period of an n-storey building can be conveniently estimated by

T = 0:1n (5:1)

For a ﬁve-storey building, for example, the time period will therefore be 0.5 s.

Building codes provide alternative expressions for the fundamental period of structures. For example, in
Eurocode 8 (EC8; CEN, 2004), the period is expressed as a function of the height of the building H, which
must be measured from the foundation or from the top of a rigid basement, and a factor Ct that depends on the
type of structural frame

T = Ct H 3/ 4 (5:2)

where Ct is equal to 0.085 for moment-resistant steel space frames; 0.075 for moment-resistant concrete space
frames and for eccentrically braced steel frames; and 0.050 for all other structures. Figure 5.1 shows a
comparison between Equations 5.1 and 5.2.

103

Seismic Design of Foundations

Figure 5.1 Time period of a structure based on EC8 (Courtesy of N. A. Alexander)

1.4

1.2

1.0
T = 0.1n (n is number of storeys, storey
height = 3 m)
0.8
Period: s

Moment frames
(concrete)
0.6
Moment frames
(steel)
0.4
Other
structures
0.2

0
0 5 10 15 20 25 30 35
Height: m

A few points may be noted

■ The formula given in Equations 5.1 and 5.2 makes no distinction between types of foundation.
In other words, a five-storey building supported by a piled foundation or by a shallow foundation
will have the same time period. The flexibility provided by the foundation and any soil–structure
interaction effects are neglected. The rationale behind this is that flexibility considerations would lead
to an increase in the time period of the structure, and this is considered beneficial because seismic
forces would reduce.
■ It will be shown in Chapters 9 and 10 that with seismic subsurface liquefaction the fundamental
period of a pile-supported structure will increase drastically.

5.3. Fundamentals of modal analysis

Modal analysis is undertaken to compute the vibration characteristics of structures: vibration mode shapes
and the corresponding period of vibration. Figure 5.2 shows examples of modal analysis with the ﬁrst mode
of vibration only.

104

Seismic Design of Foundations

Seismic analysis methods related to foundation design

Figure 5.2 Examples of modal analysis for two buildings (Courtesy of N. A. Alexander)

The differential equation governing the free vibration motion of an undamped linear system with multiple
degrees of freedom (MDOF) is given by

M€
x + Kx = 0 (5:3)

where M is the diagonal mass matrix; K denotes the stiffness matrix; x is the displacement vector, 0 is a zero
vector; and the double dot indicates the second derivative with respect to time. Assuming a harmonic free-
vibration motion, the mode shape F is mathematically expressed by

xðt Þ = F sinðwt + q Þ (5:4)

where w is the circular frequency (rad/s) and q is the phase angle at time t.

Readers are referred to standard dynamics books for examples and further details.

A modal analysis is usually undertaken by solving the generalised eigenvalue problem that is mathematically
expressed by

[K − w 2 M]F = 0 (5:5)

The roots of this equation provide the fundamental frequency of the structure; that is, the inverse of the
fundamental period of the undamped system.

In routine practice, modal analysis is normally carried out with the aid of commercially available ﬁnite-
element software, although it can be done using available analytical solutions for structural models idealised
as single-degree-of-freedom systems. Figure 5.3 shows the ﬁrst three modes of vibration of a portal frame
structure where the mode shape as well as the frequencies are provided.

105

Seismic Design of Foundations

Figure 5.3 Three modes of vibration

1.5 0.72
2.78

2.17 0.5
0.92

0.1 0.1 0.1

First mode, frequency = 2.04 Hz Second mode, frequency = 5.04 Hz Third mode, frequency = 6.58 Hz

5.4. The concept of response spectrum

The response spectrum is the backbone of simpliﬁed seismic analysis and is a lynchpin for most (if not all)
codes of practice. This concept was ﬁrst proposed in 1932 by Biot in his PhD thesis as a practical method for
characterising ground motion and its effect on structures. However, Housner (1959) was instrumental for its
worldwide acceptance.

The response spectrum is the contour of peak responses of all possible single-degree-of-freedom (SDOF)
systems to a particular ground motion. The response can by represented in any form: acceleration, velocity or
displacement. However, most codes suggest the acceleration response spectrum, as designers need to con-
sider forces on the structure. For foundation design, the force is very important.

The response of an SDOF mechanical system can be fully described by a single coordinate (i.e. the response
of the top mass). The natural period T of vibration of such an SDOF system is given by
rffiffiffiffi
m
T = 2π (5:6)
k

where m is the mass of the system providing the inertial force and k is the stiffness of the system providing the
restoring force. Unlike idealised systems, real systems do not vibrate indeﬁnitely when excited, and the
motion decays due to energy dissipation through a phenomenon known as damping.

The concept of response spectra is illustrated in Figure 5.4, which shows a series of SDOF systems (deﬁned
by time periods T1 to T7) with a given level of structural damping subjected to an acceleration time history
acting at their base. The main assumption is that the ground behaves as a giant shaking table, and all seven
SDOF systems are being shaken. Each mass will respond differently based on its natural period, and the
frequency of the applied ground motion. The maximum absolute value of the response of each SDOF
oscillator can be calculated and plotted against the corresponding value of period, T. The resulting plot is
traditionally known as the response spectrum corresponding to a damping level. If instead of seven, hundreds
of SDOF systems are used, the piecewise linear line will become a smooth curve.

106

Seismic Design of Foundations

Seismic analysis methods related to foundation design

Figure 5.4 Concept of response spectra

Maximum

Maximum m

Maximum Maximum

Maximum
Maximum
Maximum

T1 T2 T3 T4 T5 T6 T7
Maximum amplitude

m
T = 2p
k

T1 T2 T3 T4 T5 T6 T7
Time period: s

Seed et al. (1976) found that the near-surface geology under the structure can have a signiﬁcant effect on the
accelerations at the ground surface and thereby the response spectrum would look different for different
ground type. In other words, if the same experiment shown in Figure 5.4 is conducted in different types of
ground, the response curve would look different. Readers are referred to Chapter 4 for details. Figure 5.5
shows the response spectra for a magnitude 7.5 earthquake at a distance of 35 miles (55 km) or magnitude 6.5
at 20 miles (30 km) where the accelerograms of similar ‘soil class’ are grouped (e.g. rock site, deep cohe-
sionless soils, soft-to-medium clay and sand).

5.4.1 Response spectra in codes of practice

EC8 (CEN, 2004) contains many elastic acceleration spectra, and there are two groups (Figure 5.6)
■ Horizontal spectra:
– type I (for large events >5.5 ML (local magnitude) and unknown deep geology)
– type II (for smaller events <5.5 ML and known deep geology).
■ Vertical spectra:
– type I (for large events >5.5 ML and unknown deep geology)
– type II (for smaller events <5.5 ML and known deep geology).

Soil data-classiﬁcation – Table A.8 (Appendix).

107

Seismic Design of Foundations

Figure 5.5 Effect of ground type on the response spectrum

0.7

0.6

Spectra for 5% damping

0.5
Total acceleration: g

0.4 Soft to medium clay and

sand

0.3 Deep cohesionless

soils

0.2
Stiff sites
0.1 Rock

0
0 0.5 1.0 1.5 2.0 2.5
Period, T: s

Figure 5.6 Generalised form for EC8 spectra

Se/ag

2.5 Sη

TB TC TD

108

Seismic Design of Foundations

Seismic analysis methods related to foundation design

The equations for the response spectra are

T
0 ≤ T ≤ TB : Se (T) = ag S 1 + (2:5h − 1) (5:7)
TB

TB ≤ T ≤ TC : Se (T) = ag S 2:5h (5:8)

Tc
TC ≤ T ≤ TD : Se (T) = ag S 2:5h (5:9)
T

TT
TD ≤ T ≤ 4 s : Se (T) = ag S 2:5h c 2D (5:10)
T

where Se(T) is the elastic spectral acceleration; T is the vibration period of a linear SDOF system; ag is the
design ground acceleration on type A ground (i.e. (ag = g1agR, where agR is the reference acceleration)); TB is
the lower limit of the period of the constant spectral acceleration range; TC is the upper limit of the period of
the constant spectral acceleration range; TD is the period at the beginning of the constant spectral displace-
ment range; S is the soil factor; and h = √(10/(5 + x) ≥ 0.55 is the correction factor for the damping ratio x,
other than for the reference case of 5% viscous damping.

Table 5.1 lists recommended values for the parameters from EC8.

Figure 5.7 shows the acceleration spectra used for the seismic design of structures in various countries: Japan
(‘Speciﬁcations for highway bridges’; JRA, 2002), Europe (EC8; CEN, 2004) and India (IS 1893; BIS,
2002). The response shape and magnitude of a spectrum depends on many factors, such as the soil type,
structure type, damping ratio and the ground motion itself. For example, the Indian standard (IS 1893) for the
earthquake-resistant design of structures suggests different spectra for structures on rock, medium soil and
soft soil sites with different damping ratios. EC8 and the International Building Code (ICC, 2000) also
suggest different spectra for soils, with ﬁve characterisations depending on the stiffness of the soil. In
addition, EC8 and the Japanese code (JRA, 2002) advise design engineers to consider earthquake magnitude
when choosing spectra: for example, level I and II earthquakes in the JRA code, and ground motion types I
and II in EC8.

Table 5.1 EC8 spectral response parameters

Ground type Spectrum type I Spectrum type 2

S TB: s TC: s TD: s S T B: s TC: s TD: s

A 1 0.15 0.4 2.0 1.00 0.05 0.25 1.2

B 1.2 0.15 0.5 2.0 1.35 0.05 0.25 1.2
C 1.15 0.2 0.6 2.0 1.50 0.10 0.25 1.2
D 1.35 0.2 0.8 2.0 1.80 0.10 0.30 1.2
E 1.4 0.15 0.5 2.0 1.60 0.05 0.25 1.2

109

Seismic Design of Foundations

Figure 5.7 Response spectra used in codes of different countries

2.5

Accelereation reponse spectrum 2.0 Ground type A

(motion I, EC8)
Ground type A (motion II, EC8)
1.5

1.0 Stiff soil (GM II, JRA code)

Rock type
5% Damping (IS 1893)

0.5
Stiff soil (GM I, JRA code)

0.1 1 10
Time period: s

Aside. Soil–structure interaction (SSI) effects play an important role in the seismic response. Simplistically,
SSI can be ignored if it can be demonstrated that there will be no ampliﬁcation of the seismic motion by the
soil–structure system: for example, if the soil–structure system is seismically rigid (i.e. the natural frequencies
of the system is above 33 Hz – this is a rule of thumb). In this case, the seismic input loading can be applied
directly to the base of the structure. This is referred to as ﬁxed-base analysis.

5.5. Application of response spectra analysis in foundation design

5.5.1 Introduction
This book concerns foundation design, and discussion therefore focuses around using this method for the
estimation of loads on a foundation. As will be shown, the use of response spectra for design is very
straightforward. The ﬁrst step is to ﬁnd an analogy of the given structure to a suitable SDOF system, and the
fundamental mode of the structure evaluated (see Section 5.2). Using the response spectra, the acceleration of
the equivalent SDOF system can estimated from the ordinate. The acceleration can be multiplied by the
structure mass (modal mass) to obtain the seismic forces.

The response spectrum method is a very simplified approach, and is very easy to carry out – it can even be
done on the back of an envelope! The method is based on the results of a modal analysis of a structure, and
therefore requires a modal analysis as the first step. Modal analysis is an eigenvalue procedure and therefore,
by definition, linear. The natural modes of vibration of the system are effectively decoupled in the modal
analysis process (e.g. see Figure 5.2), and the response of the structure for each of these modes is evaluated
independently. Often, input motions are defined separately in the three orthogonal directions, and the
response of the structure to each of these motions can consequently be determined independently.

In this type of analysis the seismic loading is deﬁned in the response spectra for three orthogonal directions.
The analysis in the frequency domain is calculated in terms of the peak response, and therefore the notion of
time is lost. In other words, the time at which the peak response occurs is unknown.

110

Seismic Design of Foundations

Seismic analysis methods related to foundation design

Response spectrum analysis is economical to carry out, as most of the computational effort is in the modal
analysis step. The method is especially useful for foundation design, as peak seismic loads are required for
ultimate limit state calculations (e.g. minimum size of the foundations).

The analysis steps are as follows

(a) Engineers must decide which and how many modes of vibration of a given structure are important for
the structure. This process is carried out by considering the natural frequencies and the mode shapes
and the effective masses participating in the modes.
(b) For a given mode of vibration and under a speciﬁed excitation spectrum, the peak dynamic response
of the structure, such as acceleration, is calculated from the mode shape and the mass participation
factor for that mode. The spectral acceleration can then be deﬁned by the response spectrum at the
natural frequency of the mode.
(c) The response of the structure to all possible modes of interest when subject to excitation spectra can
be calculated using the appropriate modal combination criteria.

5.5.2 Modal combination methods

There are a number of methods of carrying out the modal combination process. Some popular simpliﬁed
ones are

(a) Absolute method: the responses are summed. This approach tends to be very conservative.
(b) Square root of the sum of the squares (SRSS) method: a reasonable method if the modes are evenly
spaced. If modes are closely spaced, this method may not be conservative.
(c) Complete quadratic combination (CQC) method: this combines modes using cross-correlation
coefﬁcients depending on frequency ratios and modal damping values. It provides relative weightage.

5.5.3 Limitations of the response spectrum method

■ The spectrum shows the response of an SDOF system connected to the ground, not the actual motion
of the ground.
■ The response spectrum only shows peak response, so precise timing cannot be determined.
■ Strictly, response spectra apply only to linear systems. Any change in the frequency of the system due
to non-linearity of stiffness or damping or mass change would alter the response. For pile-supported
structures in liquefied soil, the frequency of the system decreases and damping of the system increases.
This is discussed in detail in Chapters 9 and 10.
■ As the frequency of the vibrating system increases, the spectral values for all damping levels converge
to zero-period acceleration (ZPA). In other words, an SDOF system has an infinite natural frequency
and behaves rigidly (i.e. infinite stiffness). As the ZPA is the peak response of a rigid SDOF system,
we may also infer that this is also the peak of the input acceleration.
■ The response spectra can be defined in terms of either the frequency (log scale) or the time period.

5.5.4 Effect of soil liquefaction on modal parameters

The combined effect of lengthening of the period and increase in the damping ratio caused by sub-surface
liquefaction has important repercussions on the design of a structure. This is shown schematically in
Figure 5.8. The spectral acceleration that the structure would experience during soil liquefaction could be
signiﬁcantly lower than that experienced prior to the onset of liquefaction. This suggests that in a con-
ventional force-based design approach, the reduction in acceleration demand would reduce the seismic

111

Seismic Design of Foundations

Figure 5.8 Effects of liquefaction on the acceleration response spectrum

Before liquefaction
After liquefaction

Tpre
Lfix-pre Lfix-pre < Lfix-post Tpost
Tpre < Tpost
Water
Dpre < Dpost Lfix-post Water

Klat Krot Liquefied

soil
Soil body

Klat Krot
Acceleration response spectra

Dpre

Dpost Before liquefaction

After liquefaction

Z1 Z2 Z3

Tpre Tpost Fundamental period

Z1 = Acceleration-sensitive zone
Z2 = Velocity-sensitive zone
Z3 = Displacement-sensitive zone

demand imposed on the structure. However, there will be a considerable increase in the displacement
demand. This effect is very detrimental for pile-supported bridges in liqueﬁable soils, and is discussed in
detail in Chapter 10.

5.5.5 Case study: recorded acceleration in Kawagishi-cho apartment complex

during the 1964 Niigata earthquake
Figure 5.9 shows plots of the time histories of the acceleration, velocity and displacement recorded for the
basement level of Kawagishi-cho apartment complex located 1.25 km from Showa Bridge. It is important
to note that the records from Kawagishi-cho are the only available strong motion records recovered near the
city of Niigata, and the site was fully liqueﬁed. The acceleration time history was recorded, but the velocity
and displacement time histories were obtained by numerical integration of the signal. Also, the acceleration
was recorded in two directions – east–west and north–south. The acceleration, velocity and displacement

112

Seismic Design of Foundations

Seismic analysis methods related to foundation design

Figure 5.9 Time history of acceleration (recorded) and the corresponding velocity and displacement histories

North–south component East–west component

1.5 1.5
Time frame of Time frame of
1.0 Showa Bridge collapse 1.0 Showa Bridge collapse
Acceleration: m/s2

Acceleration: m/s2
0.5 0.5
0.0 0.0
–0.5 –0.5
–1.0 –1.0
–1.5 –1.5

0.6 0.6
0.4 0.4
Velocity: m/s

Velocity: m/s
0.2 0.2
0.0 0.0
–0.2 –0.2
–0.4 –0.4
–0.6 –0.6

0.6 0.6
0.4 0.4
Displacement: m

Displacement: m

0.2 0.2

0.0 0.0

–0.2 –0.2
–0.4 –0.4

–0.6 –0.6
0 25 50 75 100 125 0 25 50 75 100 125
Time: s Time: s

spectra for this strong motion are shown in Figure 5.10 for damping ratios of 5% and 20%. Acceleration–
displacement response spectra are often useful for comparing the spectral behaviour, and provide the
response of structures for different time periods. Structures with periods of 2 and 6 s are shown in Figure 5.11;
the signiﬁcance of the numbers are discussed below. Readers are referred to standard earthquake dynamics
textbooks for details on the construction of the spectra.

The 1964 Niigata earthquake was a watershed event for earthquake geotechnical engineering, as many
lessons were learnt. Figure 5.9 also shows the window when the Showa Bridge (see Chapter 10) collapsed:
readers are referred to the detailed ﬁeld report collated by Yoshida et al. (2007) and discussed by
Bhattacharya et al. (2014). It is important to note that the time period of the bridge increased from 2 to 6 s with
soil liquefaction. Based on the latest research carried out by Lombardi and Bhattacharya (2014), the damping
of liqueﬁed soil can be as high as 20%. It is observed from Figure 5.11 that the spectral displacement reached
its peak at the period range 6–7 s. While the acceleration drops, the displacement demand increases.

113

Seismic Design of Foundations

Figure 5.10 Acceleration, velocity and displacement spectra for the time histories in Figure 5.9 for damping ratios of 5% and 20%

114
Damping, ζ = 5% Damping, ζ = 5% Damping, ζ = 5%
1000 1000 1000
North–south component North–south component North–south component

Seismic Design of Foundations

East–west component East–west component
East–west component

100
Seismic Design of Foundations

100 100

Velocity: cm/s
10

Displacement: cm

Acceleration: cm/s2
10 10 1
0.1 1 10 0.1 1 10 0.1 1 10
Period: s Period: s Period: s

Damping, ζ = 20% Damping, ζ = 20% Damping, ζ = 20%

1000 1000 1000
North–south component North–south component North–south component
East–west component East–west component East–west component

100

100 100

Velocity: cm/s
10
Displacement: cm

Acceleration: cm/s2
10 10 1
0.1 1 10 0.1 1 10 0.1 1 10
Period: s Period: s Period: s
Seismic analysis methods related to foundation design

Figure 5.11 Acceleration–displacement response spectra for the time histories in Figure 5.9 for a damping
ratio of 20%

250
North–south component
East–west component
200

Spectral acceleration: m/s2

T=2s

150

T=6s
100

0
0 30 60 90 120
Spectral displacement: cm

5.6. Example calculations for design seismic base shear of a three-storey

building
This example shows how to compute the design seismic base shear for a concrete three-storey building
following the rules for the representation of seismic actions by Eurocode 8. The building has a total mass of
1,300 tonnes, and a height of 10.5m -measured from the foundation level. The soil proﬁle at the site can be
categorised as ground type E. The corresponding spectral parameters are: soil factor S = 1.4, and period limits
TB = 0.15, TC = 0.5, and TD = 2.0; and correction factor l = 0.85. The structural type of the building is a
moment resistant concrete frame, with a behaviour factor q = 3.6 The following calculations are based on
Type 1 spectrum, which is representative of high seismicity areas of southern Europe, and reference peak
ground acceleration of 0.3g.

The ﬁrst step of the analysis consists in determining the fundamental period of the building. Although a
modal analysis would provide the vibration characteristics of the building, in this example, the structural
fundamental period T is computed through the equation proposed by Eurocode 8 for buildings with heights
less than 40m, ie

T = Ct H 3/ 4 = 0:075 10:53/ 4 = 0:44s (5:11)

where Ct is taken as 0.075 for the considered structural type, H is the height in m, and T is the period in s.

As the fundamental period of the building satisﬁes the condition TB ≤ T ≤ TC, the design spectral acceleration
is computed as

Sd = ag S2:5 / q = 0:3 1:4 2:5 / 3:6 = 2:29g = 2:86 m / s2 (5:12)

115

Seismic Design of Foundations

The design seismic base shear is ﬁnally given by

Fb = lmSd = 0:85 1300 2:86 = 3,160 kN (5:13)

5.6.1 Equivalent linear static analysis of an MDOF (Multi Degree of Freedom)

system
For a MDOF system the earthquake excitation factor is
X
Li = mj fij (5:14)
j

and the modal mass is

X 2
Mi = mj fij (5:15)
j

where i is the mode in question and j relates to degree of freedom (DOF) in the structure. fij is the value of the
mode shape i at the DOF j, and mj is the mass at DOF j.

The base shear in mode i is given by

L2i
fbi (max) = S (5:16)
Mi ai

where Li2/Mi is the participating mass in mode i and Sai is the spectral acceleration for mode i.

The storey force Fk is a function of the base shear Fb as follows

zm
Fk = Fb Xk k (5:17)
zj m j
j

where zk is the storey height, mk is storey mass for storey k and j is the total number of storeys.

For our example three-storey building (see Figure 5.12)

m 1 z1
F2 = Fb X
n
m i zi
i=1

322 (400/g) 4 (5:18)

=
(400/g) 4 + (400/g) 7 + (500/g) 10:5
1600
= 322 = 53:4 kN
9650

116

Seismic Design of Foundations

Seismic analysis methods related to foundation design

Figure 5.12 Example three-storey building

m3 = 500 t F3 = 1,719 kN

m2 = 400 t F2 = 917 kN

10.5 m

m1 = 400 t 7m F1 = 393 kN

Fb = 3,160 kN

m 1 z1
F2 = Fb X
n
m i zi
i=1

322 (400/g) 7 (5:19)

=
(400/g) 4 + (400/g) 7 + (500/g) 10:5
2800
= 322 = 93:4 kN
9650

m 1 z1
F3 = Fb X
n
m i zi
i=1

322 (400/g) 10:5 (5:20)

=
(400/g) 4 + (400/g) 7 + (500/g) 10:5
5250
= 322 = 175:2 kN
9650

117

Seismic Design of Foundations

5.6.2 Example calculations for design seismic base shear for a pile-supported
structure
This example shows how to compute the design seismic base shear for the building shown in Figure 5.13
following the rules by Eurocode 8. The building has a total weight of 10,749 kN and fundamental period of
0.6s. The soil proﬁle at the site can be categorised as ground type E. The corresponding spectral parameters
are: soil factor S = 1.4, and period limits TB = 0.15, TC = 0.5, and TD = 2.0; and correction factor l = 0.85. The
structural type of the building is a moment resistant concrete frame, with a behaviour factor q = 3.6 and a
damping correction factor h = 0.91. The following calculations are based on Type 1 spectrum, which is
representative of high seismicity areas of southern Europe, and reference peak ground acceleration of 0.24g.

As the fundamental period of the building is greater than TC and less than TD, the design spectral acceleration
is computed as

Sd = ag S2:5hTc / T / q = 0:24 1:4 2:5 0:91 0:5 / 0:6 / 3:6 = 0:18g = 1:74 m / s2 (5:21)

The design seismic base shear is ﬁnally given by

Fb = lmSd = 0:85 10749 / 9:81 1:74 = 1621 kN (5:22)

Figure 5.13 Structure on pile foundations

118

Seismic Design of Foundations

Seismic analysis methods related to foundation design

5.7. Essentials of performance-based design

The goal of earthquake-resistant design is to ensure that the seismic capacity of an engineered structure is
greater than the seismic demand. The seismic demand can be evaluated using different approaches, which can
be broadly classiﬁed into three categories

■ The intensity-based approach, in which the demand is expressed by an earthquake shaking intensity
defined by a single reference spectrum, usually provided by the relevant building code. The ground-
shaking intensity defining the seismic demand can be expressed through spectral accelerations of a
reference spectrum or a suite of accelerograms, which have been selected and scaled for consistency
with the reference spectrum.
■ Scenario-based assessment, which determines the structural performance for a single earthquake scenario
defined by the pair of magnitude and site-to-source distance. In this approach the ground-shaking
intensity is defined by means of the median acceleration response spectrum obtained from attenuation
relationships. Analyses are performed using spectral acceleration derived from the median spectrum,
or by accelerograms that are scaled to be consistent with the distribution of the median spectrum.
■ Time-based assessment, wherein the ground-shaking hazard is determined through probabilistic
seismic hazard analysis that considers all possible earthquake scenarios in terms of location and
magnitude, and the resulting levels of ground motion that each scenario may produce at the site.

5.7.1 Design considerations

The seismic performance of a structure can be fully identified in terms of four quantities: stiffness, strength,
displacement and ductility. In routine practice, two of these quantities are taken as the input parameters of the
design process, and the remaining two are obtained as an outcome. For the design of ordinary structures, the
current seismic design provisions are based on the assumption that the action exerted by an earthquake can be
represented by an equivalent static force that is proportional to the mass and the spectral acceleration at the
fundamental period of the structure. In such an approach, which is normally referred to as force-based design,
the input parameters are stiffness and strength, while displacement and ductility are obtained as a result of the
design process. The spectral accelerations required for the definition of the seismic action are estimated from
normalised spectra prescribed by the national annex of building codes based on the fundamental period and
damping ratio of the structure. It is worth noting that the period and damping of a structure are computed by
ignoring the contribution of the foundation flexibility and radiation damping due to wave propagation
through the soil because their effects on the vibration characteristics are normally considered to be beneficial.
Force-based design approaches are widely used in routine practice. However, in the presence of flexible
structures, such as bridges and tall buildings, or highly non-linear phenomena, such as soil liquefaction, it is
questionable whether the performance should be assessed in terms of the strength of structural elements, as in
a conventional force-based design.

Over recent decades there has been increasing emphasis on the development of performance-based seismic
design approaches, with the goal to design and/or assess the performance of a an existing or newly built
structure for a given target performance, which is normally speciﬁed by the relevant stakeholders. In fact,
since deformation is more fundamental to damage control than strength, there has been a tendency to adopt
displacement-based design approaches wherein displacement and ductility of the structure are the input
parameters of the design process, with strength and stiffness being the output of the design.

119

Seismic Design of Foundations

Figure 5.14 Deﬁning structural performance through a global capacity curve

Structural
Non-structural
damage Collapse
damage
Elastic

Base shear force

limit

Top displacement

A convenient way to express the performance of a structure is given by the capacity curve; this can be
visualised in a base shear force versus top displacement plot (Figure 5.14), where the performance depends on
the level of displacement of the structure. For a level of deformation within the elastic range, the structure is
expected to be fully operational with negligible damage during an earthquake. As the displacement increases,
the structure starts exhibiting minor disruption to non-essential services until damage becomes severe and life
safety is at risk, eventually leading to structural collapse at large displacement levels.

5.7.2 Representation of seismic actions through response spectra

For ordinary structures, the seismic demand is obtained from code-based design response spectra, such as
those provided in the national annex of building codes. In a force-based design approach, the response
spectrum is presented in terms of acceleration spectral ordinates, reduced by a factor that accounts for the
ductility of the structure, versus the fundamental period of an equivalent SDOF system. The code-based
spectra are anchored to the value of peak ground acceleration deﬁned by a national annex through seismic
zonation maps for different return periods.

5.7.3 Code-based response spectra versus reality

An example zonation map developed for Italy for a return period of 475 years is shown in Figure 5.15.
Acceleration response spectra of earthquakes may be highly irregular owing to the relatively wide frequency
range for strong ground motions, typically in the range 0.5–15 Hz. However, code spectra used for design
practice have more regular shapes, obtained by smoothing – averaging multiple response spectra from real
earthquakes. A comparison between code response spectra and spectra from real accelerograms recorded
during the 2012 Northern Italy earthquake sequence is provided in Figure 5.16.

For a given seismic zone, the spectral ordinates are a function of (a) surface geology, whereby softer deposits
tend to amplify the spectral accelerations, and (b) the type of structure, whereby spectral ordinates are
multiplied by the importance factor of the structure, taken as unity for ordinary buildings, which effectively
changes the return period of the spectral ordinates (Figure 5.17).

Building codes also provide spectral shapes for the vertical component of ground shaking. It is worth
mentioning, however, that often such spectra tend to underestimate the actual vertical acceleration ordinates,
especially when a site is located in the proximity of the seismic source, so they should be used with caution for
design purposes (Figure 5.18).

120

Seismic Design of Foundations

Seismic analysis methods related to foundation design

Figure 5.15 Italian seismic hazard map, in terms of peak ground acceleration (PGA) referred to horizontal ground
and bedrock with a return period of 475 years. (Italian Seismic National Annex Ordinanza PCM del 28 Aprile 2006
n.3519, All.1b)

PGA
< 0.025 g
0.025–0.050
0.050–0.075
0.075–0.100
0.100–0.125
0.125–0.150

0.150–0.175
0.175–0.200
0.200–0.225
0.225–0.250

0.250–0.275
0.275–0.300

0 50 100 150 km

Figure 5.16 Comparison of elastic horizontal acceleration response spectra with Italian code spectra for two events
recorded during the 2012 Northern Italy earthquake sequence

0.7 29 May 2012 event

20 May 2012 event
0.8 Italian code NTC. 2008 Italian code NTC. 2008
Return period 475 years Return period 475 years
0.7 0.6
Spectral acceleration Sa: g

Ground type C Ground type C

Spectral acceleration Sa: g

Ground type D Ground type D

0.6 0.5
North–south component North-south component
East–west component East-west component
0.5 0.4
0.4
0.3
0.3
0.2
0.2
0.1 0.1

0 0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Period: s Period: s

121

Seismic Design of Foundations

Figure 5.17 Importance factor vs return period where k denoted the seismicity exponent

2.50

2.00
Importance factor

1.50

k = 2.5
1.00
k = 3 (EC8)
k=4
0.50

0
0 250 500 750 1000 1250 1500 1750 2000
Return period

Figure 5.18 Comparison of elastic vertical acceleration response spectra with Italian code spectrum

Italian code NTC, 2008

2.5 Return period 475 years
Ground type C
Ground type D
Spectral acceleration Sa: g

2.0 20 May 2012

29 May 2012

1.5

1.0

0.5

0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Period: s

122

Seismic Design of Foundations

Seismic analysis methods related to foundation design

5.7.4 Uniform hazard spectra

For critical infrastructure such as nuclear power plants, dams and so on, the seismic demand is obtained
through probabilistic risk hazard analysis, with the output expressed in the form of a family of seismic hazard
curves, which give the probability of exceeding various ground motion values for different return periods.
The seismic hazard curves are subsequently used to construct uniform hazard spectra (UHSs), which,
differing from code-based spectra, have spectral ordinates with the same return period. It is worth noting that
the use of UHSs has several advantages over traditional response spectra, since their spectral ordinates are no
longer anchored to peak ground acceleration, and the probability of exceeding the spectral quantity is the
same for any fundamental period.

5.7.4.1 Case study: earthquake hazard for a hypothetical nuclear power plant, UK
A hypothetical nuclear power plant was assumed to be located on a rock site at Holyhead site in north Wales
(UK). The chosen site is in the north-western part of the European Stable Continental Region, as defined
in the Flinn–Engdhal scheme. An intensity-based approach was used in which the seismic hazard is
defined through UHSs for an annual frequency of exceedance of 10–4. The use of UHSs presents several
advantages over traditional response spectra such as PML spectra (PML, 1981) and Eurocode spectra (CEN,
2004), since their spectral ordinates are no longer anchored to peak ground acceleration, and the probability
of exceeding the spectral quantity is the same for any fundamental period. Following the ASCE 43-05
standard (ASCE, 2005), the design response spectrum was obtained by multiplying the UHS by a design
factor DF, defined at each spectral frequency as

SA0:1HD
DF = maximum 0:8, 0:6A0:8
R AR = (5:23)
SAHD

where AR is the slope factor, SAHD is the spectral acceleration SA at the annual frequency of exceedance
(i.e. HD = 10–4) and SA0:1HD is the SA at 0.1HD.

Selection and scaling of accelerograms for non-linear response history analyses. Eleven
accelerograms from the European Stable Continental Region were selected for the time response analysis.
Arguably, the use of accelerograms from active crustal zones may be overly conservative for the seismic
design of nuclear facilities in the UK. On the other hand, the use of accelerograms from small-magnitude
earthquakes recorded in the UK is not recommended owing to the paucity of recorded accelerograms and
the lack of strong-ground motion data in the region. It was assumed that the features of accelerograms caused
by intra-plate earthquakes in the UK are similar to those recorded in the north-western European Stable
Continental Region. The selection criteria for the accelerograms to be used in time-history analysis are
not trivial because of the inherent uncertainties in the source, path and site characteristics. The suite of
accelerograms was selected from Akkar et al. (2014), based on the following criteria

■ intra-plate earthquakes with moment magnitudes Mw ranging from 4 to 6.5, whose range is consistent
with the dominant seismic scenario considered by Goda et al. (2013) for the development of UHSs at
the site
■ accelerograms in the far-ﬁeld range, with a site-to-source distance Repi greater than 10 km, thus
disregarding near-fault earthquakes, which tend to produce higher-frequency content in the
accelerograms (Mavroeidis and Papageorgiou, 2003).

Other characteristics such as the duration of strong motion, the rupture mechanism and the directivity of
seismic waves were not considered. The selected accelerograms (Table 5.2) were subsequently scaled at the
intensity level deﬁned by the target acceleration calculated from the design response spectrum corresponding

123

Seismic Design of Foundations

Table 5.2 Selected accelerograms compatible with UHSs for Holyhead

ID Station code/name F–E regiona Mw Year Repi Component

1 0008/Forgaria Cornino Northern Italy 5.2 1976 19 North–south

2 0008/Forgaria Cornino Northern Italy 5.2 1976 19 West–east
3 0009/Buia Northern Italy 5.2 1976 15 West–east
4 0013/Breginj-Fabrika Igli Slovenia 5.9 1976 19 North–south
5 0008/Forgaria Cornino Northern Italy 6.0 1976 18 North–south
6 0008/Forgaria Cornino Northern Italy 4.7 1976 10 West–east
7 0011/Tarcento Northern Italy 4.1 1976 18 North–south
9 0008/Forgaria Cornino Northern Italy 4.5 1976 19 West–east
10 0008/Forgaria Cornino Northern Italy 4.1 1976 10 West–east
11 0013/Breginj-Fabrika Igli Slovenia 5.3 1977 35 North–south
12 3017/Novellara Northern Italy 5.4 1996 13 North–south

a
Flinn–Engdahl Regionalisation Scheme (Young et al., 1996)

to the natural period of the internal structure (i.e. 0.18 s). Figure 5.19 shows the calculated design response
spectrum and the response spectra of the 11-scaled accelerograms. The comparison between the design
response spectrum and the median spectrum of the 11-scaled accelerograms shows good agreement,
conﬁrming the acceptable selection of the time series.

Figure 5.19 Calculated design response spectrum and the response spectra of the 11-scaled accelerograms

0.3 Design response spectrum

11-scaled accelerograms
Spectral acceleration: g

0.2

0.1

0.18 s
0
0.05 0.10 0.5 1
Period: s

124

Seismic Design of Foundations

Seismic analysis methods related to foundation design

5.8. Beam on a non-linear Winkler foundation (SSI)

Piles are the most common form of foundation in seismic areas. This section provides a summary of their
methods of analysis.

Practising engineers often use the beam on a non-linear Winkler foundation (BNWF) approach (Hetényi,
1946; Winkler, 1867) to model laterally loaded piles. This method is based on the hypothesis that the reaction
exerted by the soil at a given depth on the pile shaft is proportional to the relative pile–soil lateral deflection.
In the BNWF method, a pile is modelled by a series of consecutive beam–column elements, whereas the
lateral pile–soil interaction is analysed using non-linear springs that are attached to nodal points between two
consecutive elements. Each spring can be defined by means of a non-linear relationship between the soil
reaction (per unit length of the pile) p and the corresponding relative soil–pile horizontal displacement y. The
coefficient of proportionality between p and y is referred to as the modulus of subgrade reaction k, with
dimensions of pressure divided by length. This relationship is normally known as the p–y or reaction curve.
Figure 5.20 shows a diagram with examples of p–y curves.

In recent years, non-linear p–y curves have been extensively used in the offshore industry. Speciﬁcally, codes
of practices such as those published by API (2014)) and DNV GL (2014))] provide procedures for the
construction of non-linear p–y curves for sandy and clayey soils. It is worth noting that these procedures have
been developed based on a limited number of tests carried out on full-scale small-diameter steel piles sub-
jected to slow-cyclic loading.

Despite the limitations inherent to the discrete nature of the method, the BNWF method is used extensively in
practice owing to its mathematical convenience and ability to incorporate non-linearity of the soil and ground
stratiﬁcation.

Figure 5.20 SSI in a piled foundation and its relationship to the shape of the p–y curve: (a) p–y curves for
non-liqueﬁable sandy soils; (b) p–y curves for fully liqueﬁed soils

Axial load Small-amplitude

vibrations Large displacement

Lateral load
p p

y y
Soil spring p p
(p–y spring)
Pile

y y
p p

y y
End-bearing (a) (b)
soil spring

125

Seismic Design of Foundations

5.8.1 Aside: why does this model work?

The validity of the BNWF approach is based on the assumed similarity between two mechanical system
responses

■ the load–deformation response of the pile, which takes into account the overall macro-behaviour of
the soil–pile system
■ the stress–strain response of the adjacent soil being sheared as the pile moves laterally.

The latter is related to the micro-behaviour of the deforming material. In theory, the transformation from the
micro-scale to the macro-scale can be made by applying appropriate scaling factors, whereby stress is
converted into the equivalent soil reaction, p, and strain is converted into the equivalent relative pile–soil
displacement, y. In routine practice, however, p–y curves are constructed by means of empirical relationships,
developed in the 1970s–1980s based on a relatively limited number of full-scale tests carried out on small-
diameter steel piles (Matlock, 1970; O’Neil and Murchison, 1983; Reese et al., 1974, 1975).

5.8.2 Models for construction of p–y curves for liqueﬁed soils

One of the first methods for the construction of p–y curves for liquefiable soils was that of Dobry et al. (1995).
The method, qualitatively illustrated in Figure 5.21, consists of applying a degradation factor mp to the
conventional p–y curve for non-liquefied sand. This so-called p-multiplier can be conveniently determined
from the equivalent clean sand blow count (N1)60 using empirical charts, such as that in Figure 5.21(b). An
alternative method is to use p–y curves for soft clays, for example those recommended by API (2014)), but
replacing the undrained shear strength of the clay with the residual strength of the liquefied soil Sr (Goh and
O’Rourke, 1999) (Figure 5.22). This method is normally referred to as the residual strength approach. The
residual strength of the liquefied soil can be estimated from empirical charts proposed by several researchers
(Brandenberg, 2005; Olson and Stark, 2002; Pillai and Salgado, 1994; Seed and Harder, 1990).

It can be concluded that, in routine practice, p–y curves for liqueﬁed soils exhibit a strain-softening behaviour,
characterised by a relatively high stiffness at small displacements that gradually reduces upon shearing. This
response, however, is substantially different from the strain-stiffening behaviour observed in both element
and physical model tests. In fact, a number of studies (Ashford and Rollins, 2002; Boulanger et al., 2003;
Tokimatsu et al., 2001; Wilson et al., 2000) have shown that back-calculated p–y curves for liqueﬁed soils
have a concave-upward shape, characterised by practically zero stiffness at small displacements, but
increasing stiffness and strength upon shearing. It is worth noting that this strain-stiffening response –
hereafter referred to as strain hardening – is consistent with the post-liquefaction behaviour of sands observed
in element tests by several researchers (Dash, 2010; Lombardi et al., 2014b; Sitharam et al., 2009;
Sivathayalan and Vaid, 2004; Vaid and Thomas, 1995; Yasuda et al., 1998).

We now consider the effect of different shapes of p–y curves on the seismic response of piled foundations.
Starting from a concave-downward p–y curve, when the lateral displacement is relatively small, the soil–pile
interaction depends on the initial stiffness of the reaction curve. For a large displacement, however, the
response is inﬂuenced by the ultimate value of the soil reaction rather than the foundation stiffness. On the
other hand, if the shape of the p–y curve is concave upward, the response of the pile is highly non-linear and
exhibits practically zero stiffness at small displacements. Moreover, as a result of the limited resistance
offered by the liqueﬁed soil, the pile behaves as an unsupported column, which may be prone to buckling
instability under large axial loads and in the presence of geometrical imperfections.

126

Seismic Design of Foundations

Seismic analysis methods related to foundation design

Figure 5.21 Construction of p–y curves for liqueﬁable soils according to the p-multiplier approach: (a) application
of the degradation factor mp to p–y curves for non-liqueﬁed sands prescribed by API (2014); (b) degradation factor
mp versus the equivalent clean sand blow count from the SPT test (N1)60 (Brandenberg, 2005)

p
Non-liquefied soil
pu
kz
p = pu tanh y
Pu

Liquefied soil
mppu
kz
p = mppu tanh y
Pu

(a) y

1.0

Brandenberg (2005)
0.8 AIJ (2001)
p-multiplier: β

0.6

0.4

0.2

0.0
0 5 10 15 20 25 30
(N1)60
(b)

127

Seismic Design of Foundations

Figure 5.22 Construction of p–y curves for liqueﬁable soils according to the residual strength approach:
(a) schematic p–y curves for soft clays prescribed by API (2014); (b) residual strength Sr versus the equivalent clean
sand blow count from the SPT test (N1)60 (Cubrinovski and Bradley, 2008)

Residual shear strength, Sr: kPa

p
50

pu 40
d
30 un
Sr r bo
0.72 p pe
20 Up d
oun
10 erb
w
Lo
0
0 5 10 15 20
y Equivalent clean sand SPT blow count, (N1)60
(a) (b)

5.8.2.1 p–y curves for liqueﬁed soils

p–y curves for liqueﬁable soils can be obtained by multiplying the conventional p–y curves for sandy soils by
a degradation factor mp that is referred to as the p-multiplier. mp decreases almost linearly with increasing
excess pore pressure, attaining a limiting value of mp = 0.1 under the full liquefaction condition (Figure 5.23).
Other studies recommend values of mp ranging from 1/80 to 1/30, or express mp as a function of blow counts

Figure 5.23 p-multiplier versus the excess pore pressure ratio (Boulanger et al., 2003)

1.0

0.8
p-multiplier, mp

0.6

0.4

0.2
Dobry et al. (1995)

0
0 20 40 60 80 100
Excess pore water pressure ratio, ru: %
mp = 1 – 0.9ru Boulanger et al. (2003)

128

Seismic Design of Foundations

Seismic analysis methods related to foundation design

Figure 5.24 Comparison between the conventional and proposed p–y curves for liqueﬁed soils

No liquefaction
(conventional API (2014) p–y curve)

Soil resistance, p
Full liquefaction
(proposed p–y curve)

Transient to liquefaction
(p-multiplier method)

Pile deflection, y

obtained from SPT tests. It should be noted that the disparities between the different values proposed for
mp by different researchers can probably be attributed to the numerous uncertainties associated with the
procedures used for back-calculating mp.

A different method for modelling p–y curves under the liquefaction condition is based on the assumption that
the liquefied soil behaves as a soft clay. According to this approach, the p–y curves can be obtained from those
recommended by API (2014)) for soft clays, but replacing the undrained shear strength with the residual
strength of the liquefied soil, Sr. The existing p–y curves for liquefiable soils imply a significant stiffness in the
initial phase of the pile–liquefied soil interaction, which is followed by the strain-softening response.
However, this response is in contrast to the response observed in element tests, which is strain hardening. This
is characterised by practically zero stiffness at low strains and increasing stiffness upon shearing. This unusual
strain-hardening response can be attributed to the tendency of the liquefied soil to dilate upon shearing.
In fact, as the liquefied soil tends to dilate, the excess pore pressure gradually dissipates, which in turn leads to
increasing strength and stiffness upon shearing.

In view of these findings, a new set of p–y curves is recommended that can be obtained by modifying the
conventional p–y curves (for non-liquefied soils) in such a way as to replicate the aforementioned strain-
hardening behaviour. A comparison between the proposed p–y curves and those obtained according to the
p-multiplier method is plotted in Figure 5.24. In the following sections, the application of both the
conventional and modified p–y curves is illustrated by means of a series of numerical analysis whose
results are compared against experimental data obtained from large-shaking-table tests.

5.8.2.2 Simplified method for the construction of p–y curves for liquefiable soils
This section presents a step-by-step procedure for the construction of p–y curves for liquefiable soils based on
typical field bore log data.

129

Seismic Design of Foundations

The construction of a p–y curve for a liqueﬁable soil involves four major steps

(a) evaluation of the soil parameters from the bore log data
(b) definition of the geometry of the pile foundations
(c) construction of a simplified stress–strain curve for the liquefied soil
(d) construction of a p–y curve for the liquefied soil from the stress–strain curve evaluated in step (c) above.

(a) Evaluation of the soil parameters from the bore log data. Normally, the bore log data from a site
provides the variation of the SPT N values with depth. However, additional soil parameters are necessary to
evaluate the p–y curves for the liquefaction condition, namely the corrected SPT value (N1), the relative
density (Dr), the critical state stress ratio (Mc) and the residual shear strength of the liqueﬁed soil (Sr) – see
Figure 5.22. These parameters can be estimated from element testing or through standard correlations
available in the literature (given below). However, the limitations and applicability must be checked for the
actual site conditions for the adopted correlations. A list of empirical correlations that can be used to construct
p–y curves for liqueﬁable soils are as follows.

■ For sandy soil, the field SPT value can be corrected for overburden pressure as
N
N1 = pffiffiffiffiffiffiffiffiffiffiffi (5:24)
sv0 /98

where s 0v is the geostatic vertical stress (kPa).

■ The relative density of the soil can be estimated from the corrected SPT value N1 through a number of
correlations available in the literature. The relationship proposed by Curbinovski and Ishihara (1999) is
pffiffiffiffiffiffiffiffiffiffiffiffiffi
Dr = N1 /CD (5:25)

where CD = [9/(emax – emin)]1.7. CD is usually taken as 20 for sand with a ﬁnes content of approximately
20%, 41 for clean sand and 70 for gravelly sand.

■ The maximum shear modulus of the soil (i.e. the shear modulus at small deformations) can be
determined either from ﬁeld tests, such as cross-hole and down-hole tests, or using suitable empirical
correlations, such as the expression proposed by Imai and Tonouchi (1982) below (Gmax is in kPa)

Gmax = 14400N 0:68 (5:26)

■ The residual shear strength of the liqueﬁed soil Sr corresponds to the shear strength mobilised by the
liqueﬁed soil at large deformations. This can be computed based on empirical correlations available in
the literature as a function of results from SPT or cone penetration tests – see, for example, Figure 5.22.

(b) Consideration of pile foundation data. The pile diameter is a necessary parameter for p–y curve
construction. The friction between the pile and soil will inﬂuence the soil resistance to pile displacement, with
a rough pile–soil interface providing more resistance than a smooth interface. Depending on the pile material,
the interface is classiﬁed as either smooth or rough. Typically, a concrete pile interface is considered to be
rough, and a steel pile to be smooth.

130

Seismic Design of Foundations

Seismic analysis methods related to foundation design

(c) Deﬁnition of the geometry of the pile foundations. In addition to the pile diameter, the p–y curve
is a function of the interface friction between the soil and the pile shaft, whereby a rough interface results in
higher resistance than smooth piles. Typically, a concrete pile interface is considered to be a rough interface,
whereas steel piles may be treated as smooth piles.

(d) Construction of a simplified stress–strain curve for the liquefied soil. In the absence of results
from element tests, the following simplified stress–strain curve for the liquefied soil can be considered for the
construction of p–y curves for the liquefaction condition. The stress–strain curve can be defined by four
parameters, described below.

■ Take-off strain gto. This is defined as the engineering shear strain required to mobilise a shear strength
of 1 kPa. It can be determined by the stress–strain response of liquefied samples subjected to undrained
monotonic loading, which have been previously liquefied by the application of cyclic loads. In the
interpretation of the results from triaxial tests, the deviator stress q and the axial strain ϵa need to be
converted into the equivalent shear stress t = q/2 and the engineering shear strain g = 1.5ϵa,
respectively. In the absence of experimental results from representative soil samples, gto can be
obtained from experimental data using

gto = 74:34 − 17:71 ln Dr (%) (5:27)

■ Initial shear modulus G1. Following the deﬁnition of the take-off strain, the initial shear modulus can be
estimated by G1 = 1/gto, where G1 is in kPa. In the absence of experimental results from representative
soil samples, G1 can be obtained using
G1 = 1/gto (kPa) (5:28)

■ Critical state shear modulus G2. This is the shear modulus exhibited by the liquefied soil at large strains.
It may be derived from theoretical considerations that should take into account the effect of the rate of
dilation and the level of confinement. Such an approach, however, would involve a rather laborious
mathematical derivation that seems excessive for the present initial simplified analysis. Looking for
alternative correlations that may exist between G2 and another soil parameter, G2 can be computed based
on the ratio G2/Gmax, where Gmax denotes the tangent shear modulus at small strains (i.e. <10–4%). The
latter can be estimated from the initial void ratio of the sample e and the confining stress s 0c , according
(2:17 − e)2 0 0:5
Gmax = 8400 (sc ) (sc0 and Gmax in kPa) (5:29)
1+e
proposed by Kokusho (1980). Figure 5.25 plots G2/Gmax versus Dr for the collated data. The figure
shows that the computed data are randomly scattered, presumably due to variations in excess pore
pressure between samples prepared at different relative densities. The figure, however, identifies three
different ranges of G2/Gmax, namely 0.1, 0.01 and 0.001, which broadly depend on the degree of packing
of the sample. Considering that G2 is inversely proportional to effective mean principal stress p′, and
directly proportional to Gmax, it is suggested that G2 is obtained using
1 Gmax
G2 = pffiffiffiffi (5:30)
5 p0

131

Seismic Design of Foundations

Figure 5.25 Shear modulus of liqueﬁed soil at large strains, normalised by the maximum shear modulus, G2/Gmax
versus the initial relative density Dr

1.0 Legend
Redhill 110 Toyoura
Fraser River Syncrude
Narita
Normalised shear modulus, G2/Gmax

0.1

0.01

0.001
10 20 30 40 50 60 70 80
Relative density, Dr: %

■ Ultimate shear strength tmax. The proposed stress–strain model for liquefied soils depicted in Figure 5.26(a)
implies that the increase in the shear stress cannot continue indefinitely upon shearing because the
dilatative behaviour exhibited by the liquefied soil terminates at large strains. From a theoretical point
of view, the maximum shear is attained when the pore pressure reaches a negative value equal to
–100 kPa, upon which cavitation of the pore water occurs in the sample. According to this reasoning,
the limiting value of shear stress can be computed as
qmax M ðp0c + 100 kPaÞ
tmax = = (5:31)
2 2
in which M denotes the stress ratio at a critical state under conditions of triaxial compression, and p0c
is the mean effective stress in the geostatic condition. An alternative approach consists of equating
tmax to the residual strength Sr of the liquefied soil
t = Sr (5:32)
Sr can conveniently be estimated based on empirical correlations, such as those given in Figure 5.22(b).

132

Seismic Design of Foundations

Seismic analysis methods related to foundation design

Figure 5.26 Schematic representation of obtaining a p–y curve: (a) simpliﬁed stress–strain curve for liqueﬁed soil;
(b) linearly scaled p–y curve model for the stress–strain model; (c) the smoothed p–y curve model

p = N sτD

(a) (b) (c)

Shear stress, τ : kPa

Soil resistance, p: kN/m

Soil resistance, p: kN/m
Proposed p–y curve
Simplified stress–shear strain curve Linearly scaled p–y curve
pu pu

G2 Proposed
p–y curve

G1 p1 p1
γ10 y1 yu y1 yu
Shear strain, γ : %
Soil–pile displacement, γ : m Soil–pile displacement, γ : m

y = γ D/Mc

5.8.2.3 Construction of p–y curve for liquefied soil from stress-strain curve
The proposed method for the construction of p–y curves from stress–strain curves relies on the similarity
between the load–deflection characteristics of the pile and the mechanical behaviour of the deforming soil.
This involves scaling of the stress and strain into the compatible soil reaction p and the pile deflection y,
respectively. It is assumed that plane strain conditions are established around the pile at any depth. As a result,
soil adjacent to the foundation is expected to flow around the pile from the front to back. Although such an
assumption is acceptable for liquefied soils, the same may not be valid prior to the onset of liquefaction, when
wedge-type failure is likely to occur, particularly at shallow depths. Once the simplified stress–strain model of
liquefied soil is formulated, the corresponding p–y curve parameters are calculated by using the scaling
factors Ms and Ns – see Dash (2010), Bouzid et al. (2013) and Lombardi et al. (2017) for the derivation
of these factors. Figure 5.26 schematically illustrates the process involved in deriving the p–y curves of a
liquefied soil from its stress–strain simplified model. The proposed method requires the derivation of scaling
factors for stress Ns, which can be assumed to be equal to 9.2 for a smooth interface and 11.94 for a rough
interface. The scaling factor for the strain Ns can be taken as equal to 1.87 for a liquefied soil. The other
parameters can be obtained as below.

■ The initial lateral resistance corresponding to the take-off strain is

p1 = Ns 1:25gto G1 D (5:33)

133

Seismic Design of Foundations

■ The initial lateral displacement corresponding to the take-off strain is

1:25g to D
y1 = (5:34)
Ms

■ The ultimate lateral resistance is

pu = Ns tmax D (5:35)

■ The ultimate lateral displacement is

t − (G1 1:25g to ) D
yu = 1:25g to + max (5:36)
G2 Ms

■ To make the p–y curve model effective in the numerical analysis and to represent the behaviour
in a pragmatic way, the p–y curve obtained in the above step is smoothened at its transition using
weighting factors. The smooth p–y relationship can be obtained by using

p1 p + p1 pu − p1 2π y +y
p=w y + A(1 − w) u + tanh y− u 1 (5:37)
y1 2 2 3(yu − y1 ) 2

where A = 0 for y = 0, and A = 1 for y ≠ 0 and w is a weight function given by

1 6π 4y + y
w= 1 − tanh y− 1 u (5:38)
2 yu 6

5.8.2.4 Example: construction of p–y curves for a liquefied soil from a typical ground
profile
This section provides a stepwise description of the construction of p–y curves for a liquefied soil based on the
number of SPT blows N for a depth of 5 m (Figure 5.27).

Steps to obtain p–y curve for liqueﬁed soil from ground proﬁle.
Step 1. The soil is considered at a depth of 5 m, with SPT N value = 5 and unit weight = 17 kN/m3.

(a) The SPT value is corrected for overburden = N1 = 8, using Equation 5.24.
(b) The effective initial vertical stress is s 0v = (17 kN/m3 – 10 kN/m3) × 5 m = 35 kPa.
(c) The effective mean principal stress p′ is taken as (2/3) s 0v = 24 kPa.
(d) Assuming the soil is clean sand, the relative density Dr is computed through Equation 5.25, which
yields 0.45 = 45%.
(e) The residual strength Sr is calculated from the corrected SPT value N1 = 8, based on the plot given in
Figure 5.22(b). Two values corresponding to the lower and upper bounds are then calculated: 1.8 and
20 kPa.

Step 2. The pile consists of a steel pile of 0.6 m diameter, for which a smooth interface can be assumed.

134

Seismic Design of Foundations

Seismic Design of Foundations
Figure 5.27 Construction of p–y curves for a liquefied soil based on SPT data: (a) SPT N-value profile at the site; (b) pile geometry; (c) computed p–y curve for
the liquefied soil layer at a depth of 5 m

Depth: Soil SPT N value

γ : kN/m3
m type 10 20 30 40 50 Ground level
140

Medium 120
5 to course
sand
17
100

Steel pile
10 Medium sand 80

L = 16 m
18
60
15 Upper bound
40
Fine
sand
Pile resistance, p: kN/m Lower bound
20 20 20
0.6 m
0

Non-liquefiable layer Liquefiable layer

25 0 0.01 0.02 0.03 0.04 0.05 0.06
(a) (b) Soil–pile displacement, y: m
(c)

135
Seismic analysis methods related to foundation design
Seismic Design of Foundations

Step 3. The stress–strain parameters for the liqueﬁed soil are derived as follows.

(a) The take-off strain gto is calculated from Equation 5.27, which for a relative density of Dr leads to
gto = 6.97% = 0.0697.
(b) The initial shear modulus is calculated: G1 = 1/gto = 14 kPa.
(c) The critical state shear modulus is calculated using Equation 5.30, which yields G2 = 1758 kPa.

Step 4. From the estimated simpliﬁed stress–strain curve, the p–y curve parameters are calculated as follows.

(a) The scaling factor for stress Ns = 9.2 (assuming a smooth interface for the steel pile).
(b) The scaling factor for strain Ms is 1.87 for fully liqueﬁed soil.
(c) The p–y curve parameters are calculated using Equations 5.33–5.36.
Lower bound

p1 = 6.9 kPa, y1 = 27.5 mm

pu = 80 kN/m, yu = 28.1 mm

Upper bound

p1 = 6.9 kPa, y1 = 27.5 mm

pu = 124.1 kN/m, yu = 29.1 mm

(d) The smooth p–y curve is obtained using Equation 5.37.

The above calculation enables plotting of the p–y curve at 5 m depth, for both the lower-bound and upper-
bound conditions, as shown in Figure 5.27(c). Similar calculation can also be carried out for other depths.

5.9. Summary
Earthquake engineers need to carry out various calculations in order to satisfy code requirements and also to
make sure designs are safe against failure mechanisms. To carry out calculations, certain design parameters
are required, depending upon the area of interest. This is undertaken through seismic hazard analysis.

Seismic hazards can be classified as primary and secondary, depending on the progressive impacts they are
associated with. Primary hazards are the very initial reactions such as ground shaking and fault movement/
ground deformations. Secondary hazards are a result of the primary hazards, and include progressive fault
ruptures, landslides, infrastructural damage, soil liquefaction, ground motion amplification, floods, fire,
nuclear radiation release, tsunamis and cascading earthquakes such as aftershocks (travelling seismicity). A
typical trend of hazards with increasing size of the earthquake (in terms of the moment magnitude Mw) is
shown in Figure 5.28.

The different types of seismic analysis that can be used for structures are summarised in Table 5.3.

136

Seismic Design of Foundations

Seismic analysis methods related to foundation design

Figure 5.28 Typical trend of seismic hazards with increasing magnitude

Associated hazard
2.0
2017 Pahal (Hawaii)
earthquake, 2.6

to be felt
3.0 Unlikely

4.0
2017 Delaware (USA)
earthquake, 4.4
Moment magnitude, Mw

further triggering mechanisms

Ground shaking initiates with

5.0 2017 Tripura (India)

earthquake, 5.7

6.0
ground deformation
Liquefaction and

Landslides
possible
2016 Kumamoto (Japan)
7.0
earthquake, 7.3
2004 Sumatra (Indonesia)
earthquake, 7.8
8.0 Tsunamis
likely

2011 Tōhoku (Japan)

9.0 earthquake, 9.1

As earthquake magnitude increases hazards gets multiplied and often may result in cascading events.

137

Seismic Design of Foundations

Table 5.3 Methods for the seismic analysis of structures

Analysis method Form of excitation Analysis domain Linear vs Comments

(type of response) non-linear

Equivalent static Response spectra Frequency Linear Simplistic but very conservative
method (peak response) approach used when no information
on the dynamic characteristics of the
structure are available.
Modal Frequency Linear Necessary to determine the dynamic
characteristics. Essential part of any
dynamic analysis.
Response spectrum Response spectra Frequency Linear Requires a modal analysis as the
(peak response) ﬁrst step. Used when peak
loads/displacements are required
for design or structural integrity
assessment purposes.
Complex frequency Acceleration time Frequency Linear Not commonly used in dynamic analysis.
response histories (response Useful for the analysis of soil behaviour.
at each time step)
Modal superposition Acceleration time Time Linear Requires a modal analysis as the ﬁrst
histories (response step. Used when secondary response
at each time step) spectra are required and/or when
phasing is of interest.
Implicit direct Acceleration time Time Non- Use when the response is predomi-
integration histories (response linear nantly linear, but where non-linear
at each time step) effects over a short duration
(e.g. impacting of buildings) are of
particular interest.
Explicit direct Acceleration time Time Non- Used when the response is highly
integration histories (response linear non-linear, e.g. when the response is
at each time step) dependent on non-linear soil behaviour.

REFERENCES
AIJ (Architectural Institute of Japan) (2001) Recommendations for Design of Building Foundations. AIJ, Tokyo,
Japan.
Akkar S, Sandıkkaya MA, Şenyurt M et al. (2014) Reference database for seismic ground-motion in Europe.
Bulletin of Earthquake Engineering 12(1): 311–339.
ASCE (American Society of Civil Engineers) (2005) ASCE 43-05. Seismic design criteria for structures, systems,
and components in nuclear facilities. ASCE, Reston, VA, USA.
API (American Petroleum Institute) (2014) Recommended Practice for Planning, Designing, and Constructing
Fixed Onshore Platforms – Working Stress Design, 22nd edn. API, Washington, DC, USA.
Ashford SA and Rollins KM (2002) TILT: The Treasure Island Liquefaction Test. Department of Structural
Engineering, University of California, San Diego, CA, USA. Final Report SSRP 2001/17.

138

Seismic Design of Foundations

Seismic analysis methods related to foundation design

Bhattacharya S, Tokimatsu K, Goda K, Sarkar R, Shadlou M and Rouholamin M (2014) Collapse of Showa Bridge
during 1964 Niigata earthquake: a quantitative reappraisal on the failure mechanisms. Soil Dynamics and
Earthquake Engineering 65: 55–71.
Biot MA (1932) Transient Oscillations in Elastic Systems. PhD thesis, California Institute of Technology,
Pasadena, CA, USA.
BIS (Bureau of Indian Standards) (2002) IS 1893-1. Indian standard criteria for earthquake resistant design of
structures. Part 1: General provisions and buildings. BIS, New Delhi, India.
Boulanger RW, Kutter BL, Brandenberg SJ, Singh P and Chang D (2003) Pile Foundations in Liquefied and
Laterally Spreading Ground During Earthquakes: Centrifuge Experiments and Analyses. Center for Geotech-
nical Modeling, Department of Civil and Environmental Engineering, University of California, Davis, CA,
USA, Report UCD/CGM-03/01.
Bouzid DJ, Bhattacharya S and Dash SR (2013). Winkler Springs (p–y curves) for pile design from stress-strain of
soils: FE assessment of scaling coefficients using the mobilized strength design concept. Geomechanics and
Engineering 5(5): 379–399.
Brandeberg SJ (2005) Behaviour of Pile Foundations in Liquefied and Laterally Spreading Ground. PhD thesis,
University of California at Davis, CA, USA.
CEN (Comité Européen de Normalisation) (2004) EN 1998-1. Design of structures for earthquake resistance.
Part 1: General rules, seismic actions and rules for buildings. CEN, Brussels, Belgium.
Chopra AK (2007) Dynamics of Structures: Theory and Applications to Earthquake Engineering. Pearson/Prentice
Hall, Upper Saddle River, NJ, USA.
Cubrinovski M and Bradley B (2008) Assessment of seismic performance of soil–structure systems. Proceedings
of the 18th New Zealand Geotechnical Society 2008 Symposium, Auckland, New Zealand, pp. 111–127.
Cubrinovski M and Ishihara K (1999) Empirical correlation between SPT N-value and relative density for sandy
soils. Soils and Foundations 39(5): 61–71.
Dash S (2010) Lateral Pile–Soil Interaction in liquefiable Soils. PhD thesis, University of Oxford, Oxford, UK.
DNV GL (2014). DNV-OS-J101. Design of offshore wind turbine structures. DNV, Høvik, Norway.
Dobry R, Taboada V and Liu L (1995) Centrifuge modeling of liquefaction effects during earthquakes. Proceedings
of the 1st International Conference on Earthquake Geotechnical Engineering, Tokyo, Japan, pp. 14–16.
Goda K, Aspinall W and Taylor CA (2013) Seismic hazard analysis for the UK: sensitivity to spatial seismicity
modelling and ground motion prediction equations. Seismological Research Letters 84(1): 112–129.
Goh S and O’Rourke TD (1999) Limit state model for soil–pile interaction during lateral spread. Proceedings of the
7th US–Japan Workshop on Earthquake Resistant Design of Lifeline Facilities and Countermeasures Against
Soil Liquefaction, Seattle, WA, USA, pp. 237–260.
Hetényi M (1946) Beams on Elastic Foundation. Theory with Applications in the Fields of Civil and Mechanical
Engineering. University of Michigan Press, Ann Arbor, MI, USA.
Housner GW (1959) Behavior of structures during earthquakes. Journal of the Engineering Mechanics Division,
ASCE 85(4): 109–130.
Housner GW and Jennings PC (1982) Earthquake Design Criteria. Earthquake Engineering Research Institute,
Berkeley, CA, USA.
ICC (International Code Council) (2000) International Building Code. International Conference of Building
Officials, Whittier, CA, USA.
Imai T and Tonouchi K (1982) Correlation of N-value with S-wave velocity and shear modulus. Proceedings of the
2nd European Symposium of Penetration Testing, Amsterdam, the Netherlands, pp. 57–72.
JRA (Japan Road Association) (2002) Specifications for highway bridges. Part V: seismic design. JRA, Tokyo,
Japan. (In Japanese.)
Kokusho T (1980) Cyclic triaxial test of dynamic soil properties for wide strain range. Soils and Foundations
20(2): 45–60.

139

Seismic Design of Foundations

Lombardi D and Bhattacharya S (2014) Modal analysis of pile-supported structures during seismic liquefaction.
Earthquake Engineering and Structural Dynamics 43(1): 119–138.
Lombardi D, Bhattacharya S, Hyodo M and Kaneko T (2014) Undrained behaviour of two silica sands and
practical implications for modelling SSI in liquefiable soils. Soil Dynamics and Earthquake Engineering
66: 293–304.
Lombardi D, Dash SR, Bhattacharya S, Ibraim E, Wood DM and Taylor CA (2017) Construction of simplified
design py curves for liquefied soils. Géotechnique 67(3): 216–227.
Matlock H (1970) Correlation for design of laterally loaded piles in soft clay. Offshore Technology Conference,
Houston, TX, USA, pp. 77–94.
Mavroeidis GP and Papageorgiou AS (2003) A mathematical representation of near-fault ground motions. Bulletin
of the seismological society of America 93(3): 1099–1131.
Olson SM and Stark TD (2002) Liquefied strength ratio from liquefaction flow failure case histories. Canadian
Geotechnical Journal 39: 629–747.
O’Neill and Murchison JM (1983) An Evaluation of p–y Relationships in Sands. American Petroleum Institute,
University of Texas at Austin, Austin, TX, USA, Report.
Pillai VS and Salgado FM (1994) Post-liquefaction stability and deformation analysis of Duncan Dam. Canadian
Geotechnical Journal 31: 967–978.
PML (Principia Mechanica Ltd) (1981) Seismic Ground Motions for UK Design. PML, Epsom, UK, Report.
Reese L, Cox W and Koop F (1974) Analysis of laterally loaded piles in sand. Offshore Technology Conference,
pp. 95–105.
Reese L, Cox W and Koop F (1975) Field testing and analysis of laterally loaded piles in stiff clay. Offshore
Technology Conference, Houston, TX, USA.
Seed RB and Harder LF (1990) SPT-based analysis of cyclic pore pressure generation and undrained residual
strength. Proceedings of the H.B. Seed Memorial Symposium. Bitech, Richmond, BC, Canada, vol. 2, 351–391.
Seed HB, Ugas C and Lysmer J (1976) Site-dependent spectra for earthquake-resistant design. Bulletin of the
Seismological society of America 66(1): 221–243.
Sitharam TG, Vinod JS and Ravishankar BR (2009) Post-liquefaction undrained monotonic behaviour of sands:
experiments and DEM simulations. Géotechnique 59(9): 739–749.
Sivathayalan S and Vaid YP (2004) Cyclic resistance and post liquefaction response of undisturbed in-situ sands.
Proceedings of the 13th World Conference on Earthquake Engineering, Vancouver BC, Canada, paper 2940.
Tokimatsu K, Suzuki H and Suzuki Y (2001) Back-calculated py relation of liquefied soils from large shaking table
tests. Proceedings of the 4th International Conference on Recent Advances in Geotechnical Earthquake
Engineering and Soil Dynamics, San Diego, CA, USA, pp. 1–6.
Vaid YP and Thomas J (1995) Liquefaction and post liquefaction behavior of sand. Journal of Geotechnical
Engineering 121(2):163–173.
Wilson DW, Boulanger RW and Kutter BL (2000) Observed seismic lateral resistance of liquefying sand. Journal
of Geotechnical and Geoenvironmental Engineering 126(10): 898–906.
Winkler E (1867) Die Lehre von der Elasticitaet und Festigkeit: mit besonderer Rücksicht auf ihre Anwendung
in der Technik für polytechnische Schulen, Bauakademien, Ingenieue, Maschinenbauer, Architecten, etc.
Dominicus, Prague, Bohemia.
Yasuda S, Terauchi T, Morimoto M, Erken A and Yoshida N (1998) Post liquefaction behavior of several sands.
Proceedings of the 11th European Conference on Earthquake Engineering, Balkema, Rotterdam, the Netherlands.
Yoshida N, Tazoh T, Wakamatsu K et al. (2007) Causes of Showa Bridge collapse in the 1964 Niigata earthquake
based on eyewitness testimony. Soils and Foundations 47(6): 1075–1087.
Young JB, Presgrave BW, Aichele H, Wiens DA and Flinn EA (1996) The Flinn–Engdahl regionalisation scheme:
the 1995 revision. Physics of the Earth and Planetary Interiors 96(4): 223–297.

140

Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

ISBN 978-0-7277-6166-8
https://doi.org/10.1680/sdof.61668.141
ICE Publishing: All rights reserved

Chapter 6
Liquefaction: theoretical aspects
6.1. Introduction
Liquefaction is a phenomenon in which the strength and stiffness of a soil are reduced by earthquake shaking
or other rapid cyclic loading. Liquefaction and related phenomena have been responsible for tremendous
amounts of damage in historical earthquakes around the world. Following the twin earthquakes in 1964 that
occurred in Niigata, Japan, and in Alaska, USA, research on soil liquefaction has progressed signiﬁcantly in
terms of understanding the mechanism behind the phenomenon, especially the effect of cyclic loading on
saturated sands, its impact on the built environment and the ways to mitigate the associated problems. The
frequent observation of liquefaction-induced damage during major earthquakes has made liquefaction a
major concern in geotechnical earthquake engineering.

6.2. Concept of effective stress and dilatancy

In many instances, soils have been analysed using continuum mechanics principles, where their behaviour is
considered as a continuous mass; that is, the matter in the body is continuously distributed and ﬁlls the entire
region of space it occupies. Being a continuum, the body can be continually subdivided into inﬁnitesimal
elements with properties being those of the bulk material. As a result, both linear and non-linear behaviour
can be used to model the soil response, for example in analysing subsurface stresses and seismic response.

However, soil consists of grains and, therefore, fundamentally speaking, does not ﬁt the concept of a con-
tinuum. There are many cases where the soil behaves in a manner that is quite different from a continuous
material. The phenomenon of soil liquefaction occurs because sand is a discrete material whose behaviour is
governed by the principle of effective stress and by dilatancy.

Effective stress is deﬁned as the difference between the total stress and the pore water pressure in a fully
saturated soil, that is

s0 = s − u (6:1)

where s is the total stress, s′ is the effective stress and u is the pore water pressure. The physical interpretation
of effective stress is illustrated in Figure 6.1. Consider plane X–X′ in a fully saturated soil, which passes
through the interparticle contacts only. Such plane is actually a curved surface, but because of the relatively
small scale of the particles, it is indistinguishable from a true plane; hence the plane represented in the ﬁgure is
a good approximation. The effective normal stress is therefore approximated as the sum of all the vertical
components P within the area, divided by the area. On the other hand, the pore water pressure will act on the
plane over the entire area. So, from Equation. 6.1, the total normal stress is the sum of the effective stress and
pore water pressure. In effect, the effective stress is the sum of the contact forces among sand grains per unit
area; it is not equal to the contact stress between two particles (where the contact area is small, and therefore
would be much higher than the effective stress).

141

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 6.1 Effective stress in soil

h
P P P P

P P P P
P P P P
P P P P

1.0
1.0

On the other hand, dilatancy is the volume change observed in granular materials when they are subjected to
shear deformations. Figure 6.2(a) illustrates the volume change in an initially loose sand that is sheared under
drained conditions. The individual grains fall into the pores, leading to compaction or volume decrease. The
only way for dense sand to rearrange itself when sheared is for the individual grains to move on top of each
other, as shown in Figure 6.2(b); this results in volume expansion. A material is called dilative if its volume
increases with increasing shear (positive dilatancy), and contractive if the volume decreases with increasing
shear (negative dilatancy). Whether positive or negative, dilatancy is a property speciﬁc to granular materials.

Figure 6.2 Concept of dilatancy as illustrated by the shearing of (a) initially loose sand and (b) initially dense sand
under drained conditions

Decrease in volume

(a)

Increase in volume

(b)

142

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

6.3. Mechanism of pore pressure development

It is widely recognised that the basic mechanism of liquefaction in a deposit of loose saturated sand during
earthquakes is the progressive build-up of excess pore water pressure due to the application of cyclic shear
stresses induced by the upward propagation of shear waves from the underlying rock formation (Ghabousi
and Wilson, 1973, Marcuson, 1978; Seed and Lee, 1966; Terzaghi and Peck, 1948).

As mentioned above, loose dry granular material will tend to densify when subjected to cyclic loading, or
other loadings such as a single large shock. Densiﬁcation occurs when loosely packed granular particles are
rearranged to become more tightly packed, reducing the void space between the particles. As illustrated in
Figure 6.2(a), when loose dry sand is shaken, the sand particles adjust their positions, occupying available
voids, and the top surface falls as the sand particles become more tightly packed. Thus, when subjected to
earthquake shaking, dry sand simply densiﬁes.

On the other hand, if the loose granular material is totally saturated (i.e. water ﬁlls the voids between the
particles), then densiﬁcation cannot occur until some of the water is displaced, as the water in the voids is
incompressible. Thus, the initial response of loose saturated sand to cyclic loading, before there is any change
in volume, is an increase in the pore water pressure.

Before cyclic loading, the soil element is subjected to a confining stress due to the weight of the overlying soil
(Figure 6.3(a)). When cyclic shear stresses are applied, the element of loose sand tends to reduce in volume.
However, if the duration of loading is short compared with the time for drainage to occur, volume contraction
cannot occur immediately. In order to keep the volume of contracting sand constant, some change in the
existing stress system must take place. This stress change is achieved in the form of a reduction in the existing
confining stress, accompanied by an equal increase in the pore water pressure. Hence, the contact forces
between the individual soil particles are lost, resulting in softening and weakening of the soil particle skel-
eton. When the state of sand packing is loose enough and the magnitude of the cyclic shear stress is great
enough, the pore pressure builds up to a point where it becomes equal to the initial confining stress (Ishihara,
1985). At this state, no effective stress (i.e. intergranular stress) is acting on the sand, and the individual
particles lose contact. Such a state is called liquefaction (Figure 6.3(b)). After liquefaction, the individual

Figure 6.3 Behaviour of saturated loose sand when subjected to earthquake shaking under undrained conditions:
(a) before liquefaction; (b) during liquefaction; (c) after liquefaction

Sand
Pore
grain
water
Settlement
τ=

Total stress σ = constant Total stress σ = constant Total stress σ = constant

Effective stress σ ′ Effective stress σ ′ = 0 Effective stress σ ′
(a) (b) (c)

143

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

particles start to sediment, expelling pore water towards the boundary of the deposit. After sedimentation has
taken place throughout the depth, the excess pore water pressures reduce to zero, there is positive effective
stress in the sand particle skeleton and there is a new particle arrangement, which may be denser than before
the liquefaction (Figure 6.3(c)).

As seen from Figure 6.3, the process of soil liquefaction generally involves three phases: (a) pore pressure
build-up, (b) a state of liquefaction with nearly zero effective stress in the soil mass and (c) dissipation of
excess pore water pressures. The liquefaction process occurring at a certain point within the saturated deposit
is depicted in Figure 6.4.

The first phase in which the pore pressure rises to the level of the initial effective overburden stress (and
triggers liquefaction) is very short. During the 1995 Kobe earthquake (Japan), for example, only two or three
cycles (seconds) of intense ground shaking were required to trigger liquefaction in 15–20 m-thick reclaimed
deposits. The rate of pore pressure build-up is primarily influenced by the deformational behaviour of the soil
and ground motion characteristics (amplitude and number of intensive cycles of ground shaking). The
duration of the second phase (during which the effective stress is nearly zero and the soil is liquefied) is much
longer, and depends on a number of factors (soil stratification, permeability of soils, presence and thickness of
impermeable layers, duration of shaking and ground distortion caused by the shaking). During this phase,
significant unbalanced excess pore pressures develop, causing flow of pore water and liquefied soil mass
towards the ground surface. Sand boils are a typical consequence of such non-equilibrium pressures. The final
phase is quite long and, again, very complex. It involves dissipation of excess pore water pressures, sedi-
mentation of soil particles, and solidification and reconsolidation of the deposit. During liquefaction and
dissipation of pore pressures, phenomena such as secondary and progressive liquefaction may occur due to
the upward flow of water and the redistribution of the void ratio (loosening of the soil).

Figure 6.4 The liquefaction process and the development of excess pore water pressure (EPWP)

Large ground
deformation

Initial effective
EPWP overburden stress, σ ′vo = σ ′v-in

Excess pore
water pressure (EPWP)

(seconds) (minutes; hours) Time

144

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

6.4. Laboratory undrained tests

The mechanism of soil liquefaction has been investigated mostly in the laboratory by performing monotonic
or cyclic undrained tests on specimens using either triaxial or torsional shear stress devices. During earth-
quakes, the time of application of seismic load is much faster than the time it takes for water to drain from the
potentially liqueﬁable layer. Thus, the condition is considered to be undrained, and this is simulated in the
laboratory by closing the drainage valves during the load application. By measuring the development of pore
water pressures and shear strains, the mechanism of liquefaction can be best observed.

6.4.1 Undrained behaviour under monotonic loading

The undrained behaviour of sandy soils under monotonic shearing is conventionally used to investigate the
liquefaction mechanism. Generally, the behaviour under monotonic shearing is characterised by three types
of response, as illustrated in Figure 6.5. Very loose sand shows fully contractive behaviour where, following
the peak stress, the sand strain softens until the steady state is reached at large strains. Steady state is the idea
that sand sheared in an undrained manner can undergo unlimited shear deformation (>20% strain) under a
constant magnitude of shear stress and constant volume (Castro, 1969). This strain-softening phase of the
response resembles a flow-type behaviour, and therefore it is commonly referred to as flow or flow defor-
mation (Ishihara, 1993). The shear stress at large strains, called residual strength, is said to be independent of
the initial level of confining pressure. If the density of the sand is somewhat higher, then the strain softening is
followed by strain hardening in which the sand recovers its strength and restores stability. In this case, the
flow takes place over a limited range from the peak stress to the point of phase transformation where dilative
behaviour is initiated (phase transformation is the state where the soil changes its behaviour from contractive

Figure 6.5 Monotonic behaviour of sand: (a) effective stress path; (b) shear stress–shear strain relation
Shear stress, (σ1 – σ3)/2

Shear stress, (σ1 – σ3)/2

Limited flow No flow

No flow
Limited flow

Flow Flow

Mean effective stress, p′ Shear strain, ε1 – ε3

(a) (b)

145

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

to dilative, as defined by Ishihara et al. (1975)). This type of response is also known as flow with limited
deformation or limited flow. Finally, in the case of medium-dense and dense sands, a strain-hardening
response is observed, and ever-increasing shear stress is needed to induce shear strain and eventually obtain
the steady state of deformation. In this case, flow is not induced.

The results of a series of undrained triaxial compression test on loose samples of Toyoura sand with a relative
density of 16% are presented in Figure 6.6(a). It can be seen that the strain-softening behaviour appears
prominently for the specimen subjected to a high confining stress of 0.1 MPa, while under a low-confining
stress of 0.01 MPa such behaviour disappears, and only strain-hardening behaviour is observed. Moreover, it
can be seen that at small strains the stress–strain curves appear to be different, but at large strains, of the order
of axial strains >20%, the curves seem to merge, indicating almost identical behaviour. At this range of strain,
the deviator stress is seen to stay approximately constant, with confining stress also constant, indicating the
steady state condition.

The results for another series of tests, this time on dense samples with a relative density of 64%, are presented
in Figure 6.6(b), where the same tendency is observed in terms of the overall behaviour with respect to the
inﬂuence of the initial conﬁning stress on the stress–strain and pore pressure response. This indicates that
relative density alone is not a good indicator of whether a soil will liquefy and attain large deformation (or a
high pore water pressure) when subjected to earthquake shaking.

Figure 6.6 Undrained behaviour of Toyoura sand: (a) Dr = 16%; (b) Dr = 64% (Ishihara, 1993; courtesy of Oxford
University Press)

Shear stress–shear strain Toyoura sand Shear stress–shear strain

Deviator stress, q = σ 1 − σ 3: MPa

0.15 e = 0.916, Dr = 16% 4.0

σ ′0 = 0.1 MPa
e = 0.917
0.06 MPa
3.0 σ ′0 = 3 MPa
0.10
0.915 e = 0.735
2.0 MPa
0.02 MPa
2.0 0.735
0.916
1.0 MPa
0.05 0.735 Toyoura sand
0.1 MPa e = 0.735, Dr = 64%
1.0
0.01 MPa 0.735
0.917
0 0
0 5 10 15 20 25 30 0 10 20 30
Axial strain, εl: % Axial strain, εl: %

Effective stress path Effective stress path

Deviator stress, q = σ 1 − σ 3: MPa

Steady
0.15 Toyoura sand Toyoura sand
4.0 state
e = 0.916, Dr = 16% e = 0.735, Dr = 64%
}

Steady
3.0
0.10 state
}

e = 0.917
2.0 e = 0.735

0.05 0.915
0.917 1.0 0.735
0.916 0.735
e = 0.735
0 0
0.01 0.02 0.06 0.1 0.1 1.0 2.0 3.0
Effective confining stress, p′ = (σ ′1 + 2σ ′3 )/3: MPa Mean principal stress, p′ = (σ ′1 + 2σ ′3 )/3: MPa
(a) (b)

146

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

Based on the above test results, two conclusions can be made

■ the state of stress of sand at steady state deformation is determined uniquely by the void ratio alone
■ whether a soil specimen will exhibit contractive or dilative behaviour (i.e. the pore water pressure will
increase or decrease) during undrained shear depends not only on the void ratio of the specimen but
also on the initial conﬁning pressure.

6.4.2 Critical state soil mechanics framework

The discussion presented above conﬁrms the basic behaviour of sands, as proposed by Castro (1969) and
embodied in critical-state soil mechanics (Roscoe et al., 1958). It is accepted (e.g. Been et al., 1991) that the
steady state and the critical state are the same condition and are independent of the stress path followed to
reach the ultimate condition (i.e. the ultimate end point when a soil is sheared).

The steady state (or critical state) can be represented by a line in the void ratio (e) versus the mean effective
stress (p′) plot, as illustrated in Figure 6.7, which represents the behaviour of a granular soil loaded in either
undrained or drained triaxial compression. The signiﬁcance of such a critical line on the e–log p′ plot is that it
distinguishes what would be described as ‘loose’ soils from ‘dense’ soils. The term ‘loose’ is used because,
when a soil is sheared, it will rearrange into a denser arrangement, and vice versa with a ‘dense’ soil. The soil
can also be described as ‘contractive’ when it is loose and lies to the right of the steady state line, and ‘dilative’
when it is dense and lies to the left of the line due to the nature of the soil when it is sheared. Under undrained
condition – that is, the volume remains unchanged (i.e. the void ratio is constant) – shearing of sand initially
located ‘loose of critical’ will result in the development of excess pore water pressure (leading to a reduction
in the effective mean stress) as the undrained path moves to the left to reach the steady state line. Similarly,
when the sand is initially located ‘dense of critical’, shearing will involve an undrained path towards the right,
and results in the development of negative pore water pressure (resulting in an increase in the mean effective
stress) upon reaching the steady state condition. For the drained condition, any excess pore water pressure
will not develop; therefore, shearing of sand initially located ‘loose of critical’ will result in densiﬁcation
(decrease in e), with a drained path going vertically down, while shearing of sand initially positioned ‘dense
of critical’ will dilate and loosen (increase in e), as indicated by the drained path going vertically up.

Figure 6.7 Monotonic behaviour of sand in triaxial compression

Loose of critical

Drained path
Undrained path

Undrained path
Void ratio, e

Dense of critical Drained path

Critical state line

(or steady state line)

Mean effective stress, p′ (log scale)

147

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Contractive and dilative behaviour of the sand can be expressed by the state parameter (y) of the sand, which
is represented by

y = e − ec (6:2)

where e is the current void ratio and ec is the void ratio at the steady (critical) state of the soil. A positive state
parameter indicates that the soil is located to the right of the steady state line and can be associated with
contractive behaviour, and vice versa (i.e. negative state parameter for dilative behaviour).

As explained earlier, liquefaction occurs when the excess pore water pressure generated equals the total stress
(i.e. when the effective stress equals zero). This generation of excess pore water pressure becomes signiﬁcant if
the soil layer is subjected to an undrained boundary condition, and can be brought about by two methods:
through statically induced stresses or through cyclic-induced stresses. The difference between these two
methods is the way in which plastic strains are generated. In static liquefaction, the soil needs to be in a ‘loose’
state (i.e. it requires a state parameter greater than zero), and thus exhibit contractive behaviour. In exhibiting
contractive behaviour, the soil will generate excess pore water pressure when shear stress is applied to it, which
will then exert a stress against the soil skeleton, reducing the contact pressure between soil particles, and thus
induce liquefaction. In the case of cyclic liquefaction, the plastic strain on the soil is generated through the
densiﬁcation process brought on by cyclic stress changes, which tend to consolidate the soil particles further.
Unlike static liquefaction, cyclic liquefaction affects all types of soil, regardless of denseness and cohesiveness.

6.4.3 Cyclic undrained response of loose and dense sand

Figure 6.8(a) illustrates the typical result of a cyclic undrained shear test on loose sand using a torsional shear
apparatus. The upper graph indicates the time history of cyclic shear stress ratio applied to the specimen
(harmonic at constant amplitude) while the middle and lower graphs indicate the development of shear strain
and excess pore water pressure (ru = u=s00 ) with time. It is observed that during the ﬁrst stages of loading that
there is negligible shear strain and a low value of ru. However, as the loading progresses, the pore pressure
builds up and eventually approaches a value equal to the initially applied conﬁning pressure (i.e. ru = 1.0).
Such state is usually referred to as initial liquefaction or simply liquefaction (Seed, 1979). At this stage, the

Figure 6.8 Cyclic undrained torsional shear test on (a) loose sand and (b) dense sand
EPWP ratio, Shear strain, Cyclic stress ratio,

EPWP ratio, Shear strain, Cyclic stress ratio,

0.30 0.30
0.15 0.15
τ/σ0′

τ/σ0′

0.00 0.00
–0.15 –0.15
–0.30 –0.30
4.00 4.00
0.00 0.00
γ (%)

γ (%)

–4.00
–8.00 –4.00
–12.00 –8.00
1.00 1.00
u/σ0′

u/σ0′

0.50 0.50

0.00 0.00
(a) (b)

148

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

shear strain develops rapidly, reaching values >10%. For loose sand, the initial liquefaction is taken as a state
of softening, because very large deformation is produced suddenly with complete loss of strength during or
immediately following the 100% pore pressure build-up.

The cyclic response of dense sand under undrained conditions is illustrated in Figure 6.8(b). It is observed
that as cyclic loading progresses, a state of softening is also produced with 100% pore water pressure
build-up (ru = 1.0); however, the deformation thereafter does not grow indeﬁnitely large and complete loss
of strength does not take place. Thus, although dense sand develops excess pore water pressure during
cyclic loading, just as for loose sand, the amplitude of the shear strain does not increase drastically, and the
effective stress decreases rather slowly due to the moderate rate of pore pressure development. Thus,
compacted dense sand has good resistance against liquefaction.

The characteristic behaviour of sand is more clearly understood in terms of the stress path and the shear stress–shear
strain curve. The data obtained from the same tests shown in Figure 6.8 are illustrated in this fashion in Figure 6.9.
It can be seen that the cyclic behaviour of loose sand (Figure 6.9(a)) indicates a steady decrease in the effective
conﬁning stress with continuous cyclic stress application; once the peak point of the cyclic shear stress touches the
phase transformation line (Ishihara et al., 1975), the stress path goes to the right upwards during loading and is
directed left downwards during unloading. In addition, the stress–strain curve during stress application shows that,
with the cyclic loading, the sand deforms signiﬁcantly, producing shear strains in excess of 5%.

On the other hand, the cyclic behaviour of dense sand is depicted in Figure 6.9(b). Similar to loose sand, the
effective confining stress decreases and the shear strain increases with cyclic load application, albeit in a
slower manner. When the phase transformation line is reached, the stress path goes to the right towards the
failure line during loading, and down to the left during unloading until a state of near-zero effective confining
stress is reached. With further application of cyclic load, the rate of shear strain development is further
reduced, eventually reaching a certain limiting value. This state of near-zero (non-zero) effective confining
stress is called cyclic mobility. During this state, the effective stress path exhibits ‘butterfly’ loops with
repeated reduction in the effective stress and subsequent dilation (increase in the effective stress). The latter is
reflected in the stress–strain curve of the soil through a sharp increase in stiffness and a sudden arrest of strain
development.

6.4.4 Liquefaction resistance of sand

The resistance of sand to soil liquefaction is evaluated by performing cyclic undrained tests on sand speci-
mens using either triaxial or torsional shear apparatus. In the tests, samples of saturated sand are consolidated
under a specified confining pressure and subjected to a sequence of constant amplitude cyclic stress until the
specimen either deforms to a certain amount of strain, or the excess pore water pressure reaches a value close
to the initial confining pressure. At this stage, the specimen can be considered to be in a state of cyclic
instability, and liquefaction has occurred. In essence, liquefaction can be defined in two ways: (a) devel-
opment of high pore water pressure, usually 95% of the initial confining pressure, or (b) development of high
strain, typically expressed as 5% double-amplitude axial strain for triaxial tests and 7.5% double-amplitude
shear strain for simple shear tests (Figure 6.10).

In order to deﬁne the onset of liquefaction (the development of 95% pore water pressure or 5% double-
amplitude axial strain), the number of cycles needs to be speciﬁed for a given constant-amplitude uniform
cyclic loading. In practice, three or four tests are performed, with varying amplitudes of the shear stress, and

149

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
150
Figure 6.9 Stress path and stress–strain curve from the cyclic torsion shear test on (a) loose sand and (b) dense sand (Ishihara, 1985)

Torsional shear test Torsional shear test

0.6 τd /σ0′ = 0.229 Dr = 47% 0.6 τd /σ0′ = 0.717
σ0′ = 98 kN/m2, K0 = 1.0 Dr = 47%, K0 = 1.0
Seismic Design of Foundations

0.4 Fuji river sand 0.4 σ0′ = 98 kN/m2

Fuji river sand
0.2 0.2
30 50 80 σ0’ 100 50 σ0′
0 Effective
0
Effective 100

τd /σ0′
τd /σ0′
confining stress confining stress σ0′: kN/m2
0.2 σ’: kN/m2 0.2

0.4 Line of phase 0.4 Line of phase

transformation transformation

0.6 0.6
Stress path Stress path

Torsional shear test τd /σ0′

Torsional shear test
τd /σ0’ τd /σ0′ = 0.717 0.6
τd /σ0′ = 0.229 Dr = 75%
0.6
Dr = 47% σ0′ = 98 kN/m2 0.4
σ0′ = 98 kN/m2 Fuji river sand
0.4 0.2
Fuji river sand
5 3
0.2
3 3 5
7 5
0.2 γ:%
3 3
γ:% 0.4
0.2
0.6

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Stress–strain curve Stress–strain curve

(a) (b)
Liquefaction: theoretical aspects

Figure 6.10 Response of saturated sand under cyclic loading in the undrained condition

20
Shear stress: kPa 10

–10

–20

–30

Initial liquefaction
100
Excess pore pressure: kPa

95% of the initial confining pressure

5 Liquefaction based on strain

Shear strain: %

0
Double-amplitude strain 7.5% (simple shear)
5% (triaxial)

−5

the number of cycles to attain liquefaction is noted (Figure 6.11). Then, each of the shear stress ratios,
obtained by normalising the shear stress by the initial confining pressure, is plotted against the number of
cycles required to achieve either a 95% pore pressure ratio or 5% double-amplitude axial strain. The locus of
points defining each state is called the liquefaction resistance curve. Note that the attainment of liquefaction as
defined in terms of the 95% pore pressure ratio or 5% double-amplitude axial strain may not occur simul-
taneously, and, therefore, two separate curves can be drawn; however, the strain amplitude curve is usually
preferred. In principle, it is customary to consider 15–20 cycles in view of the typical number of significant
cycles present in many time histories of accelerations recorded during past earthquakes. Thus, the lique-
faction resistance (or cyclic strength) is specified in terms of the magnitude of the cyclic stress ratio required
to produce 5% double-amplitude axial strain in 15–20 cycles of uniform load application (Ishihara, 1993;
National Research Council, 1985).

151

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 6.11 Determination of the cyclic resistance curve

Shear stress, τ

Liquefaction
Shear stress ratio, τd /σ ′0

15−20
No. of cycles to liquefaction, N

6.5. Role of numerical modelling

The results of the cyclic undrained triaxial or torsional tests indicate how a small element of the soil material
will behave under uniform initial stresses and uniform changes in stress conditions to identify the conditions
that will lead to liquefaction (density of material, cyclic shear stress, number of cycles of shearing, degree of
saturation, etc.). On the other hand, the response of the material in the ﬁeld depends not only on these factors
but also on various boundary conditions (depth of the layer under consideration, properties of the layers
below and above the target layer, drainage conditions, permeability of the layer, etc.) and characteristics of the
applied earthquake motion (amplitude and duration of shaking).

Thus, the objective of any numerical modelling is to evaluate how a large volume of soil in the ﬁeld will
respond to earthquake motion while making use of the element behaviour observed in the laboratory cyclic
tests. The effects of various factors, such as the soil proﬁle properties and the intensity and frequency of
earthquake shaking, among others, can be evaluated. Finally, modelling can provide a means of evaluating
the state of the soil during shaking, such as degradation in stiffness, generation (and dissipation) of excess
pore water pressure and development of deformation (transient and/or residual).

6.5.1 Non-linear time domain ﬁnite-element method (FEM) analysis

The dynamic behaviour of saturated granular material under cyclic shearing is highly non-linear and changes
dramatically as it goes through the different phases of cyclic loading. Such analyses are generally performed
by means of a mathematical model composed of ﬁnite elements (FEM analysis). The main steps in this
approach can be summarised as follows

■ discretisation of the model ground by means of a ﬁnite-element mesh

■ dynamic characterisation of the materials (shear modulus, damping factor, etc.)
■ choice of the input motion
■ FEM analytical procedure, in which the fundamental equation of motion is
u + C u_ + Ku = R(t)
M€ (6:3)

where M is the mass matrix, C is the damping matrix, K is the rigidity matrix, u is the displacement matrix
u and u_ are the acceleration and velocity vectors) and R(t) is the loading vector simulating the cyclic motion.
(€

152

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

In the realm of earthquake geotechnical engineering, there are two sets of procedures that can perform time
domain analysis for the purpose of evaluating the liquefaction potential of soil structures: (a) total stress and
(b) effective stress analyses.

In a total stress analysis approach, the focus of the assessment is the dynamic response of the soil structure
(stresses, strains, etc.) without considering (or oversimplifying) the effect of pore water pressure developed
during cyclic loading. The results provide accurate modelling of the stress–strain behaviour of the material
and the resulting deformation in the absence of liquefaction; as a result, it provides feedback on some aspects
of the dynamic response. However, total stress analysis cannot accurately simulate the effects of excess pore
pressures and liquefaction. Hence, strictly speaking, it should be applied to non-liquefaction cases only where
excess pore water pressure development will not occur.

On the other hand, a rigorous effective stress analysis permits the evaluation of the dynamic response of soil
structures while considering the effects of excess pore water pressure and eventual soil liquefaction on the
resulting ground deformation. Whereas total stress analysis is a ‘decoupled approach’, effective stress
analysis ‘couples’ the dynamic response with the pore water pressure generation (and dissipation), and
therefore more accurately represents the behaviour of the soil structure under dynamic motion.

Whereas the predictive capacity of effective stress analysis has been verified in many studies, its application in
engineering practice is constrained by two requirements: the required high-quality and specific data on the in
situ conditions, physical properties and mechanical behaviour of materials; and the high demands on the user
regarding the knowledge and understanding – both of the phenomena considered and of particular features in
the adopted numerical procedure. Provided that the above requirements are met, however, effective stress
analysis is an excellent tool for the assessment of the dynamic performance of soil structures in liquefiable soils.

6.5.2 Constitutive modelling

Recent trends in computing liquefaction and the associated ground deformations involve representing the
behaviour of liqueﬁed soil through constitutive models and applying them through ﬁnite-element-based
numerical techniques. One early technique proposed by Lee (1974) is based on the assumption that the
deformation is due to softening of the soil as a result of seismic shaking so that after the earthquake the deposit
will deform to a new condition compatible with the new softened stiffness of the soil. The softened parameters
are based on laboratory tests in which the in situ stress conditions before and during the earthquake are
simulated. Finite-element analyses are then performed using the normal and reduced/softened moduli to
calculate the liquefaction-induced ground displacement.

Many of the presently available techniques for evaluating soil liquefaction and permanent displacements
employ non-linear dynamic stress–strain behaviour of soil in conjunction with plasticity theory. A number of
numerical codes (with appropriate constitutive models for soils) have been developed to model the liquefaction
behaviour of soils. However, it should be emphasised that the ability of the seismic effective stress analysis to
accurately simulate soil behaviour during earthquakes essentially depends on the capability of the constitutive
model to represent real soil behaviour. For cases involving soil liquefaction, this is very challenging because
the soil behaviour is very complex, involving signiﬁcant temporal and spatial variation of the in situ conditions/
state (including changes in the effective stress due to generation/dissipation of excess pore water pressures),
loads (irregular cyclic stresses) and the subsequent stress–strain behaviour (highly irregular stress paths).

In the 1990s, the VELACS (Veriﬁcation of Liquefaction Analysis by Centrifuge Studies) project presented a
good opportunity to verify the accuracy of various analytical procedures. In this project, veriﬁcation and
validation of the various analysis procedures were carried out by comparing their predictions (i.e. ‘class A’

153

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

predictions were made before the relevant experiments were performed) with the measurements recorded in
centrifuge experiments (in terms of excess pore water pressure, acceleration and displacement time histories).
Details of this can be found in the proceedings of the VELACS meeting (Arulanandan and Scott, 1993). Since
then, better understanding of the physical phenomena involved in soil liquefaction and improved compu-
tational capability have led to enhanced constitutive models that can simulate soil behaviour; developments
such at these have resulted in seismic effective stress analysis being an important component in many liq-
uefaction studies. More recently, the approach has been adopted as state-of-the-art practice for the evaluation
of liquefaction problems, including the assessment of liquefaction-induced displacements, effects on struc-
tures and the effectiveness of countermeasures against liquefaction (ISO, 2005; Greater Vancouver Lique-
faction Task Force, 2007). To supplement this, an international collaborative effort called LEAP
(Liquefaction Experiments and Analysis Projects) has been established to produce high-quality experimental
data sets and to undertake a systematic exercise to validate existing computational models of the response and
liquefaction of saturated granular soils (Kutter et al., 2015; Manzari et al., 2014).

6.5.3 Example: numerical simulation of laboratory element tests

To illustrate the necessary features of a constitutive model appropriate for liquefaction analysis, the elastic–
plastic constitutive model called the stress–density (S-D) model (Cubrinovski and Ishihara, 1998a, 1998b) is
presented here. The model permits consideration of excess pore water pressure, flow of pore water through
the soil and detailed modelling of the stress–strain behaviour of soils. It has the capability to accurately
simulate highly non-linear stress–strain behaviour under both monotonic loading (from small strains to large
strains or the steady state of deformation) and irregular cyclic loading; therefore, it is specifically tailored for
the analysis of liquefaction problems, and it has been extensively verified through case studies (Cubrinovski
et al., 2001) and large-scale shake table tests (Cubrinovski et al., 2005).

Figure 6.12 illustrates the liquefaction resistance curve for Toyoura sand corresponding to the attainment of
3% double-amplitude shear strain. The open symbols in the ﬁgure are the results from six cyclic torsional
shear tests on similar specimens of Toyoura sand with a relative density Dr ≈ 60%. Based on these test data,
the corresponding experimental liquefaction resistance curve is approximated by the dashed line.

Figure 6.12 Experimental and simulated liquefaction resistance curves (Cubrinovski, 2011)

0.5

γ DA = 3%

0.4
Simulation
Shear stress ratio, τ/σ ′v0

Experiment
0.3
Model simulation
shown in Figure 6.13

0.2

0.1

Toyoura sand (Dr = 60%)

0
1 10 100
No. of cycles to liquefaction, N

154

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

The purpose of the exercise is to accurately simulate the experimental liquefaction resistance curve. This
is done by varying several parameters of the constitutive model in a trial-and-error procedure to identify
the best-ﬁt value of the parameter(s) to provide the most accurate simulation of the target experimental
curve. Using the S-D model, this is achieved by using one of the dilatancy parameters (Sc). For example,
in Figure 6.12 the solid line shows the S-D model simulated the liquefaction resistance curve by taking
Sc = 0.0055. It is of note that this line was established through a number of element test simulations of
the cyclic behaviour of the soil under various levels of cyclic shear stresses.

Figure 6.13 illustrates one of those simulations, where the effective stress path and the stress–strain curve
simulated by the model are shown. In this model simulation, the soil element was subjected to a uniform
cyclic stress ratio of 0.20, and 11 cycles were required to cause liquefaction and develop 3% double-
amplitude shear strain. The simulation of the liquefaction resistance curve is considered sufﬁciently accurate
for liquefaction analysis. The experimental and analytical liquefaction resistance curves practically coincide
for cyclic stress ratios of <0.2, while there is a small discrepancy at higher stress ratios. It is important to
achieve a good level of accuracy across wide range of cyclic stress ratios corresponding to 1–30 cycles,
because these are loading levels and cycles relevant for strong ground motions of earthquakes of different
magnitudes (Cubrinovski, 2011).

Another illustration of the satisfactory modelling of the essential features of cyclic mobility is shown in
Figure 6.14, which compares the stress paths and stress–strain relations of dense Fujii River sand (Dr = 75%)
as observed experimentally using a torsional shear apparatus (Ishihara, 1985) and numerically (Iai, 1991).
The constitutive model employed is of the generalised plasticity type deﬁned in the strain space, and the
concept of multiple mechanisms is used as a vehicle for decomposing the complex mechanism into a set of
simple mechanisms deﬁned in one-dimensional space. The good agreement shown indicates that the model is
capable of representing the rapid as well as gradual increase in shear amplitude greater than several per cent in
a manner consistent with laboratory data.

Figure 6.13 Effective stress path and the stress–strain curve obtained in element test simulations with the S-D
model for a cyclic stress ratio of 0.20 (Cubrinovski, 2011)

30 Element test simulation: τ /σ ′v0 = 0.20 30 Toyoura sand: Dr = 60% (e = 0.765)

20 20
Shear stress, τ : kPa

Shear stress, τ : kPa

10 10

0 0

–10 –10

–20 –20
Cyclic 10 5 1st cycle 12 11 10
–30 mobility –30 Cyclic mobility

0 20 40 60 80 100 120 –8 –4 0 4 8
Effective overburden stress, σ ′v: kPa Shear strain, γ : %

155

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Figure 6.14 Shear behaviour of Fuji River Sand: (a) measured (Ishihara, 1985); (b) numerically simulated (Iai, 1991)

156
τxy
(–σm0′)
Torsional shear test
Torsional shear test τxy /(–σ
σmm00′) = 0.717
σmm00′) = 0.717
τxy /(–σ 0.6
0.6 Dr = 47%, K0 = 1.0
Dr = 75%
–σ
σmm00′ = 98 kPa
–σ
σmm00′ = 98 kPa Fuji river sand
0.4 0.4
Fuji river sand

0.2
Seismic Design of Foundations

0.2
50
5 3
0
Effective 100 3 5
confining stress − σm′: kPa
0.2 0.2 γxy: %

τxy /(–σm0′)
0.4 Line of phase 0.4
transformation

0.6 0.6
Stress path

(a)

0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0

τxy /(–σm0′)
0.2
0.2
0.4 0.4

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
0.6
0.6
0.8
0.8
0 20 40 60 80 100 6.5 4 2 0 2 4 6.5
–σm′: kPa γxy: %

(b)
Liquefaction: theoretical aspects

6.5.4 Examples: numerical simulation of case histories

Seismic effective stress analysis is also very useful in investigating and examining the real performance of soil
structures during earthquakes. Two examples using case histories from Japan are presented below.

6.5.4.1 Port Island (1995 Hyogoken Nambu earthquake)

Many researchers have analysed the performance of reclaimed islands and port structures during the 1995
Hyogoken Nambu earthquake. Port Island, one of the reclaimed islands off the coast of Kobe City, was
constructed between 1966 and 1980 by transporting and dumping soil in sea water using bottom-dump
barges. The soil used in the reclamation was decomposed weathered granite (known as Masado or Masa soil),
which is well graded with a fairly large amount of gravel. In the northwest corner of the island, a vertical array
of accelerometers had been installed in 1991 at four different depths (ranging from the ground surface to 83 m
depth). The soil proﬁle at the site of the vertical array and the location of the accelerometers are shown in
Figure 6.15.

Figure 6.15 Soil proﬁle at the site of the vertical array in Port Island, showing the location of the accelerometers

Depth: m Soil Shear wave

type velocity Instrument A
0
(0 m)
170 m/s
Gravels
10 and
sands 210 m/s
(reclaimed) Instrument B
(16 m)
20
Alluvial 180 m/s
clay

30 Alluvial 245 m/s Instrument C

sand
(32 m)

Diluvial
40 gravels 305 m/s
and
sands
50

Sand 350 m/s

70 Stiff 303 m/s

clay

80
Instrument D
(83 m)

157

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

These accelerometers were able to record the ground motions induced by the main shock of this earthquake.
The records have been used by many researchers to examine the liquefaction of the artificial fills in the island,
as well as the general ground response of the site following the earthquake. For example, using the recorded
ground motions as well as field investigation results and other material properties obtained prior to the
earthquake, Cubrinovski et al. (1996) performed a fully coupled one-dimensional effective stress analysis to
simulate the acceleration records and to clarify the characteristics of the ground response at this site, which
underwent severe liquefaction. For this purpose, they used the S-D constitutive model, discussed in the
previous section.

It is well-known that the directionality of the ground motion was very pronounced, with the northwest–
southeast and northeast–southwest directions being approximately the directions of the maximum and
minimum shaking intensities, respectively. In one of their simulations, Cubrinovski et al. (1996) used the
recorded motion in the northwest–southeast direction at a depth of 32 m as the base input motion, and the
simulated time histories of acceleration at the ground surface and at 16 m depth are illustrated in Figure 6.16,
together with the actual recorded motions. It is seen that there is very good agreement between the computed
and recorded motions at both depths.

Similarly, Cubrinovski et al. (1996) simulated the development of excess pore water pressure within the
Masado backﬁll. Figure 6.17(a) illustrates the simulated time history of the pore water pressure development

Figure 6.16 Comparison between computed and recorded northwest–southeast acceleration time histories: (a) at
the ground surface; (b) at 16 m depth (Cubrinovski et al., 1996)

0.4
Acceleration: g

0.2

–0.2 Recorded
Computed
–0.4
(a)

0.6
0.4
Acceleration: g

0.2

–0.2

–0.4

–0.6
0 5 10 15 20
Time: s
(b)

158

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

Figure 6.17 (a) Computed excess pore water pressure time history at 10.5 m depth; (b) computed depth distri-
bution of the maximum excess pore water pressure; and (c) computed depth distribution of shear strains
(Cubrinovski et al., 1996)

Excess pore
water pressure: kPa Shear strain: %
200
0 100 200 0 2 4 6
Excess pore water pressure: kPa

Northwest–southeast GL – 10.5 m
0 0
150 σ ′v σ ′v Northwest–southeast

5 5

Depth: m
100

10 10
50

15 15
0
0 10 20
Northwest–southeast
Time: s 20 20
(a) (b) (c)

at a depth of 10.5 m, where it is observed that the excess pore water pressure reached the initial effective
overburden pressure at about 6–7 s after the start of shaking, indicating the onset of liquefaction. The dis-
tribution of the maximum excess pore water pressures and the maximum shear strains with depth in the
reclaimed soil, computed using the northwest–southeast motion, are shown in Figures 6.17(b) and 6.17(c),
respectively. The simulation indicated that the layer at a depth of between 5 and 15 m underwent extensive
liquefaction, where the excess pore water pressure generated reached the initial effective vertical stress,
resulting in cyclic softening, generally after only 1.5–2 cycles of intense shaking. Such a pore pressure
response was associated with maximum strains of 3–4%.

6.5.4.2 Kawagischi-cho (Niigata earthquake)

Another simulation of an actual case history, this time the liquefaction observed at the Kawagishi-cho site
following the 1964 Niigata earthquake, was conducted by Ishihara and Towhata (1980). The constitutive
model used involved integration of the stress path model and the continuum theory of two-phase media to
establish a one-dimensional level ground response analysis procedure in which the time-dependent stiffness
degradation effect caused by the generation and dissipation of excess pore water pressure was taken into
account. Details of the constitutive model are presented elsewhere (Ishihara and Towhata, 1980, 1982).

Following the 1964 earthquake, detailed soil investigations were conducted at the Kawagishi-cho site, where
extensive damage to engineering structures were observed as a result of soil liquefaction, including the
settlement and tilting of apartment buildings. For the simulation, the results from cyclic triaxial tests con-
ducted on undisturbed sand samples, obtained by large-diameter sampler and Osterberg piston samplers were
used, to determine the input parameters. In addition, the acceleration record obtained at a non-liqueﬁed site
180 km from the target proﬁle was used as the base input motion, with the amplitude scaled using an
appropriate attenuation relation.

159

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 6.18 illustrates the computed surface accelerations, based on input parameters derived from the
Osterberg and large-diameter samples, respectively, as well as the actual recorded motions (two components)
at the basement of an apartment building located about 50 m from the sampling site. As seen from the ﬁgure,
the computed and recorded motions on the ground surface resemble each other in that there is an initial period
of high-frequency motion followed by longer-period motions.

Using data from a large-diameter sampler, the distribution of the computed excess pore water pressure versus
depth at several instants of time during seismic shaking is shown in Figure 6.19. It is seen that the pore water
pressure builds up gradually as the shaking proceeds, and at 8.0 s after the start of shaking the loose layer near
the surface is brought into a state of liquefaction to a depth of approximately 12 m, consistent with what was
observed during the actual earthquake.

These simulation studies of actual case studies highlight the role of constitutive modelling not only in
understanding ground response during earthquakes but also in explaining what happens in loose saturated
sandy layers during periods of high intensity seismic shaking. These models are now currently being used not
only to investigate pore water pressure generation but also in the prediction of liquefaction-induced ground
deformations and their effects on structures.

Figure 6.18 Comparison of ground accelerations between those computed (upper two plots) and recorded (lower
two plots) (Ishihara and Towhata, 1980)

200 Acceleration computed based on data from the Osterberg samples

Acceleration: g

100
0
100 Time: s
1 2 3 4 5 7 8 9 10 11 12 13
200
200 Acceleration computed based on data from the large-diameter samples
Acceleration: g

100
0
100

200
200 Acceleration recorded at the basement of the Kawagishi-cho apartment complex
Acceleration: g

100 north–south component

0
100

200
200 Acceleration recorded at the basement of Kawagishi-cho apartment complex
Acceleration: g

100 east–west component

0
100 Time: s
1 2 3 4 5 6 7 8 9 10 11 12 13
200

160

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

Figure 6.19 Distributions of pore water pressures versus depth computed using the soil parameters from a large-
diameter sampler (Ishihara and Towhata, 1980)

Excess pore water pressure: kN/m2

0 50 100
0

5
Initial effective
vertical stress

10
Depth: m

15 10.0 s
4.0 s
8.0 s

2.0 s
6.0 s
20

Kawagishi-cho
(Niigata)
25 Data from large-
diameter sampling

6.6. Actual earthquake records as manifestations of in situ soil

liquefaction
The occurrence of liquefaction within the subsurface layers can affect ground surface motions in a number of
ways. As the excess pore water pressure builds up, the effective stresses decrease and the shear modulus of the
soil, which tends to vary with approximately the square root of the effective stress, also decreases. As a result,
there is a reduction in soil stiffness which in turn, decreases the tendency to transmit high frequency waves
and increases the tendency to transmit low frequency waves.

6.6.1 Case histories of liquefaction in sites with loose sandy deposits

When the subsoil comprises loose sands, the rate of pore water pressure generation increases as the stress
paths begin to approach the phase transformation line, as indicated by laboratory element tests (e.g. see
Figure 6.8(a)). As a result, the softening of the subsoil caused by sharp degradation in stiffness, occurs
rapidly, resulting in rapid change in the frequency content of the surface motion. As illustrated in the
ground motion recorded at the basement of an apartment building in Kawagishi-cho (see lower plots in
Figure 6.18), the high frequencies that dominate the initial stages of the motion are transformed to motions

161

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

with prolonged periods with reduced amplitude. In essence, the liquefied subsoil, which remains very soft
over the duration of shaking and cannot sustain any shear stresses once liquefied, acted as damper and
filtered the high frequency components of the seismic waves as well as reducing their amplitude. Other
acceleration records in liquefied deposits during earthquakes, such as in Port Island vertical array during
the 1995 Kobe earthquake (see Figure 6.16(a)), are consistent with laboratory observations showing clear
evidence of liquid-like behaviour including elongation of the period of the ground motion and loss of
high-frequency content.

6.6.2 Case histories of liquefaction in sites with dense sandy deposits

For dense sands, liquefaction appears as a transient phenomenon during which the effective stress temporarily
drops to zero and the soil develops large shear strains due to an instantaneous reduction in stiffness. The shear
strains progressively increase with cyclic stresses, and typically reach several per cent at the end of the test.
Several acceleration records at liquefied dense sites show clear evidence of cyclic mobility in the field, with
acceleration spikes occurring at the time of a sudden increase in soil stiffness due to a reduction in excess pore
pressures. For example, the north–south accelerations recorded at the ground surface and at a depth of 77 m at
the Port and Harbour Research Institute (PHRI) station, a former beach line, following the 1993 Kushiro-oki
earthquake (Japan) is shown in Figure 6.20. From the record, the following characteristics are evident: most
of the high-frequency components are filtered out after 30 s, after which the motion has an elongated period,
highlighted by spikes at each peak of the motion. Iai et al. (1995) interpreted this spiky acceleration as cyclic
mobility of the dense sand deposit, in which contraction and dilation of sand alternately make the stiffness of
the sand soften and stiffen during the cyclic loading.

Figure 6.20 North–south components of the acceleration time histories recorded at the PHRI station during the
1993 Kushiro-oki earthquake (Japan) (Iai et al., 1995; reprinted with permission from the author)

5
Ground surface North South

0
Acceleration: m/s2

Maximum –4.68
–5

5
Base (–77 m) North South

Maximum 2.04
–5
10 20 30 40 50 60
Time: s

162

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

Other popular and widely studied records are those obtained at the Wildlife Liquefaction Array following the
1987 Superstition Hills earthquake (USA) (Holzer and Youd, 2007). The magnitude 6.6 event produced sand
boils, ground ﬁssures and permanent lateral displacements at the site. The site was instrumented with accel-
erometers (one at the ground surface and another at 7.5 m depth) and six piezometers. Figure 6.21 depicts the
north–south components of the recorded accelerations at the ground surface and at 7.5 m depth, and the
associated excess pore water pressure measured at 2.9 m depth. It is clear that the site experienced a clear and
gradual stiffness degradation associated with an increase in the pore water pressure. It is also clear from the
surface acceleration record that the high-frequency portion of the motion appears to diminish signiﬁcantly at
about 18 s; this is followed by a series of isolated high-frequency pulses of acceleration. Zeghal and Elgamal
(1994) correlated these pulses to temporary drops in the measured pore pressures, and proposed that they were
caused by dilation-induced stiffening. In many cases, these pulses have amplitudes smaller than those of the
pulses that occur prior to liquefaction, but in some cases the peak ground acceleration of the entire motion is
associated with dilation pulses (Kramer et al., 2011).

Figure 6.21 North–south components of the acceleration time histories and associated pore water pressure
recorded at Wildlife site during the 1987 Superstition Hills earthquake (Zeghal and Elgamal, 1994; reproduced with
permission from the publisher, ASCE)

200
North South (surface)
Acceleration: m/s2

100

0
(10) (12) (13)
(3) (8) (11)
–100 (6) (7)
(5)
(1) (2)
–200
200
Acceleration: m/s2

North South (7.5 m depth)

100

–100
Stage 2

Stage 1 Stage 3 Stage 4

–200
1
Pore pressure ratio, ru

0.8
(9) (10) (12)
(13)
(6) (7) (8) (11)
0.6
(4) (5)
(3)
0.4
(2)
(1)
0.2
P5 (2.9 m depth)
0
0 10 20 30 40 50 60 70 80 90 100
Time: s

163

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

6.7. Effects on the built environment

The liquefaction phenomenon by itself may not be particularly damaging or hazardous. Only when lique-
faction is accompanied by some form of ground displacement or ground failure does it becomes destructive to
the built environment. For engineering purposes, it is not the occurrence of liquefaction that is of prime
importance, but its severity or its capability to cause damage. Adverse effects of liquefaction can take many
forms.

6.7.1 Flow failures

Flow failures are the most catastrophic ground failures caused by liquefaction. These failures commonly
displace large masses of soil laterally tens of metres, and, at times, large masses of soil have travelled tens of
kilometres down long slopes at velocities of up to tens of kilometres per hour. Flows may comprise com-
pletely liqueﬁed soil or blocks of intact material riding on a layer of liqueﬁed soil. Flows develop in loose
saturated sands or silts on relatively steep slopes, usually of greater than 3° (Figure 6.22).

Figure 6.22 Flow failure at Niteko Dam during the 1995 Kobe earthquake (Japan). (Photo by S. Kawase, courtesy of
EDMC, Kobe University Library)

164

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

6.7.2 Lateral spreads

Lateral spreads involve the lateral displacement of large superﬁcial blocks of soil as a result of the liquefaction
of a subsurface layer (Figure 6.23). Displacement occurs in response to the combination of gravitational
forces and inertial forces generated by an earthquake. Lateral spreads generally develop on gentle slopes
(most commonly less than 3°) and move towards a free face such as an incised river channel. Horizontal
displacements commonly range up to several metres. The displaced ground usually breaks up internally,
causing ﬁssures, scarps, horsts and grabens to form on the failure surface. Lateral spreads commonly disrupt
the foundations of buildings built on or across the failure, sever pipelines and other utilities in the failure
mass, and compress or buckle engineering structures, such as bridges, founded on the toe of the failure.

6.7.3 Ground oscillation

Where the ground is ﬂat or the slope is too gentle to allow lateral displacement, liquefaction at depth may
decouple overlying soil layers from the underlying ground, allowing the upper soil to oscillate back and forth
and up and down in the form of ground waves (Figure 6.24). These oscillations are usually accompanied by
the opening and closing of ﬁssures and the fracture of rigid structures such as pavements and pipelines.

6.7.4 Tilting and settlement of structures

When the soil supporting a building or other structure liqueﬁes and loses strength, large deformations can
occur within the soil that may allow the structure to settle and tip (Figure 6.25). For example, many buildings
settled and tipped during the 1964 Niigata earthquake (Japan). The most spectacular bearing failures during

Figure 6.23 Lateral ground movements observed near the Avon River following the 2011 Christchurch earthquake
(New Zealand)

165

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 6.24 Buckling of road pavements during the 2007 Niigataken Chuetsu-oki earthquake (Japan) as a result of
lateral ground oscillation caused by liquefaction

Figure 6.25 Sinking of a commercial building during the 2011 Christchurch earthquake (New Zealand) due to loss
of bearing strength

166

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

that event were in the Kawagishi-cho apartment complex where several four-storey buildings tipped as much
as 60° (see Figure 1.6). Apparently, liquefaction ﬁrst developed in a sand layer several metres below the
ground surface and then propagated upwards through overlying sand layers. The rising wave of liquefaction
weakened the soil supporting the buildings and allowed the structures to slowly settle and tip.

6.7.5 Ground settlement

In many cases, the weight of a structure will not be great enough to cause the large settlements associated with
the soil bearing capacity failures described above. However, smaller settlements may occur as soil pore water
pressures dissipate and the soil consolidates after the earthquake, as shown in Figure 6.26. These settlements
may be damaging, although they would tend to be much less so than the large movements accompanying ﬂow
failures, lateral spreading and bearing capacity failures. The eruption of sand boils (fountains of water and
sediment emanating from the pressurised, liqueﬁed sand) is a common manifestation of liquefaction that can
also lead to localised differential settlements.

6.7.6 Buoyant rise of buried structures

Soil liquefaction can also induce buoyant rise of underground structures. Normally, the uplift of a buried
structure is prevented by resistance from the adjacent soil. However, when soil liquefies, the soil loses its
resistance and starts to behave like a liquid with a unit weight almost twice that of water. When the unit weight
of the buried structure is less than that of the liquefied soil, floating of underground structures can occur.
Figure 6.27 shows a manhole that was uplifted by 1.5 m.

̄ hoku earthquake (Japan)

Figure 6.26 Settlement of soil adjacent to a pile-supported building following the 2011 To

167

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 6.27 A manhole uplifted during the 1993 Kushiro-oki earthquake (Japan)

6.7.7 Increased lateral pressure on retaining walls

If the soil behind a retaining wall liqueﬁes, the lateral pressures on the wall may greatly increase. As a result,
retaining walls may be laterally displaced, tilt or structurally fail, as has been observed for waterfront walls
retaining loose saturated sand in a number of earthquakes (Figure 6.28).

6.7.8 Sand boils

Although not strictly a form of ground failure because they alone do not cause ground deformation, sand boils
are diagnostic evidence of an elevated pore water pressure at depth, and are indications that liquefaction has
occurred. During earthquakes, sand boils are formed by water venting to the ground surface from zones of
high pressure generated at shallow depth by the compaction of granular soils during seismic shaking. The
water, which may ﬂow violently, usually transports considerable suspended sediment that settles and forms a
conically shaped sand boil deposit around the vent (Figure 6.29).

6.8. Summary
Liquefaction is a process that involves the reduction in strength and stiffness of the soil due to earthquake
shaking or other rapid cyclic loading. Loose saturated sands are prone to liquefaction because they tend to
contract with shearing; dense sands, on the other hand, dilate under undrained loading. The process of soil
liquefaction generally involves three phases: (a) pore pressure build-up; (b) a state of liquefaction with nearly
zero effective stress in the soil mass; and (c) dissipation of excess pore water pressures.

168

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

Figure 6.28 Lateral movement of a retaining wall in Port Island (Japan)

Figure 6.29 Sand boil observed during the 2000 Tottori-ken Seibu earthquake (Japan)

169

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Laboratory element tests, either through triaxial or torsional shear apparatus, are important tools in under-
standing the response of sands to undrained cyclic loading, such as in the generation of pore water pressure
and the development of strains. In loose sand specimens, the pore pressure builds up and eventually
approaches a value equal to the initially applied conﬁning pressure; this is called initial liquefaction. For
dense sands, on the other hand, further application of a cyclic load results in a reduction in the rate of shear
strain development; such a state of near-zero (non-zero) effective conﬁning stress is called cyclic mobility.

Constitutive models have been developed to simulate the response of sands, and many of these models have
been validated through comparison with laboratory tests and case studies. Acceleration time histories
recorded following actual earthquakes have shown manifestations of the effects of liquefaction and cyclic
mobility on subsurface deposits.

The adverse effects of liquefaction can take many forms. Generally, only when liquefaction is accompanied
by some form of ground displacement or ground failure does it become destructive to the built environment.
Thus, for engineering purposes, it is not the occurrence of liquefaction that is of prime importance but its
severity or its capability to cause damage.

REFERENCES
Arulanandan K and Scott RF (1993) Verification of numerical procedures for the analysis of soil liquefaction
problems. Proceedings of the International Conference on the Verification of Numerical Procedures for the
Analysis of Soil Liquefaction Problems. Balkema, Rotterdam, the Netherlands, vol. 1.
Been K, Jefferies MG and Hachey J (1991) The critical state of sands. Géotechnique 41(3): 365–381.
Castro G (1969) Liquefaction of Sands. Harvard Soil Mechanics Series, No. 81. Harvard University, Cambridge,
MA, USA.
Cubrinovski M (2011) Seismic effective stress analysis: modelling and application. 5th International Conference
on Earthquake Geotechnical Engineering, Santiago, Chile.
Cubrinovski M and Ishihara K (1998a) Modelling of sand behaviour based on state concept. Soils and Foundations
38(3): 115–127.
Cubrinovski M and Ishihara K (1998b) State concept and modified elastoplasticity for sand modelling. Soils and
Foundations 38(4): 213–225.
Cubrinovski M, Ishihara K and Tanizawa F (1996) Numerical simulation of Kobe Port Island liquefaction. 11th
World Conference on Earthquake Engineering, Acapulco, Mexico, paper 330.
Cubrinovski M, Ishihara K and Kijima T (2001) Effects of liquefaction on seismic response of a storage tank on pile
foundations. 4th International Conference on Recent Advances in Geotechnical Earthquake Engineering and
Soil Dynamics, San Diego, CA, USA.
Cubrinovski M, Sugita H, Tokimatsu K, Sato M, Ishihara K, Tsukamoto Y and Kamata T (2005) 3-D numerical
simulation of shake-table tests on piles subjected to lateral spreading, Proceedings of the TC4 Satellite
Conference on Recent Developments in Earthquake Geotechnical Engineering, Osaka, Japan.
Ghabousi J and Wilson EL (1973) Liquefaction and analysis of saturated granular soils. Proceedings of the 5th
World Conference on Earthquake Engineering, Rome, Italy, vol. 1, 380–389.
Greater Vancouver Liquefaction Task Force (2007) Task Force Report. Geotechnical Design Guidelines for
Buildings on Liquefiable Sites in Accordance with NBC 2005 for Greater Vancouver Region. Greater
Vancouver Liquefaction Task Force, Vancouver, Canada.
Holzer TL and Youd TL (2007) Liquefaction, ground oscillation, and soil deformation at the Wildlife Array,
California. Bulletin of the Seismological Society of America 97(3): 961–976.
Iai S (1991) A strain space multiple mechanism model for cyclic behavior of sand and its application. Earthquake
Engineering Research Note No. 43. Port and Harbor Research Institute, Ministry of Transport, Japan.

170

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: theoretical aspects

Iai S, Morita T, Kameoka T, Matsunaga Y and Abiko K (1995) Response of a dense sand deposit during 1993
Kushiro-Oki earthquake. Soils and Foundations 35(1): 115–131.
Ishihara K (1985) Stability of natural deposits during earthquakes. Proceedings of the 11th International
Conference on Soil Mechanics and Foundation Engineering, San Francisco, CA, USA, vol. 11, pp. 321–376.
Ishihara K (1993) Liquefaction and flow failure during earthquakes – the 33rd Rankine Lecture. Géotechnique
43(3): 351–415.
Ishihara K and Towhata I (1980) One-dimensional soil response analysis during earthquake based on effective
stress method. Journal of the Faculty of Engineering, University of Tokyo 35(4): 654–700.
Ishihara K and Towhata I (1982) Dynamic response analysis of level ground based on the effective stress method.
In Soil Mechanics—Transient and Cyclic Loads (Pande GN and Zienkiewicz OCs (eds)). Wiley, New York,
NY, USA, pp. 133–172.
Ishihara K, Tatsuoka F and Yasuda S (1975) Undrained deformation and liquefaction of sand under cyclic stresses.
Soils and Foundations 15(1): 29–44.
ISO (International Organization for Standardization) (2005) ISO 23469. Bases for design of structures. Seismic
actions for designing geotechnical works. ISO, Geneva, Switzerland.
Kramer SL, Hartvigsen AJ, Sideras SS and Ozener PT (2011) Site response modeling in liquefiable soil deposits.
4th IASPEI/IAEE International Symposium: Effects of Surface Geology on Seismic Motion, Santa Barbara,
CA, USA.
Kutter BL, Carey TJ, Hahimoto T, Manzari MT A Vasko A, Zeghal M and Armstrong RJ (2015) LEAP databases
for verification, validation and calibration of codes for simulation of liquefaction. Proceedings of 6th
International Conference on Earthquake Geotechnical Engineering, Christchurch, New Zealand.
Lee KL (1974) Seismic Permanent Deformation in Earth Dams. School of Engineering and Applied Science,
University of California, Los Angeles, CA, USA, Report UCLA-ENG-7497.
Manzari MT, Kutter BL, Zeghal M et al. (2014) LEAP projects: concept and challenges, 4th International
Conference on Geotechnical Engineering for Disaster Mitigation and Rehabilitation, Kyoto, Japan.
Marcuson WF (1978) Definition of terms related to liquefaction. Journal of Geotechnical Engineering Division,
ASCE 104(9): 1197–1200.
National Research Council (1985) Liquefaction of Soils during Earthquakes. National Academy Press, Washing-
ton, DC, USA.
Roscoe KH, Schofield AN and Wroth CP (1958) On the yielding of soils. Géotechnique 9: 71–83.
Seed HB (1979) Soil liquefaction and cyclic mobility evaluation for level ground during earthquakes. Journal
Geotechnical Engineering Division, ASCE 105(GT2): 201–255.
Seed HB and Lee KE (1966) Liquefaction of saturated sands during cyclic loading. .Journal of the Soil Mechanics
and Foundations Division, ASCE 92(SM6): 105–134.
Terzaghi K and Peck RB (1948) Soil Mechanics in Engineering Practice. Wiley, New York, NY, USA.
Zeghal M and Elgamal AW (1994) Analysis of site liquefaction using earthquake records, Journal of Geotechnical
Engineering, ASCE 120(6): 996–1017.

171

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

ISBN 978-0-7277-6166-8
https://doi.org/10.1680/sdof.61668.173
ICE Publishing: All rights reserved

Chapter 7
Liquefaction: practical aspects
7.1. Introduction
Past large-scale earthquakes have demonstrated that buildings, bridges and other civil engineering struc-
tures located at sites of shallow groundwater and/or adjacent to bodies of water are highly susceptible to
liquefaction-induced damage. In order to prevent and/or minimise such damage, there is a need to evaluate the
liquefaction potential at the site, and to predict the associated ground deformations in order to evaluate how
they affect the structures.

The evaluation of soil liquefaction and its effects involve several steps, from determining the seismic demand
at the site to evaluating liquefaction susceptibility, liquefaction triggering and liquefaction-induced defor-
mations. These steps may involve either simpliﬁed or detailed analysis procedures

(a) liquefaction susceptibility (are the soils at the site liqueﬁable or not?)
(b) triggering of liquefaction (is the ground motion of the adopted ‘design earthquake’ strong enough to
trigger liquefaction at the site?)
(c) liquefaction-induced ground deformation (if liquefaction occurs, what will be the extent and
magnitude of the resulting ground deformation?)
(d) effects of liquefaction on structures (what will be the effect of liquefaction on the seismic performance
of structures?)
(e) mitigation of liquefaction and its consequences (if liquefaction-induced ground displacements are
intolerable, what countermeasures should be used against soil liquefaction?).

This chapter will focus on the ﬁrst three questions: susceptibility, triggering and ground deformations. The
fourth question will be answered in Chapters 8–10, and the last question will be addressed in Chapter 12.

7.2. Liquefaction susceptibility

The susceptibility of soil deposits to liquefaction is determined by a combination of various factors to which
they may be subjected, such as soil properties, geological conditions and ground motion characteristics.
These factors are summarised in Table 7.1. Note that soil properties and geological conditions determine the
resistance of the deposit to liquefaction, while earthquake characteristics control the seismic loading
conditions.

Among the factors listed in the table, the three most important ones are

■ the ground is a loose sandy deposit

■ the groundwater table is high and the ground is saturated
■ the earthquake intensity is sufﬁciently large and the duration of shaking is sufﬁciently long.

173

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Table 7.1 Factors affecting liquefaction susceptibility

Soil properties Geological conditions Earthquake characteristics

Unit weight, grain size character- Geological history (ageing, Intensity of ground shaking
istics, relative density, fines cementation), effective confining (horizontal acceleration,
content, soil structure, shear pressure, initial static shear stress, magnitude), duration of shaking
modulus, damping ratio, degree overconsolidation ratio, boundary (or number of cycles), direction of
of saturation conditions (drainage, seepage, shearing, strain level
deformation), lateral earth
pressure coefficient

7.2.1 Geological and geomorphological criteria

Geomorphology helps not only in identifying the most likely physical characteristics of a deposit but also
provides clues regarding the groundwater depth, surface slope, depth and aerial extent of subsurface strata.
Geomorphological and history-based techniques have been employed to identify liquefaction-susceptible
units in regions of high seismic hazards. The geomorphological classiﬁcation focused on key factors such as

■ correlation between geomorphological units and geotechnical properties

■ identiﬁcation of former stream channels, approximate dates, materials and methods of reclamation
■ the groundwater conditions, for example the depth of the groundwater table, the distribution of springs
and artesian conditions.

A detailed investigation was conducted by several researchers to correlate the geomorphology to the
liquefaction susceptibility of various regions. Tables 7.2 and 7.3 show the estimated susceptibility to
liquefaction of various geomorphological conditions based on studies in the USA and Japan, respectively.
Note from these tables that recently deposited sands and silts in areas with high groundwater levels have
high liquefaction susceptibility. Generally, the younger and looser the sediment, and the higher the water
table, the more susceptible the soil is to liquefaction. Sediments most susceptible to liquefaction include
Holocene (less than 10 000 years old) delta, river channel, flood plain and aeolian deposits, and poorly
compacted fills. Liquefaction has been most abundant in areas where groundwater lies within 10 m of the
ground surface; few instances of liquefaction have occurred in areas with groundwater deeper than 20 m.
Dense soils, including well-compacted fills, have low susceptibility to liquefaction.

7.2.2 Compositional criteria

Compositional characteristics associated with high volume-change potential, such as particle size, shape and
gradation, tend to be associated with high liquefaction susceptibility. Uniformly graded soils have higher
volume-change potential and, therefore, tend to build up pore pressure more readily than well-graded soils.
Based on this concept and on field evidence, the Ministry of Transport in Japan has proposed a zone
expressed in terms of grain size distribution separating potentially liquefiable soils from fine-grained soils that

174

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

Table 7.2 Estimated susceptibility of sedimentary deposits to liquefaction during strong seismic shaking

Type of deposit General distribution of Likelihood that cohesionless sediments, when saturated,
cohesionless sediments would be susceptible to liquefaction (by age of deposit)
in deposits
<500 years Holocene Pleistocene Pre-Pleistocene
Continental deposits
River channel Locally variable Very High High Low Very low
Flood plain Locally variable High Moderate Low Very low
Alluvial fan and plain Widespread Moderate Low Low Very low
Marine terraces and Widespread — Low Very low Very low
plain
Delta and fan-delta Widespread High Moderate Low Very low
Lacustrine and playa Variable High Moderate Low Very low
Colluvium Variable High Moderate Low Very low
Talus Widespread Low Low Very low Very low
Dunes Widespread High Moderate Low Very low
Loess Variable High High High Unknown
Glacial till Variable Low Low Very low Very low
Tuff Rare Low Low Very low Very low
Tephra Widespread High High ? ?
Residual soils Rare Low Low Very low Very low
Sebka Locally variable High Moderate Low Very low
Coastal zone
Delta Widespread Very high High Low Very low
Estuarine Locally variable High Moderate Low Very low
Beach
High energy Widespread Moderate Low Very low Very low
Low energy Widespread High Moderate Low Very low
Lagoonal Locally variable High Moderate Low Very low
Foreshore Locally variable High Moderate Low Very low
Artificial
Uncompacted fill Variable Very high — — —
Compacted fill Variable — — — —

Youd and Perkins (1978); reproduced with permission from the publisher, ASCE

175

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Table 7.3 Susceptibility of detailed geomorphological units to liquefaction subjected to ground motion of Japan
Meteorological Agency intensity V or moment magnitude VIII

Geomorphological condition Liquefaction potential

Classiﬁcation Speciﬁc condition

Valley plain Valley plain consisting of gravel and cobble Not likely
Valley plain consisting of sandy soil Possible
Alluvial fan Vertical gradient >0.5% Not likely
Vertical gradient <0.5% Possible
Natural levee Top of natural levee Possible
Edge of natural levee Likely
Back marsh Possible
Abandoned river channel Likely
Former pond Likely
Marsh and swamp Possible
Dry river bed Dry river bed consisting of gravel Not likely
Dry river bed consisting of sandy soil Likely
Delta Possible
Bar Sand bar Possible
Gravel bar Not likely
Sand dune Top of dune Not likely
Lower slope of dune Likely
Beach Beach Not likely
Artiﬁcial beach Likely
Inter-levee road Likely
Reclaimed land by drainage Possible
Reclaimed land Likely
Spring Likely
Fill Fill on boundary zone between sand and lowland Likely
Fill adjoining cliff Likely
Fill on marsh or swamp Likely
Fill on reclaimed land by drainage Likely
Other types of ﬁll Possible

Wakamatsu (1992)

176

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

Figure 7.1 Relationship between grain size characteristics and liquefaction potential: (a) soil with a low coefﬁcient
of uniformity, Cu < 3.5; (b) soil with a high coefﬁcient of uniformity, Cu > 3.5 (Tsuchida, 1971)

Percentage finer by weight: %

100

75 High possibility
of liquefaction
50 Possibility of
liquefaction
25

0 0.01 0.1 1.0 10

Grain size: mm
Clay Silt Sand Gravel
0.005 0.074 2.0
(a)
Percentage finer by weight: %

100

75 High possibility
of liquefaction
50 Possibility of
liquefaction
25

0 0.01 1.0 10
0.1
Grain size: mm
Clay Silt Sand Gravel
0.005 0.074 2.0
(b)

are not vulnerable to liquefaction (Figure 7.1). As seen in the figure, two alternative charts can be used,
depending on whether the uniformity coefficient (Cu = D60/D10) is less or more than 3.5. Note that D60 and
D10 are the diameters corresponding to 60% and 10% fines passing, respectively. Subsequent studies,
however, suggest that even finer soils than those proposed possessed a high degree of potential to liquefaction
if the fines are non-plastic. Moreover, well-graded reclaimed fills containing 30–60% gravels, well beyond
the zones proposed, liquefied following the 1995 Kobe earthquake (Japan). Thus, based on current knowl-
edge, grading criteria alone are not a reliable indicator of liquefaction susceptibility.

7.2.3 Liquefaction susceptibility of ﬁne-grained soils

It is widely accepted that sands, non-plastic silts, gravels and their mixtures form soils that are susceptible to
liquefaction. Clays, on the other hand, do not exhibit typical liquefaction features (even though they may
significantly soften and fail under cyclic loading) and therefore are considered non-liquefiable. The key
difference in the context of liquefaction is that the pore pressure rise in clayey soils is typically limited to 60–
80% of the overburden stress (whereas liquefiable soils generate 100% rise in the excess pore water pressure).

177

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

The largest problem exists in the evaluation of liquefaction susceptibility of fine-grained soils that are in the
transition zone between the above-mentioned liquefiable soils and non-liquefiable clays. Based on the pre-
dominant matrix structure of fines-containing sands, we can distinguish several sub-groups of these soils as
below

■ sands (gravelly sands) with 10–20% ﬁnes

■ sands (gravelly sands) with 20–40% fines
■ fine-grained soils with over 50% fines.

The first group is basically ‘controlled’ by a sand matrix, with the fines filling the voids in between sand
particles. The third group is a fine-grained matrix structure with large sand particles ‘floating’ in the fine-
grained matrix. Soils in the second group are in the transition zone between those with a sand matrix and those
with a fine-grained matrix. This classification is useful for further discrimination between liquefiable and non-
liquefiable soils based on plasticity criteria; that is, what threshold value of the plasticity index (PI) should be
adopted for each of the three soil groups in order to discriminate between liquefiable and non-liquefiable
soils?

The first group of soils is closer to clean sands, and hence rigorous plasticity criteria (higher PI value) are
needed in order to deem the soil non-liquefiable (especially because the amount of fines is very low, 10–20%).
The second group of ‘transitional soils’ is the most difficult to characterise; in the absence of specific data and
evidence, these soils should be conservatively evaluated using the same criteria as for the first group. The soils
in the third group are fine-grained (silts or clays) and hence should be assessed using different plasticity
criteria (lower PI value can be justified by a high fines content of over 50%). Studies by Boulanger and Idriss
(2006) and Bray et al. (2004) provide criteria based on the fines content and the PI for fines-containing sands
and fine-grained soils. It is important to point out that

■ there are no widely accepted criteria for liquefaction susceptibility

■ soils that are deemed non-liquefiable can undergo significant deformation due to cyclic softening
(evaluation of seismic deformation should not be limited to liquefiable soils).

7.3. Liquefaction triggering

The evaluation of whether liquefaction will occur or not for a given site is very important in the design of
structures. For this purpose, two approaches are currently available

■ the use of semi-empirical methods where a set of independent procedures are used to estimate the
stresses induced by the earthquake together with empirical charts/equations to calculate the
liquefaction triggering
■ numerical methods where a complete evaluation of liquefaction, including triggering, ground
deformation and effects on structures in a complex time history analysis, is performed.

For the semi-empirical, or ‘simpliﬁed’, methods, the calculation, or estimation, of two variables is required
for the evaluation of the liquefaction triggering of soils

■ the seismic demand on a soil layer, expressed in terms of the cyclic shear stress ratio (CSR)
■ the capacity of the soil to resist liquefaction, expressed in terms of the cyclic resistance ratio (CRR).

178

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

7.3.1 Cyclic shear stress ratio (CSR)

The most common approach to characterisation of earthquake loading is through the use of cyclic shear
stresses. By normalising the cyclic shear stress amplitude by the initial effective vertical stress, a cyclic stress
ratio (CSR) can represent the level of loading induced at different depths in a soil proﬁle by an earthquake.
There are, in general, two different procedures for evaluating the cyclic shear stresses

■ site response analyses may be performed, using different computer programs, such as SHAKE
(Schnabel et al. 1972), NERA (Bardet and Tobita, 2001), STRATA (Kottke and Rathje, 2008) and
DeepSoil (Hashash and Park, 2001)
■ a ‘simpliﬁed’ approach may be used to estimate the CSR as a function of the peak ground surface
acceleration amplitude (e.g. AIJ, 2001; Idriss and Boulanger, 2008; JRA, 2002; Seed and Idriss, 1971,
1982; Youd et al., 2001).

In this procedure, the CSR developed at a particular depth beneath a level ground surface is estimated by

tave amax s0
0 = 0:65 r (7:1)
s0 g s00 d

where tave is the average cyclic shear stress during a particular time history, s00 is the effective overburden
stress at the depth in question, s0 is the total overburden stress at that depth, amax is the peak horizontal ground
acceleration generated by the earthquake, g is the acceleration due to gravity and rd is the stress reduction
factor.

This equation has remained largely unchanged since its development. The stress reduction factor, initially
presented graphically by Seed and Idriss (1971) and plotted as a range of possible values that decreased with
depth z, as shown in Figure 7.2, has been subsequently reﬁned by various researchers to incorporate the
effects of other parameters, such as earthquake magnitude, acceleration and soil shear wave velocity. Typical
forms of rd equations are presented by Liao and Whitman (1986), Youd et al. (2001), Idriss (1999) and Cetin
et al. (2004), among others.

Regardless of which rd correlation is ‘most correct’, rd curves cannot be ‘mixed and matched’; that is, each
correlation should only be used within its recommended procedure (Idriss and Boulanger, 2008; Seed, 2010).

7.3.2 Cyclic resistance ratio (CRR)

The dynamic stress that the soil can withstand before liquefying is deﬁned in terms of the CRR. The eval-
uation of the cyclic resistance to liquefaction of in situ deposits can be grouped into two methods: (a) using
ﬁeld-obtained data, typically from the penetration resistance obtained using standard penetration tests (SPTs)
or cone penetration tests (CPTs); and (b) from laboratory testing of high-quality samples (Figure 7.3). Note
that in both cases the in situ state of the soil, which has large effect on its liquefaction resistance, is preserved.

In North America, for example, charts have been proposed by correlating the SPT N value (Seed and Idriss,
1971; Seed et al., 1985), the CPT qc value (Moss et al., 2006; Robertson and Wride, 1998; Suzuki et al.,
1995) and shear wave velocity Vs (Andrus and Stokoe, 1997, 2000) and the estimates of CSR of a number
sites that had or had not manifested evidence of liquefaction during major earthquakes in the past (e.g. Seed
et al., 2003; Youd et al., 2001). By plotting the CSR–(N1)60 (or CSR–qc) pairs for cases in which lique-
faction was and was not observed, a curve that bounds the conditions, at which liquefaction has historically
been observed, can be drawn ((N1)60 is the normalised SPT blow count). This curve, when interpreted as the
maximum CSR for which liquefaction of a soil with a given penetration resistance can resist liquefaction,
can be thought of as a curve of the CRR.

179

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 7.2 Stress reduction factor rd versus depth curves developed by Seed and Idriss (1971). (Adapted from
Youd et al., 2001; reproduced with permission from the publisher, ASCE)

Stress reduction coefficient, rd

0 0.2 0.4 0.6 0.8 1.0
0

Average values
by Seed and
Depth: m 5 Idriss (1971)

10
Range for different
soil profiles by
Seed and Idriss (1971)
15
Simplified procedure
not verified with
case history data
in this region
20

Figure 7.3 Methods of estimating the CRR

Estimation of CRR

Field-based Laboratory-based
approach approach

Empirical correlation
0.6
0.5
0.4 Liquefaction
Liquefaction resistance curve
CRR Boundary CRR 0.3
observed

0.2

No liquefaction 0.1
observed
0
1 10 100
Field parameter Number of cycles to 5%
(N1)60, qc1N or VS1 double-amplitude strain, Nc

180

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

7.3.2.1 Evaluation of the CRR using an SPT-based procedure

The SPT is used to determine the SPT N value, which gives an indication of the soil stiffness and can be
empirically related to many engineering properties. The test is conducted inside a borehole. A split-spoon
sampler is attached to the bottom of a core barrel and lowered into position at the bottom of the borehole. The
sampler is driven into the ground by a drop hammer weighing 68 kg falling through a height of 76 cm. The
number of hammer blows is counted. The number required to drive the sampler three successive 150 mm
increments is recorded. The ﬁrst increment (0–150 mm) is not included in the N value, as it is assumed that the
top of the test area has been disturbed by the drilling process. The SPT N value is the number of blows
required to achieve penetration from 150 to 450 mm. The hammer weight, drop height, spoon diameter, rope
diameter and so on are standard dimensions. After the test, the sample remaining inside the split spoon is
preserved in an airtight container for inspection and description.

Criteria for the evaluation of liquefaction resistance based on the SPT are largely embodied in the CSR
versus (N1)60 plot, reproduced in Figure 7.4. (N1)60 is the SPT blow count normalised to an overburden
pressure of approximately 100 kPa (1 ton/ft2) and a hammer energy ratio or hammer efﬁciency of 60%.
The normalisation is usually done using

ERm
ðN1 Þ60 = Cn N (7:2)
60 m

n
Pa
Cn = 0 (7:3)
sv0

where Cn is a correction coefficient for the overburden pressure, ERm is the actual energy efficiency delivered
to the drill rod, and Nm is the measured N value. In the overburden correction factor, Pa = 1 atm of pressure in
the same units as s0v0 , and n is an exponent that varies with the soil type. Hence, knowing (N1)60, the CSR
required to induce liquefaction for a an earthquake with a moment magnitude (Mw) of 7.5 can be obtained
from the figure.

Note that Figure 7.4 is for ‘clean sand’ (fines content < 5%). Research has shown that there is an apparent
increase in the CRR with an increase in the fines content. Whether this increase is caused by an increase in
liquefaction resistance or a decrease in penetration resistance is not clear. The effect of the fines content on
the CRR is considered by adjusting (N1)60 to an equivalent clean sand value (N1)60,cs; many researchers
have proposed various ways to apply this correction. For example, Youd et al. (2001) recommended the
corrections

ðN1 Þ60cs = a + b ðN1 Þ60 (7:4a)

8
>
> 0 for Fc ≤ 5%
<
a = exp 1:76 − 190/Fc2 for 5% < Fc < 35% (7:4b)
>
>
:
5 for Fc ≤ 35%

181

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 7.4 Curves relating stress ratios causing liquefaction in clean sands and the (N1)60 value for Mw = 7.5 earthquakes
(Idriss and Boulanger, 2008; reproduced with permission from the publisher, Earthquake Engineering Research Institute)

0.6
Curves derived by 3
1 Seed (1979) 5 4
2 Seed and Idriss (1982)
0.5
3 Seed et al. (1984) and NCEER (1997)
4 Cetin et al. (2004) 2 1
5 Idriss and Boulanger (2004)
0.4
CSR

0.3

0.2

Fines content ≤5%

0.1 Liquefaction
Marginal liquefaction
No liquefaction
0
0 10 20 30 40
Corrected SPT, (N1)60

8
>
> 1:0 for Fc ≤ 5%
<
b= 0:99 + Fc1:5 / 1000 for 5% < Fc < 35% (7:4c)
>
>
:
1:2 for Fc ≤ 35%

where Fc is the ﬁnes content.

7.3.2.2 Evaluation of the CRR using a CPT-based procedure

The CPT is conducted to obtain the cone resistance, the side friction and, if there is a piezocone, the pore
pressure in the deposit. A primary advantage of the CPT is that a nearly continuous profile of penetration
resistance is developed for stratigraphic interpretation. The continuous profile also allows a more detailed
definition of soil layers. In the test, a 60° cone with a face area of 10 cm2 and a 150 cm2 ‘friction sleeve’ is
hydraulically pushed into the ground at a constant speed (ranging from 1.5 to 2.5 cm/s). The force required to
maintain this penetration rate, and the shear force acting on the friction sleeve, are recorded.

The liquefaction potential can also be evaluated using the results of CPTs. In fact, procedures that use CPT
resistance are very similar to those using SPT resistance. Based on in situ investigation, a boundary can be
defined separating liquefiable from non-liquefiable conditions, as shown in Figure 7.5. In the graph, qc1N is
the cone tip resistance qc normalised by the overburden pressure s0v0 . So
n
qc1 qc Pa
qc1N = = 0 (7:5)
Pa Pa sv0

182

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

Figure 7.5 Curves relating stress ratios causing liquefaction in clean sands and the qc1N value for Mw = 7.5 earthquakes
(Idriss and Boulanger, 2008; reproduced with permission from the publisher, Earthquake Engineering Research Institute)

0.6
Shibata and Teparaksa (1988)
Robertson and Wride (1997)
Suzuki et al. (1997)
0.5
Moss et al. (2006) – 5% probability
Idriss and Boulanger (2004)

0.4
CSR

0.3

0.2

0.1
Clean sands
Liquefaction
No liquefaction
0.0
0 50 100 150 200 250
Normalised corrected CPT tip resistance, qc1N

where Pa and n are similar to the values discussed above for correcting the SPT N value. Thus, once the cone
penetration resistance qc for a deposit is known, the liquefaction resistance can be estimated from the chart
and, consequently, the liquefaction potential can be evaluated.

7.3.2.3 Evaluation of the CRR using a Vs-based procedure

Shear wave velocity (Vs) tests measure the small-strain shear modulus (stiffness) of the soil, and thus they
represent an engineering property measurement rather than an index test. There are several different methods
for measuring Vs in situ, such as cross-hole, downhole, seismic CPT, seismic dilatometer and spectral analysis
of surface waves (SASW). A general advantage of Vs tests is that they can be used for sites underlain by soils
that are difﬁcult to penetrate (gravel, cobble, etc.); on the other hand, they provide very limited spatial res-
olution for characterising site stratigraphy and heterogeneity, and sampling is not possible.

The shear wave velocity of the deposit can also be used to evaluate its liquefaction potential. Essentially, the
procedure is similar to other penetration-based approaches, and the boundary curves separating liquefaction
and non-liquefaction obtained by various researchers are shown in Figure 7.6. In the plot, the shear wave
velocity is normalised using
0:25
Pa
Vs1 = Vs 0 (7:6)
sv0

where Vs1 is the overburden stress-corrected shear wave velocity, Pa is the atmospheric pressure (100 kPa)
and s0v0 is the initial vertical stress in the same units as Pa. The above equation assumes a constant coefﬁcient
of earth pressure of K0 = 0.5.

183

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 7.6 Curves relating stress ratios causing liquefaction in clean sands and Vs1 value for Mw = 7.5 earthquakes
(Youd et al. 2001; reproduced with permission from the publisher, ASCE)

0.6
*Curve adjusted using scaling Mw = 7.5
factor of 1.19 for Mw 7 earthquakes
**Approximate curve for clean
sand and 15 cycles of loading,
assuming emin = 0.65, Lodge
K0 = 0.5, r0 = 0.9 (1994)*

0.4
Best fit
Tokimatsu and
Uchida (1990)**
Lower
CRR

bound
Andrus et al. Andrus and
Liquefaction (1999) Stokoe
(1997)
0.2
Kayen et al.
Robertson et al. (1992)*
(1992)

No
liquefaction

0.0
0 100 200 300
Overburden stress-corrected shear wave
velocity, VS1: m/s

7.3.2.4 Evaluation of the CRR using laboratory element testing

In Japan, the results of cyclic shear tests on high-quality undisturbed samples from in situ sand deposits are
used to determine the resistance of soil to liquefaction. The use of undisturbed soil samples is recommended
because age, soil fabric and stress history play very important roles in deﬁning the liquefaction strength of
soils; however, obtaining undisturbed samples of loose granular materials is very challenging because the soil
structure must not be disturbed during sampling. The best method is to freeze the ground using liquid nitrogen
before carving out large-diameter samples by rotary coring (Hatanaka et al., 1985; Sego et al., 1999; Yoshimi
et al., 1978). More recently, advanced sampling techniques have been developed, such as the ‘gel-push’
sampling method (Lee and Chen, 2013; Mori and Sakai, 2016). As the name suggests, this involves the
injection of a water-soluble polymeric lubricant (gel) from the sampler shoe to lubricate and reduce friction
between the cut sample and the tube, both during sampling as well as extrusion in the laboratory. As with
regular piston samplers, the sampler is lowered down a cased, mud-ﬁlled borehole before being hydraulically
pushed into undisturbed soil.

Once the undisturbed samples are obtained, liquefaction tests are performed in the laboratory, as outlined in
Section 6.4.4. Once the liquefaction resistance curve is obtained, the CSR corresponding to 15–20 cycles is
read from the curve, and this is taken as the CRR. Note that an Mw 7.5 earthquake is usually considered to
induce 15–20 equivalent cycles.

184

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

However, considering that the condition of an element of soil in the field and the soil specimen in the lab-
oratory are under different conditions (in terms of the state of stress and earthquake loading), many
researchers have suggested that the CRR obtained in the laboratory should be corrected to represent the CRR
in the field, (CRR)field. For example, Towhata (2008) proposed the following for corrections to the CRR
obtained using the triaxial apparatus, (CRR)triaxial

(CRR)field = C1 C2 C3 C4 C5 (CRR)triaxial (7:7)

where C1 is the correction due to the difference in the consolidation stress (C1 = (1 + 2K0)/3, where K0 = 0.5
for normally consolidated soils); C2 is the correction due to the difference in the loading conditions, where
earthquake loading is irregular while laboratory specimens are subjected to sinusoidal waves (C2 = 1/0.65 or
1/(0.55–0.70); C3 is the correction due to sample disturbance (C3 > 1, but this is not clearly understood yet);
C4 is the correction due to densiﬁcation during handling (C4 < 1, but this is not clearly understood yet);
and C5 is the correction due to the loading direction, where earthquake loading is at least two components,
east–west and north–south (C5 = 0.80–0.90).

In many design codes in Japan, empirical charts, similar to those depicted in Figure 7.4 are used. However,
the curves in those charts were not established from case histories but rather based on correlation between the
CRR obtained from high-quality soil samples and the penetration resistance obtained at nearby sites (e.g.
Ishihara, 1993; 1996; JRA, 2002).

7.3.2.5 Correction factors for magnitude and overburden pressure

The empirical charts discussed above are expressed in terms of CRR7.5 and 1 atm (i.e. the CRR of the soil
adjusted to 1 atm of effective overburden pressure for an Mw 7.5 earthquake). Note that the various rela-
tionships use different factors to account for earthquake magnitude and overburden stresses. It is common for
CRR relationships to be normalised for this base case, and then site-speciﬁc adjustments are made via
magnitude scaling factors (MSFs) and overburden correction factors (Ks) to account for the earthquake
magnitude under consideration and the overburden stresses at the depth of interest. The site-adjusted CRR
(i.e. for a given earthquake magnitude M and overburden pressure s 0v0 ) is obtained from CRR7.5,1 atm
according to

CRRM,sv0 = CRR7:5,1 atm MSF Ks (7:8)

The adjustment for magnitude is necessary because of the strong correlation between the CRR and the
number of loading cycles that the earthquake imparts to the soil. Many researchers proposed expressions for
the MSF as a function of the magnitude Mw; for example, Youd et al. (2001) recommended

MSF = 102:24 /Mw 2:56 (7:9)

Other commonly used expressions were proposed by Cetin et al. (2004) and Idriss and Boulanger (2008).
More recently, Boulanger and Idriss (2014) recommended a revised expression for the MSF that takes into
account the soil type and the denseness of the soil.

The overburden correction factor (Ks) is used to extrapolate liquefaction-triggering relationships to

high overburden pressures. Typical expressions for Ks were proposed by Seed and Harder (1990),

185

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Hynes and Olsen (1999) and Boulanger (2003). For example, based on previous research, Youd et al.
(2001) recommended
0 f −1
sv0
Ks = ≤ 1:0 (7:10)
pa

For this equation, pa = 100 kPa and f = 0.6 for Dr = 80%, f = 0.7 for Dr = 60% and f = 0.8 for Dr = 40%.

It should be mentioned that the various liquefaction-triggering procedures cannot be compared solely on CRR
plots. For example, even if the CRR relationships appear to be very similar, one would not necessarily obtain
the same CRR value for given raw penetration data (N value or tip resistance) because the factors used to
correct the raw data to the penetration index (say the normalised clean sand N value (N1)60,cs or the normalised
clean sand tip resistance qc1,cs) can vary signiﬁcantly between relationships.

7.3.3 Factor of safety against liquefaction

Once the CSR and CRR are known, the liquefaction potential of the soil can be estimated. It is usually
expressed in terms of the factor of safety against liquefaction, FL, and is given by
CRR
FL = (7:11)
CSR
If FL < 1.0, the shear stress induced by the earthquake exceeds the liquefaction resistance of the soil and,
therefore, liquefaction will occur. Otherwise, when FL > 1.0, liquefaction will not occur.

The ﬂowchart of the whole empirical procedure is shown in Figure 7.7 for the case of the SPT-based
approach; a similar approach is also applicable for the other penetration-based methods.

Figure 7.7 Typical ﬂowchart for evaluating the soil liquefaction potential based on the SPT N value

Peak ground Earthquake SPT Effective

Fines content,
acceleration, amax magnitude, N value stress σ ′v
Fc (%)
M
(N1)60 = CN(σ ′v )N
Magnitude scaling
factor, MSF
(CRR)M = 7.5 = (τcyc /σ ′v 0 )M = 7.5

Cyclic shear stress ratio Cyclic resistance ratio

τ cyc a σ
CSR = = 0.65 max v0 rd CRR = (CRR)M = 7.5 × MSF × Kσ
σ ′v0 g σ ′v0

Factor of safety
CRR
FL =
CSR

186

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

Figure 7.8 Calculation of the soil liquefaction potential of a ground proﬁle

SPT N value FL
0
Sand

Liquefiable
5 Sand

10
Sandy
silt

15 Gravel

Thus, for a given profile, the FL value is calculated at several depths within the profile, typically at locations
where the penetration resistance (N value or qc tip resistance) or Vs is known. This is depicted in Figure 7.8,
where layers with FL < 1 are deemed liquefiable for the specified amax and Mw. Note that if the penetration
resistance is obtained in a continuous manner, as in the CPT, then a continuous profile of the factor of safety
against liquefaction can be obtained.

7.3.4 Liquefaction evaluation through effective stress-based ground response

analyses
A more detailed investigation of the response of soils to cyclic loading can be performed using effective stress
analyses. These involve seismic response and liquefaction process modelling using fully coupled equations
for the granular solid and the pore fluid to simulate the response of soils (as obtained from laboratory testing)
in terms of the generation and dissipation of excess pore water pressure and the associated development of
ground deformation. These are typically implemented through the finite-element or finite-difference method.
Some of the more popular computer programs used are FLAC (Itasca, 1998), PLAXIS (Plaxis, 2011) and
DIANA (DIANA Analysis, 2001). More recently, many researchers have been using the Open System for
Earthquake Engineering Simulation (OpenSees) (Mazzoni et al., 2011) as a ground response analysis tool.
OpenSees is a fully non-linear object-orientated code that provides a platform for the modelling of geo-
technical and structural systems under static and dynamic loading. Further details of the role of numerical
modelling in liquefaction potential evaluation were discussed in Section 6.5.

7.4. Evaluation of the effects of liquefaction

While the FL concept discussed above could be used to estimate the liquefaction potential at a speciﬁed depth
within a soil proﬁle, the FL value alone at the indicated depth cannot be used to evaluate the possible impact of
liquefaction at the target site. For this purpose, other indices are used to indicate the possible severity of
damage that liquefaction of the subsoil layers can induce at a particular site.

187

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 7.9 Proposed boundary curves for the surface manifestation of liquefaction-induced damage
(Ishihara, 1985)

12
Non-liquefiable H1
Thickness of liquefiable sand layer, H2: m
11 surface layer Sand H1
Liquefiable sand
10 H2 Sand (N £ 10) H2
layer

9
Peak acceleration
8 ~0.2g Peak acceleration
~0.4–0.5g
7 Peak
Liquefaction-induced

acceleration Unliquefiable soil H1

ground damage

6 ~0.3g
Sand (N £ 10) H2
5
4

3
Unliquefiable soil H1
2
1 Sand (N £ 10) H2
0
0 1 2 3 4 5 6 7 8 9 10
Thickness of surface layer, H1: m

7.4.1 Damage in the presence of an unliqueﬁable surface layer or crust

Depending on the relation between the thickness of the liquefiable layer H2 and the thickness of the
surface crust H1, it can be estimated whether or not liquefaction will exert damage on the ground surface
using Figure 7.9. The definitions of H1 and H2 are indicated in the figure. This chart was developed by
considering various case histories of recent large earthquakes where liquefaction-induced ground damage
has been observed.

7.4.2 Liquefaction potential index

The severity of liquefaction at a site can be quantiﬁed in terms of the liquefaction potential index (LPI), given
by (Iwasaki et al. 1982)
ð 20
LPI = F(z) w(z) dz (7:12)
0

where z is the depth below the ground surface (m) and F(z) is a function of the liquefaction resistance
factor FL
(
1 − FL FL < 1:0
F(z) = (7:13)
0 FL ≤ 1:0

and w(z) = 10 – 0.5z. Equation 7.12 gives values of LPI ranging from 0 to 100. Sites with LPI > 15
would result in severe damage, while if LPI < 5, the effects are minor. Using this criterion, LPI has
been used to assess liquefaction hazards worldwide.

188

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

7.4.3 Ishihara-inspired LPI

Many researchers evaluating the LPI at various sites during recent earthquakes have found its performance
to be inconsistent, ranging from generally promising (Toprak and Holzer, 2003) to largely erroneous (Lee
et al., 2004). Using the observed ground performance following the 2010–2011 Canterbury earthquake
sequence (New Zealand), Orense et al. (2014) and Lees et al. (2015) showed that while LPI performance
was generally good, the liquefaction severity was significantly over-predicted for a portion of the study
area. Given the inconsistent efficacy of the existing LPI framework and the criterion for assessing risk due
to liquefaction, Maurer et al. (2015) modified the LPI framework using the H1 – H2 concept proposed by
Ishihara (1985); that is, they assumed that each boundary curve in Figure 7.9 may be approximated by two
straight lines having slopes m and ∞ and that FL for liquefied strata is assumed to be uniform with depth
(i.e. F in Equation 7.13 is constant while assuming a power-law depth weighting function w′. The revised
index is termed LPIISH, and

ð 20
25:56
LPIISH = F(FL ) dz (7:14)
H1 z

where
(
1 − FL FL ≤ 1:0 ∩ H1 m(FL ) ≤ 3
F(FL ) = (7:15a)
0 otherwise

5
m(FL ) = exp −1 (7:15b)
25:56(1 − FL )

7.4.4 Liquefaction severity number

The liquefaction severity number (LSN) is a parameter developed by Tonkin & Taylor (2013) to reﬂect the
more damaging effects of shallow liquefaction on residential land and foundations. The equation used to
calculate LSN is
ð
ev
LSN = 1000 dz (7:16)
z

where ev is the calculated volumetric densiﬁcation strain in the subject layer from Zhang et al. (2002) and z is
the depth to the layer of interest in metres below the ground surface.

The depth-weighting function recognises that ground surface damage from shallow liquefied layers is more
likely than from deeper layers. The LSN considers depth-weighted calculated volumetric densification strain
within soil layers as a proxy for the severity of liquefaction land damage likely at the ground surface. The
published strain calculation techniques consider strains that occur where materials have a calculated trig-
gering FL that reduces below 2.0. This means that the LSN begins to increase smoothly as factors of safety
drop, rather than when FL reaches 1.0. One other aspect of LSN to note is that strains self-limit based on the
initial relative density as the factor of safety drops, so a given soil profile has a maximum LSN that it tends
towards as the peak ground acceleration increases. Table 7.4 summarises the expected performance of the
ground for each LSN range, based on data observed following the 2010–2011 Canterbury earthquake
sequence (New Zealand).

189

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Table 7.4 Predominant performance of sites for each LSN range

LSN range Predominant performance

0–10 Little to no expression of liquefaction; minor effects

10–20 Minor expression of liquefaction; some sand boils

20–30 Moderate expression of liquefaction; with sand boils and some structural damage

30–40 Moderate to severe expression of liquefaction; settlement can cause structural damage

40–50 Major expression of liquefaction; undulation and damage to ground surface; severe total
and differential settlement of structures

>50 Severe damage; extensive evidence of liquefaction at surface; severe total and differential
settlements affecting structures; damage to services

Tonkin & Taylor (2013)

7.5. Codes of practice-speciﬁed methods

While the methods discussed above are popularly used in geotechnical practice, there are some design codes
where the method(s) to be employed is explicitly speciﬁed. These include the design provisions in Eurocode 8
as well as those speciﬁed in various codes of practice in Japan.

7.5.1 Eurocode approach

The steps involved in this approach for the determination of the liquefaction potential using SPT data are
given below.

Step 1. The seismic shear stress (te) is estimated from the following relation

te = 0:65aSsv (7:17)

in which sv is the total overburden pressure at the depth under consideration, a is the ratio of the design ground
acceleration on type A ground (rock) to the acceleration due to gravity, and S is the soil parameter speciﬁed in
clause 3.2.2.2 of Part 1.1 of Eurocode 8 (CEN, 2004a). To determine the value of S, the ground type is
determined based on SPT data using Table 3.1 in Eurocode 8. Note that the S value is determined from Table 3.2
in the code, based on the ground type. The above expression for te is applicable up to a depth of 20 m.

Step 2. The SPT data are normalised to a reference effective overburden pressure of 100 kPa and to a 60% ratio of
impact energy over the theoretical free fall energy. For a depth of less than 3 m, the measured SPT value is reduced
by 25%. To normalise with respect to the effective overburden pressure, the measured SPT data are multiplied by a
factor of (100/ s 0v )0:5 , where s 0v (kPa) is the effective overburden pressure acting at a depth where the SPT
measurement is done. Energy normalisation is carried out by multiplying the measured SPT raw data by the factor
ER/60, where ER is the measured energy ratio (%); normalised SPT N values are referred to as N1(60).

Step 3. The CSR (t e /s 0v0 ) required to induce liquefaction is determined for a particular N1(60) for an earth-
quake of Mw 7.5 using Figure B4 in Eurocode 8 Part 5. To apply this ﬁgure to earthquake magnitudes different
from 7.5, the ratio (t e /s 0v0 ) is multiplied by the factor CM in Table B.1 of Part 5 of the code (CEN, 2004b).

190

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

Step 4. A soil layer is considered to be susceptible to liquefaction whenever the earthquake-induced shear
stress (from step 1) exceeds 80% of the critical stress (from step 3) known to have induced liquefaction in a
previous earthquake.

7.5.2 Japanese code approach

In Japan, various seismic codes and standards have been established for specific structures. These codes
specify methods of evaluating the liquefaction potential and consideration of the effects of liquefiable ground
in the design of these structures. Some of these structure-specific codes of practice are listed in Table 7.5,
together with the organisations involved.

Most of these standards use FL-based methods similar in concept to those discussed above. However, they
differ in many details regarding susceptibility criteria, CSR and CRR calculations. Some of the important
features of these codes (Japanese Geotechnical Society, 1998) are that

■ all codes require the SPT blow count and some measure for the grain-size distribution
■ the grain-size distribution is characterised by the mean grain diameter D50 and the ﬁnes content
(only two codes use the PI, and one code uses the clay content Pc for susceptibility)
■ SPT-based evaluation is prevalent (only the ‘Technical standards and commentaries for port and
harbour facilities in Japan’ (OCDI, 2009) have CPT-based criteria)
■ in some codes, the magnitude of the earthquake is directly considered when calculating the CSR;
in others, it is implicitly assumed through seismic coefﬁcients
■ only two codes require seismic response analysis (nuclear power plant and port and harbour facilities)
■ the depth of interest in liquefaction evaluation is 20 m or, in some, 15 m.

Table 7.5 Japanese design codes and standards that introduce liquefaction assessment procedures

Structures Organisation

Port and harbour facilities Japan Port and Harbour Association (JPHA)

Highway bridges Japan Road Association (JRA)

Railway structures Japan National Railway (JNR)

Building foundation structures Architectural Institute of Japan (AIJ)

Control of hazardous materials Ministry of Home Affairs (MHA); Fire Defense Agency (FDA)

Liqueﬁed natural gas in-ground storage Japan Gas Association (JGA)

Waterworks facilities Japan Water Works Association (JWWA)

Sewage facilities Japan Sewage Works Association (JSWA)

Common utility ducts Japan Road Association (JRA)

Nuclear power plants Japan Electric Association (JEA)

Japanese Geotechnical Society (1998)

191

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Many of these codes adopt historical features, and the relevant organisations make adjustments based on
lessons learned from recent earthquakes.

As an example, the method outlined in the ‘Speciﬁcations for highway bridges’ (JRA, 2002) is illustrated
here. The code advises assessment of the liquefaction potential for the following soils

■ saturated alluvial sand when the sand lies within 20 m from the ground level (GL) and the water table
is within 20 m from the GL
■ the ﬁnes content (deﬁned by less than 75 μm in size, i.e. silt and clay) is less than 35%
■ the PI is less than 15% (i.e. the apparent cohesion is small)
■ the mean grain size D50 < 10 mm while D10 > 1 mm.

In the method, the resistance of soil against liquefaction R is calculated together with the dynamic load L
induced in the soil element by the seismic motion. The factor of safety FL is then calculated based on Equation
7.11, with R being equivalent to the CRR while L is similar to the CSR.

Prior to the 1995 Hyogoken Nambu earthquake, the resistance of soil to liquefaction was calculated as the sum of
three factors that take into account the overburden pressure, grain size and ﬁnes content. However, the code has been
revised in the aftermath of the 1995 earthquake to take into account some underestimation of the cyclic strength.

The resistance against liquefaction R is calculated by

R = CW RL (7:18)

in which RL is the resistance as would be observed in cyclic undrained triaxial tests and CW represents the
effects of the type of loading. Although the triaxial strength RL should be determined by performing labo-
ratory tests on undisturbed specimens, it can be estimated through correlations with the SPT N value
( pffiffiffiffiffiffiffiffiffiffiffiffiffi
0:0882 Na / 1:7 Na < 14
RL = pffiffiffiffiffiffiffiffiffiffiffiffiffi −6
(7:19)
0:0882 Na / 1:7 + 1:6 10 (Na − 14) Na ≥ 14
4:5

where
(
c1 N1 + c2 sandy soil
Na = D50 (7:20a)
1 − 0:36 log10 2 N1 gravelly soil
8
>
>1 (0% ≤ Fc < 10%)
<
c1 = (Fc + 40)/ 50 (10% ≤ Fc < 60%) (7:20b)
>
>
:
(Fc / 20) − 1 (Fc ≤ 60%)

(
0 0% ≤ Fc < 10%
c2 = (7:20c)
(Fc − 10)/ 18 Fc ≤ 10%

192

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

In the above equations, s0v is the effective overburden stress, Na is the adjusted SPT N value, N1 is the
corrected SPT N value for s0v = 1 kgf/cm2 (N1 = 1.7N/(s0v + 0.7)), D50 is the mean grain size (mm) and Fc is
the ﬁnes content (%).

The factor CW considers the effect of the type of loading. Two types of earthquake motions are considered:
type I ground motion, which is caused by a plate boundary type earthquake (expected Mw > 7.5 and longer
duration expected, i.e. the number of loading cycles is higher), and type II ground motion caused by a very
rare near-ﬁeld earthquake (a large-amplitude earthquake with fewer cycles and short epicentral distance; may
create greater acceleration in a limited area). The factor CW is given by

type I ground motion CW = 1:0 (7:21a)

8
>
>1:0 (RL ≤ 0:1)
<
type II ground motion CW = 3:33RL + 0:67 (0:1 < RL ≤ 0:4) (7:21b)
>
>
:
2:0 (RL > 0:4)

On the other hand, the shear stress ratio L induced by the earthquake is calculated by

sv
L = rd khc (7:22)
sv0

where khc is the horizontal seismic coefﬁcient at the ground surface, sv and s0v are the total and effective
overburden stress, respectively, and rd is the reduction factor of shear stress in the vertical direction, which is
speciﬁed as

rd = 1:0 − 0:015z (7:23)

where z is the depth of interest (m). The seismic coefﬁcient khc is calculated by

khc = cz khc0 (7:24)

where cz is the seismic zone factor, which represents the regional seismicity of the target site and ranges
between 0.7 and 1. The standard seismic coefﬁcient khc0 is dependent on the subsoil class and ground motion
types. To deﬁne the subsoil class, the predominant or natural period of the ground TG is estimated using

X
n
4Hi
TG = (7:25)
1
Vs,i

in which H (m) and Vs (m/s) are the thickness and the shear wave velocity, respectively, for layer i, and
n indicates the number of layers from the ground surface to the bedrock. If Vs is not available, then it is
determined from the following empirical correlation with the SPT N value
(
80N 1/ 3 for sand
Vs = (7:26)
100N 1/ 3 for clay
Based on the value of TG, the subsoil is classiﬁed, and the value of khc0 is as shown in Table 7.6.

193

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Table 7.6 Standard seismic coefﬁcient khc0

Subsoil class TG Design earthquake

Type I Type II

Class I (hard) TG < 0.2 s 0.30 0.80

Class II 0.2 s < TG < 0.6 s 0.35 0.70

Class III (soft) TG > 0.6 s 0.40 0.60

JRA (2002)

7.6. Liquefaction-induced ground displacements

As a result of excess pore water pressure increase and loss of effective stress, a liqueﬁed soil deposit can
undergo various types of deformation (i.e. lateral displacements and vertical settlements). The development
of reliable and realistic techniques to evaluate these liquefaction-induced ground deformation phenomena
requires a clear understanding of the actual mechanisms responsible for these deformations.

Cyclic and permanent lateral displacements of liquefied soil are among the most troublesome liquefaction
hazards for buried structures and lifeline facilities. Because these types of deformation involve the movement
of competent soil, full passive pressure can be mobilised against an underground structure. Pipelines, bridges,
pile foundations and other civil engineering structures on or within the extensional zone are typically cracked
or torn apart. Almost all the alarming photographs of widespread destruction from an earthquake are damage
done by large permanent lateral movements of the soil surface. Vertical settlement, on the other hand, is due
to reconsolidation of the sand mass as the grains settle out of the liquefied state. Initially, there is no settlement
if undrained conditions are assumed, but as the excess pore water pressure dissipates, the surface will settle by
an amount roughly equal to the ejected water. Surface settlement is also associated with the appearance of
sand boils. Note that excess pore water pressure need not be equal to the effective confining pressure to
produce significant settlement. In addition, the sand deposit need not undergo a large amount of densification,
as pointed out by the evidence of subsequent liquefaction of historically liquefied sands.

7.6.1 Estimating cyclic ground displacements

As mentioned in Section 6.7.3, liquefaction at depth may decouple overlying soil layers from the underlying
ground and allow the upper soil to oscillate back and forth and up and down in the form of ground waves. This
deformation typically occurs in flat ground or on gentle slopes where lateral permanent displacement would
not occur. These oscillations are usually accompanied by opening and closing of fissures and fracture of rigid
structures such as pavements and pipelines. Such back and forth movement of the liquefied soil, with the
surface crust riding on top, is also a major hazard to pile foundations; hence, estimation of the magnitude of
the cyclic component is necessary for the efficient design of underground structures.

However, very few studies have been done to date to estimate the magnitude of cyclic ground displacements
during liquefaction. The only available approach is based on the studies of Tokimatsu and Asaka (1998)
following the 1995 Kobe earthquake (Japan). In this method, the peak cyclic displacements are expressed in
terms of simpliﬁed charts correlating the maximum cyclic shear strain that develop in the liqueﬁed layer with

194

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

the CSR and SPT blow count, as shown in Figure 7.10. In the figure, Na is the normalised SPT N value
adjusted for the effect of the fines content (i.e. see Equation 7.20a; this is the equivalent SPT N value for clean
sand, (N1)60,cs). In order to determine the peak cyclic deformation, the adjusted SPT N values (Na) and
equivalent CSRs during the earthquake (t av / s 0v0 ) at various depths are evaluated, from which the cyclic shear
strain gcy is estimated for each depth. Then, the cyclic ground displacement profile fcy(z) is determined by
integrating gcy upwards from the bottom of the liquefied layer, assuming gcy develops in the same direction.
A schematic diagram showing this procedure is shown in Figure 7.11.

Figure 7.10 Relation between the maximum CSR and the SPT N value (Tokimatsu and Asaka, 1998; reprinted with
permission from the author)

0.6

γcy = 4% 2% 1% 0.5%
0.5

0.4
Stress ratio, τav /σv0′

0.3

0.2

0.1

0
0 10 20 30 40
Adjusted SPT N value, Na

Figure 7.11 Diagram illustrating the determination of cyclic ground deformation during an earthquake

Cyclic strain, Ground displacement

Na value γcy fcy
0
Sand

5
Sand

10
Sandy
silt

15 Gravel

195

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

7.6.2 Estimating permanent lateral displacements

The evaluation of permanent lateral ground displacements is a difficult non-linear problem. However, it is
important, as the evaluation of the damaging effects of liquefaction at a site requires the ability to predict the
magnitude and spatial distribution of lateral ground displacements. It is worth mentioning that even though
the prediction if a site will liquefy or not can be done with a reasonable degree of confidence, there is no
generally accepted method for evaluating lateral displacements. Current methods can be generally classified
into the following categories.

7.6.2.1 Empirical approach

The most straightforward approach to estimating horizontal displacements due to liquefaction is to compile a
database of known displacements and to correlate those displacements with measurable parameters. Note that
choosing the parameters that should be considered is a formidable task, since there are many geotechnical,
topographical and seismological factors affecting the results. Moreover, the accurate measurement and
compilation of in situ measurements, and the relevant parameters, is also difﬁcult.

Based on a detailed mapping of liquefaction-induced ground displacements in Noshiro City after the 1983
Nihonkai-Chubu earthquake (Japan), combined with information from the 1964 Niigata earthquake (Japan)
and the 1971 San Fernando earthquake (USA), Hamada et al. (1986) carried out a regression analysis. The
result of the regression is given by
p
2
ffiffiffiffi p
3
ffiffiffi
D = 0:75 H q (7:27)

where D is the displacement (m), H is the thickness of the liquefied layer (m), and q is the greater of the
ground surface slope and the slope of the bottom boundary of the liquefied layer (%). The database used to
derive the above equation, however, is biased by several factors. There is strong influence from Noshiro City,
which had milder slopes and smaller displacements, and the large displacements in Niigata. The estimates are
unreliable at areas of rapid change, such as near river banks. The earthquakes considered were very close in
magnitude and the soils were very similar; hence, Equation 7.27 does not represent general conditions.

A multiple linear regression analysis was conducted by Bartlett and Youd (1995) to take into account a great
number of variables in the predictive equation. They analysed 43 detailed factors from eight different
earthquakes to account for seismological, geological, topographical and geotechnical effects on lateral ground
displacements due to liquefaction. Youd et al. (2002) later corrected and updated the original analysis, and
two equations for lateral ground displacement DH (m) – one for movement down a gentle slope and another
towards a free face – were derived

■ for sloping ground

log DH = −16:213 + 1:532M − 1:406log R* − 0:012R + 0:338log S

(7:28a)
+ 0:540 log T15 + 3:413 log(100 − F15 ) − 0:795log (D5015 + 0:1)

■ for a free face

log DH = −16:713 + 1:532M − 1:406 log R* − 0:012R + 0:592 log W

(7:28b)
+ 0:540 log T15 + 3:413 log (100 − F15 ) − 0:795 log (D5015 + 0:1)

196

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

Figure 7.12 Relation between measured and predicted displacements for values up to 2 m (Youd et al., 2002;
reproduced with permission from the publisher, ASCE)

2.0
Measured = 2 × predicted
1.8

1.6 Measured = predicted

1.4
Measured displacement, DH: m

1.2

1.0

0.8 Measured = 0.5 × predicted

0.6
Japan data
0.4 US data
× × Whiskey Springs data
0.2 × Moss Landing data
× Kobe earthquake data
0
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0
Predicted displacement, DH: m

where M is the earthquake moment magnitude, R is the horizontal distance from the site in question to the
nearest bound of seismic energy source (km), R* is a modiﬁed source distance (R* = R + 100.89M – 5.64), W is
the ratio of the free-face height to the distance to the free face (%), S is the ground slope (%), T15 is the
cumulative thickness (m) of saturated sandy layers with a normalised SPT N value (N1)60 < 15, F15 is the
average ﬁnes content of saturated granular layers within T15 (%) and D5015 is the average mean grain size
(mm) of layers included in T15. Figure 7.12 shows that the above formulas give results comparable to those of
observed displacements within a factor of ±2. Note again that there are limitations to the use of these for-
mulas, based on the data set used to derive the correlations.

7.6.2.2 Semi-empirical approach

Another approach is to estimate the permanent strains that are expected within the liqueﬁed zones and then
integrate those strains over the depth to obtain an estimate of the potential lateral displacement at the ground
surface. Then, the laterally displacement is estimated by calibrating the potential displacement with case
history observations.

Thus, the ﬁrst step in this method is integration of shear strain proﬁles estimated in conjunction with SPT- and
CPT-based liquefaction analyses. The maximum potential shear strains may be estimated using existing
relationships, such as the three SPT-based correlations compared in Figure 7.13. The computed ground
surface displacement is known as the lateral displacement index (LDI). Based on calibration with case
histories, two equations have been proposed for the lateral displacement LD (Zhang et al., 2004)

■ for sloping ground

LD = ðS + 0:2ÞLDI for 0:2 % < S < 3:5 % (7:29a)

■ for a free-face

LD = 6ðL /H Þ−0:8 LDI for 4 < L /H < 40 (7:29b)

197

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 7.13 Comparison of the relationship between the CSR, SPT (N1)60,cs and the maximum shear strain for three
levels of maximum shear strain (Idriss and Boulanger, 2008; reproduced with permission from the publisher,
Earthquake Engineering Research Institute)

0.6

50% 20% 5%
0.5 Maximum
shear
strains,
γmax
0.4
CSR

0.3

0.2
γmax–FSliq by Ishihara and Yoshimine
(1992) with CRR-(N1)60CS by
0.1 Idriss and Boulanger (2006)
Tokimatsu and Asaka (1998)
Wu (2002)
0
0 10 20 30 40
Corrected SPT, (N1)60cs

where H is the free-face height, L is the distance of the target point to the free face and S is the ground
slope (%). This method does not account for two- or three-dimensional effects; therefore, the results for
individual borings can be misleading on their own. A beneﬁt of the integration of the strain method is that an
estimate of the soil displacement proﬁle is obtained over the depth of the foundation ground, which can in turn
be used as an analysis input.

7.6.2.3 Direct computation

The direct computation of seismically induced ground displacements has been a topic of interest in recent
years. Numerous analytical and numerical techniques have been developed, most of which are based on limit
equilibrium; that is, if the driving stress is greater than the available strength, the structure becomes unstable
and some form of displacement will occur until enough resistance is mobilised to counteract the driving
stresses.

One of the earliest and probably the most widely used type of analysis is that proposed by Newmark
(1965), where the ground displacement is calculated based on a single-degree-of-freedom system and rigid
plastic soil, and the earthquake load is represented by a pseudo-static load. When the earthquake-induced
acceleration is less than a certain critical value, no sliding will occur; however, when the critical value is
exceeded, permanent ground displacement will take place. The displacement is then derived by integrating
the acceleration–time history in excess of the critical value. Although the procedure is quite simple, the
determination of the critical value is quite complicated, and the calculated displacements depend on the
accuracy of this critical acceleration.

Note that this method is based on the rigid-body movement of soil. When considering liqueﬁed soil, however,
the absence of rigid soil makes the application of the model difﬁcult. Moreover, in its present form, the model

198

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

cannot account for the large strains and displacements that occur within the zone of liquefaction. Never-
theless, this type of analysis has been adopted by several researchers. For example, Baziar et al. (1992) used
the model in conjunction with monotonic and cyclic laboratory tests and employed regional attenuation
relations for peak ground acceleration and velocity to evaluate in situ ground deformations. Byrne (1990)
proposed incorporating a more-thorough description of post-liquefaction behaviour of sand by using a non-
linear spring to represent the stiffness of the liqueﬁed layer, with its residual strength incorporated.

7.6.2.4 Finite-element-method approach

Many effective stress-based ground response analyses are used to numerically determine the magnitude of
cyclic and permanent ground displacements induced by liquefaction through the finite-element method or the
finite-difference method. Such numerical models can address many important aspects of liquefaction-induced
ground deformations, particularly the influence of irregular geometry. Many of the presently available
techniques for evaluating such ground displacements employ non-linear dynamic stress–strain behaviour of
soil in conjunction with plasticity theory.

However, even with the abundance of constitutive models currently available in the literature, no numerical
method can accurately model the actual physical behaviour of liquefied soil. This is because soil liquefaction
is a very complex phenomenon, and accurate constitutive modelling of liquefiable soil is hampered by the
difficulty in obtaining representative ‘undisturbed’ testing samples from an in situ deposit. As a result, there is
at present no generally agreed upon detailed understanding of what happens during and after liquefaction, and
no method exists to measure the relevant parameters, even if there were agreement. Thus, the obvious
conclusion is that, until a complete constitutive model of soil behaviour is validated for all conceivable
conditions, the only useful techniques currently available for estimating post-liquefaction lateral displace-
ments in standard practice are purely the empirical correlations and semi-empirical approaches.

7.6.3 Estimating vertical settlements

Vertical settlements of the order of up to several tens of centimetres have been observed to occur after major
earthquakes, such as the 1964 Niigata earthquake (Japan), the 1995 Hyogoken Nambu earthquake (Japan),
2010–2011 Canterbury earthquake sequence (New Zealand) and the 2011 Tōhoku earthquake (New
Zealand). Differential settlement, an associated problem, caused extensive damage to lifelines and under-
ground structures during the 1985 Chile earthquake and the 1990 Luzon earthquake (Philippines). Thus, there
is a need to estimate the magnitude of vertical settlement that can occur within a liqueﬁed soil deposit.

Based on data from a variety of laboratory tests, post-liquefaction field data and centrifuge tests, it is estimated
that volumetric strains of the order of 1.5–5.0% can occur in loose sands, while less than 0.2% is expected for
dense sand. Multiplying these values of strains by the thickness of the liquefied layer, it is clear that the
absolute vertical displacements can be sizable. For example, if the thickness of the liquefied layer is 10 m, a
settlement as large as 50 cm can be expected for loose sandy deposits.

7.6.3.1 Ishihara–Yoshimine method

One of the most widely used procedures for estimating vertical settlements of sandy deposits is that developed
by Ishihara and Yoshimine (1992). They derived a set of curves from cyclic simple shear tests on Fujii River
sand relating the volumetric strain ϵv to the relative density Dr and the factor of safety against liquefaction FL,
as shown in Figure 7.14. The values for Dr and FL are derived from correlation with the SPT blow count or
CPT resistance. In this procedure, the whole liqueﬁable deposit is subdivided into several layers, and FL is
evaluated for each layer. For each layer, the volumetric strain ev is obtained from the chart based on the values
of FL and Dr (or, if available, the normalised SPT N value or normalised CPT qc value) for that layer. The total

199

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 7.14 Chart for determining the volumetric strain as a function of the factor of safety (Ishihara and
Yoshimine, 1992; reprinted with permission from the author)

2.0

1.8 Clean sands

γmax = 1.5ε1max
1.6
Factor of safety for liquefaction, FL

1.4

1.2

3%
1.0 3.5%

4%
0.8 Dr = 40 Dr = 30
Dr = 50 N1 = 6 N1 = 3
Dr = 60 N1 = 10 qC1 = 45 qC1 = 33
6% N1 = 14 qC1 = 60
0.6 qC1 = 80
Dr = 70
8% (N1 = 20 qC1 = 110)
0.4 Dr = 80
γ max = 10% (N1 = 25 qC1 = 147)
Dr = 90%
0.2 N1 = 30
qC1 = 200 kgf/cm2

0
0 1.0 2.0 3.0 4.0 5.0
Post-liquefaction volumetric strain, εv

settlement is then calculated by integrating the strains over the layer thickness, as illustrated in Figure 7.15.
The method has been extended by Zhang et al. (2002) for use with CPT data.

7.6.3.2 Tokimatsu–Seed method

Another popular procedure that makes use of the SPT N value for estimating the settlement of sand is that
proposed by Tokimatsu and Seed (1987). In this method, the volumetric strains ϵv induced by earthquake
shaking are expressed as a function of the corrected SPT N value ((N1)60), and the CSR (tav /s 00 ), induced by
an earthquake shaking of Mw 7.5, as shown in Figure 7.16. For values of Mw other than 7.5, magnitude scaling
factors (MSFs) may be used.

From the ﬁgure, it can be observed that although the ﬁeld data are limited, the evidence is consistent with the
plots shown.

Based on the analysis of case histories, it was observed that vertical settlements of the order of 10 cm or less
correspond to the area where there was no destruction. Similarly, settlements ranging from 10 to 20 cm were
observed in area with an intermediate amount of damage, and when the settlements become larger than 30 cm,
considerable destruction always occurs on the ground surface, such as sand spurting, ﬁssures and large
offsets.

200

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

Figure 7.15 Diagram illustrating calculation of the total liquefaction-induced settlement of the ground

SPT N value FL Volumetric strain, εv

0
Sand

ΔH
5
Sand

10
Sandy
clay

15 Gravel

For each layer ΔS = ΔH εv

Total Settlement
S= Σ ΔS

Figure 7.16 Chart for the determination of volumetric strain from the normalised SPT N value and the CSR. Field
performance from selected earthquakes is indicated (Tokimatsu and Seed, 1987; reproduced with permission from
the publisher, ASCE)

0.6

Volumetric strain: %
10 5 4 3 2 1 0.5
0.5

0.4 0.2

0.1
τav /σ0′

0.3

Arahama
2 Hachinohe, PI
Hachinohe
0.2 9 P6
0.7
Niigata, C. Niigata, A.
0
2

0.1

0
0 10 20 30 40 50
(N1)60

201

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

7.7. Worked example

A geotechnical investigation of a level site was carried out to assess its liquefaction potential. The soil proﬁle
and the ﬁeld and laboratory results are indicated in Figure 7.17. The following are given

■ peak ground acceleration = 0.24g

■ Mw = 7.0
■ epicentral distance = 20 km.

It is required to

■ calculate the factors of safety FL for each layer

■ estimate the ground settlement
■ estimate the cyclic lateral displacement
■ estimate the magnitude of lateral spreading if the site is on sloping ground with a gradient of 3°.

Before proceeding with the solution, there are currently different liquefaction-triggering procedures available,
whether SPT based, CPT based or Vs based. Each procedure has its own set of correlations to determine the
CSR and CRR, and the FL values obtained from each of the procedures are not equivalent for the same input
data. Therefore, caution must be exercised when establishing an FL value. It must be stressed that the cor-
relations developed for each procedure (rd, (N1)60,cs, etc.) cannot be ‘mixed and matched’; each correlation
should only be used within its recommended procedure.

In the following calculations, the procedure outline by Youd et al. (2001) based on a report on National
Center for Earthquake Engineering Research (NCEER) workshops is adopted. For simplicity, the results are
presented in tabulated form, and explanations on the formulas used in each column are provided below the
tables.

Figure 7.17 Soil proﬁle for the worked example

Backfill
2.0 m γ = 15 kN/m3

Sand
3.0 m
Fc = 8%, γ = 18 kN/m3, D50 = 0.20 mm, (N60)ave = 8

Silty sand
2.0 m Fc = 30%, γ = 17.5 kN/m3, D50 = 0.10 mm, (N60)ave = 11

Sandy silt
2.0 m Fc = 50%, γ = 17.0 kN/m3, D50 = 0.05 mm, (N60)avet = 18

Thick clay layer

202

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

7.7.1 Calculation of the factor of safety

Column 1 2 3 4 5 6 7 8 9 10

Layer Depth: m σv: kPa σ0v : kPa rd CSR CN (N1)60 α β (N1)60,cs

1 3.5 57.0 42.3 0.98 0.21 1.54 12.32 0.3 1.01 12.74

2 6.0 101.5 62.3 0.96 0.25 1.27 13.97 4.7 1.15 20.77

3 8.0 136.0 77.2 0.94 0.26 1.14 20.52 5.0 1.20 29.62

Notes

Column 1: FL is evaluated at the midpoint of each layer. Layer 1 is the ﬁrst layer below the water table; the uppermost layer is above the
water table and therefore will not liquefy.

Column 2: The total stress is evaluated at the centre of each layer, by considering the weight of soil above the target point.

Column 3: The effective stress is calculated by subtracting the hydrostatic pressure from the total stress.

Column 4: Use ‘average data’ from Figure 7.2.

τ ave a σ0
Column 5: CSR = = 0:65 max r (Equation 7.1)
σ0v0 g σ00 d
rffiffiffiffiffiffiffiffiffi
Column 6: Cn = 100 (σ 0v in kN=m2 ) ≤ 1:7 (Equation 7.3)
σ 0v
ERm
Column 7: (N1 )60 = Cn N (Equation 7.2)
60 m
8
>
> 0 for F c ≤ 5%
>
>
<
Column 8: α = exp½1:76 − (190/F 2c ) for 5% < F c < 35% (Equation 7.4b)
>
>
>
>
:
5 for F c ≤ 35%
8
>
> 1:0 for F c ≤ 5%
>
>
<
Column 9: β = ½0:99 + (F 1:5
c /1000) for 5% < F c < 35% (Equation 7.4c)
>
>
>
>
:
1:2 for F c ≤ 35%
Column 10: (N1 )60cs = α + β(N1 )60 (Equation 7.4a)

203

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Column 11 12 13 14 15 16

Layer CRRM = 7.5 MSF Kσ CRRM = 7.0 FL

1 0.14 1.19 1.0 0.17 0.81 Liquefaction

2 0.23 1.19 1.0 0.27 1.08 No

3 0.50 1.19 1.0 0.60 2.31 No

Notes

Column 11: CRRM = 7.5 is obtained from the SPT clean sand base curve (see curve 3 in Figure 7.4).

Column 12: Use the NCEER lower bound (MSF = 102.24/Mw2.56) (Equation 7.9)

f −1
σ0v0
Column 13: Kσ = pa ≤ 1:0 (Equation 7.10)

Column 14: CRR = (CRR)M = 7.5 MSF Kσ (see Figure 7.7)

Column 15: FL = CRR (Equation 7.11)

CSR

7.7.2 Calculation of settlement

(a) Ishihara–Yoshimine method

Column 1 17 10 18 16 19 20

Layer Depth: m ΔH: m (N1)60,cs Dr : % FL εv: % ΔS: cm

1 3.5 3.0 12.74 52.6 0.81 3.2 9.6

2 6.0 2.0 20.77 67.2 1.08 0.5 1.0

3 8.0 2.0 29.62 80.2 2.31 0.0 0.0

S = 10.6

Notes
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
(N1 )60,cs
Column 18: Dr = (estimated from Idriss and Boulanger (2008))
46

Column 19: εv is obtained from Figure 7.14.

Column 20: ΔS = ΔH εv, S = ∑ΔS (see Figure 7.15)

204

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

(b) Tokimatsu–Seed method

Column 1 17 7 5 12 21 22 23

Layer Depth: m ΔH: m (N1)60 CSR MSF CSRM = 7.5 εv: % ΔS: cm

1 3.5 3.0 12.32 0.21 1.19 0.18 2.2 6.6

2 6.0 2.0 13.97 0.25 1.19 0.21 1.5 3.0

3 8.0 2.0 20.52 0.26 1.19 0.22 0.4 0.8

S = 10.4

Notes

Column 21: CSRM=7:5 = CSR (refer to Figure 7.7)

MSF

Column 22: εv is obtained from Figure 7.16.

Column 23: ΔS = ΔH εv, S = ∑ΔS (see Figure 7.15)

7.7.3 Calculation of the cyclic lateral displacement

Tokimatsu–Asaka method

Column 1 17 24 5 25

Layer Depth: m H: m (N1)60,cs CSR γcy: %

1 3.5 3.0 12.74 0.21 2.0

2 6.0 2.0 20.77 0.25 0.5

3 8.0 2.0 29.62 0.26 0.0

Notes
Column 24: Na = (N1)60,cs

Column 25: γcy is obtained from Figure 7.10.

The cyclic displacement proﬁle is given in Figure 7.18.

205

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 7.18 Worked example: cyclic displacement proﬁle

7 cm

2m
6 cm

3m 2%

1 cm

2m
0.5%

7.7.4 Calculation of lateral spreading

Youd et al. (2002) multi-linear regression method

log DH = −16:713 + 1:532M − 1:406 log R* − 0:012R

+ 0:338 log S + 0:540 log T15 + 3:413 log (100 − F15 ) (Equation 7:28; also on p: 196)
− 0:795 log (D5015 + 0:1)
M = 7.0
R = 20 km
R* = R + 100.89M − 5.64 = 20 + 100.89 × 7 − 5.64 = 23.89 km
S = tan(3°) × 100% = 5.24%
T = 3m + 2m = 5m
F15 = 3 m 8%5+m2 m 30% = 16:8%
D5015 = 3 m 0:2 mm 5+m2 m 0:10 mm = 0:16 mm

Solving for DH

DH = 0.30 m = 30 cm

With the displacement of the surface crust known, it is estimated that the liqueﬁed subsoil layer would have a
sinusoidal shape, and the proﬁle of lateral displacement is illustrated in Figure 7.19.

206

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

Figure 7.19 Worked example: lateral displacement proﬁle

30 cm

7.8. Planning and design of a major project in seismic areas

The planning and design of major structures at sites in seismically active areas, especially those susceptible to
liquefaction and lateral spreading effects, are a complex and challenging undertaking; consequently, a large
number of issues associated with the process should be considered. The key considerations and issues that are
important for the design of structures on sites prone to liquefaction and lateral spreading include

■ appropriate geotechnical investigations

■ evaluation of liquefaction and lateral spreading
■ detailed analysis of the effects of liquefaction and lateral spreading on the structure
■ ground improvement
■ structural design issues
■ post-construction monitoring issues.

The quality of the design solution that would be developed, the seismic behaviour of structures, and the ability
of the structural and geotechnical engineers to assess the post-earthquake condition of the structure are highly
dependent on whether all of these issues have been adequately addressed. Some of the major considerations
that need to be considered are discussed below.

207

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

7.8.1 Preliminary considerations

The preliminary considerations for the design of structures on sites that are susceptible to liquefaction and
lateral spreading include

■ compliance with the structure performance requirements and standards given by design codes and
standards
■ liquefaction effects on structures (total and differential settlement of a structure, effects on services
carried by the structure, reduction in the bearing capacity, soil stiffnesses and overall stiffness of the
structure; possible down-drag loading on the foundation piles, etc.)
■ lateral spreading effects on structures (loads from liquefied soils and from non-liquefiable crust,
the effect of lateral spreading on the foundations, and the opposite, reduced strength properties for
laterally spreading liquefied material, and the form of likely damage to foundations, e.g. piles)
■ the foundation capacity following liquefaction (generation of excess pore water pressure in liquefied
layers and its effect on the bearing capacity, settlement and structural capacity of the foundation).

7.8.2 Geotechnical investigations

In order to understand the full extent of the liqueﬁable deposits and the level of liquefaction susceptibility,
site-speciﬁc geotechnical investigations need to be undertaken.

7.8.2.1 Preliminary screening

A preliminary geotechnical appraisal is carried out initially to establish the site topography, stratigraphy and
location of the groundwater table. The appraisal should establish

■ the depth to the bedrock and the likely site subsoil class
■ the soil nature, structure and fabric
■ the groundwater conditions
■ the density of soils
■ the geological age (time under a signiﬁcant overburden pressure can increase liquefaction resistance;
sediments that are geologically young are generally more susceptible to liquefaction)
■ Prior seismic activity at the site and the cyclic loading history
■ the overconsolidation effects
■ the drainage conditions
■ the effective conﬁning stress
■ geological hazards such as slope stability, tsunami and so on
■ the liquefaction potential of the site based on available information.

At this stage, it would be worthwhile to compile historical evidence for the site, including information and
documents on local land use, ﬁlls, site features before construction, and old river channels, waterways or land
features associated with high liquefaction potential. The historical performance of the site in past earthquake
events should be carefully considered in the site evaluation, whenever such evidence is available. Indeed,
evidence that a site has liqueﬁed before would indicate that the site can liquefy again.

If the preliminary geotechnical appraisal indicated the presence of liquefiable materials, site-specific geo-
technical investigations may be required. A programme of geotechnical investigation should be developed to
obtain sufficient information for the assessment of the liquefaction potential of the site soils and for seismic
analysis.

208

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

7.8.2.2 Investigation plan

A suitable investigation should include the following tests, depending on the scale and type of structure to be
constructed

■ boreholes with SPTs

■ trial pits
■ CPTs
■ shear wave velocity tests
■ Scala penetrometer tests
■ seismic piezocone penetrometer tests
■ static or cyclic plate load tests
■ laboratory testing, including particle size analysis, Atterberg limits, speciﬁc gravity, organic content,
unit weight and shear box tests
■ in situ testing of the seismic behaviour of soils using truck-based vibration machines or explosives.

All of the information collected during the preliminary geotechnical appraisal and the geotechnical inves-
tigations should be interpreted and reported. The geotechnical investigation report should outline the
methodology used in the interpretation of the collected information.

The presence of liquefiable soils within the site can substantially increase project costs; therefore, it is imperative
to characterise the site as thoroughly as possible. Note that identifying the presence of liquefiable soils is
generally not enough; rather, it is important to identify the extent of liquefiable soils, in both the vertical and
lateral directions, to determine whether soil liquefaction will indeed pose a hazard to the structure.

Generally, the investigation of sites with liquefiable layers presents special problems. Conventional
approaches, such as the use of unsupported test pit excavations and hand augers, are usually incapable of
penetrating depths below the water table, especially if the layer consists of loose, cohesionless soils. Pro-
cedures that provide continuous measurement of the in situ state of the soil (e.g. the CPT) are preferred
because complex stratification is commonly associated with high-risk geomorphologies, and even relatively
thin strata of liquefiable soil may pose a significant hazard in some cases.

If obtaining high-quality undisturbed soil samples of loose, cohesionless soils is preferred, many techniques
are currently available, such as ground freezing techniques, gel-push samplers and Dames & Moore
(Osterberg-type) hydraulic ﬁxed-piston samplers (see Section 7.3.2); these techniques can generally result in
recovery of high-quality samples. However, the costs associated with such procedures may be prohibitive
when dealing with smaller projects.

7.8.3 Evaluation of liquefaction and lateral spreading

If geotechnical investigation indicates the presence of liquefiable strata, quantitative liquefaction analysis
using site-specific field and laboratory test data should be conducted using the procedures outlined in
Section 7.3. Note that in liquefaction potential evaluation, the peak ground acceleration (amax) is required.
Generally, the ground motion parameters at a site may be evaluated using one of the following methods

■ method 1 – risk-based method using the earthquake hazard presented in relevant codes or standards
■ method 2 – site-speciﬁc probabilistic seismic hazard analysis
■ method 3 – site-speciﬁc response analysis.

209

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Method 1 is appropriate for routine engineering design projects. Methods 2 and 3 are preferred for more
signiﬁcant projects, more complex sites or other cases where advanced analysis can be justiﬁed.

Once the liquefaction potential is evaluated, an assessment of the consequences of liquefaction, such as
seismic-induced settlement, cyclic ground displacement and permanent ground displacement due to lateral
spreading, in terms of their impact to the structure is performed. Simpliﬁed methods or advanced numerical
modelling procedures can be used to determine the expected ground deformations.

7.9. Summary
The evaluation of soil liquefaction and its effects involve several steps, from determining the seismic demand
at the site to evaluating liquefaction susceptibility, liquefaction triggering and liquefaction-induced defor-
mations. These steps may involve either simpliﬁed or detailed analysis procedures. A good understanding of
the liquefaction process and phenomena is critical in the evaluation of liquefaction (i.e. pore pressure
characteristics and liquefaction-induced deformation).

REFERENCES
AIJ (Architectural Institute of Japan) (2001) Recommendations for Design of Building Foundations. AIJ, Tokyo,
Japan. (In Japanese.)
Andrus RD and Stokoe KH (1997) Liquefaction resistance based on shear wave velocity. Proceedings of the
NCEER Workshop on Evaluation of Liquefaction Resistance of Soils, Buffalo, NY, USA, pp. 89–128.
Andrus RD and Stokoe KH (2000) Liquefaction resistance of soils from shear-wave velocity. Journal of
Geotechnical and Geoenvironmental Engineering, ASCE 126(11): 1015–1025.
Andrus RD, Stokoe KH and Chung RML (1999) Draft Guidelines for Evaluating Liquefaction Resistance
Using Shear Wave Velocity Measurements and Simpliﬁed Procedures. National Institute of Standards and
Technology, Gaithersburg, MD, USA.
Bardet JP and Tobita T (2001) NERA: A Computer Program for Nonlinear Earthquake Site Response Analyses of
Layered Soil Deposits. Department of Civil Engineering, University of Southern California, Los Angeles, CA, USA.
Bartlett SF and Youd TL (1995) Empirical prediction of liquefaction-induced lateral spread. Journal of
Geotechnical Engineering, ASCE 121(4): 316–329.
Baziar MH, Dobry R and Elgamal AW (1992) Engineering Evaluation of Permanent Ground Deformations due to
Seismically-induced Liquefaction. National Center for Earthquake Engineering Research, Buffalo, NY, USA,
Technical Report NCEER-92-0007.
Boulanger RW (2003) High overburden stress effects in liquefaction analyses. Journal of Geotechnical and
Geoenvironmental Engineering, ASCE 129(12): 1071–082.
Boulanger R and Idriss IM (2006) Liquefaction susceptibility criteria for silts and clays. Journal of Geotechnical
and Environmental Engineering, ASCE 132(11): 1413–1426.
Boulanger R and Idriss IM (2014) CPT and SPT Based Liquefaction Triggering Procedures. Center for
Geotechnical Modeling, University of California, Davis, CA, USA, Report UCD/CGM-14/01.
Bray JD, Sancio RB, Durgunoğlu HT et al. (2004) Subsurface characterization at ground failure sites in Adapazari,
Turkey. Journal of Geotechnical and Environmental Engineering, ASCE 130(7): 673–685.
Byrne PM (1990) A Model for Predicting Liquefaction-induced Displacements. Department of Civil Engineering,
University of British Columbia, Vancouver, Canada.
CEN (Comité Européen de Normalisation) (2004a) EN 1998-1. Design of structures for earthquake resistance. Part 1:
General rules, seismic actions and rules for buildings. CEN, Brussels, Belgium.

210

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

CEN (2004b) EN 1998-5. Design of structures for earthquake resistance. Part 5: Foundations, retaining structures
and geotechnical aspects. CEN, Brussels, Belgium.
Cetin KO, Seed RB, Kiureghain AD et al. (2004) Standard penetration test-based probabilistic and deterministic
assessment of seismic soil liquefaction potential. Journal of Geotechnical and Geoenvironmental Engineering,
ASCE 130(12): 1314–1340.
DIANA Analysis (2001) DIANA 2D Version 7.2. DIANA Analysis, Delft, the Netherlands.
Hamada M, Yasuda S, Isoyama R and Emoto K (1986) Study on Liquefaction-induced Permanent Ground
Displacements. Association for the Development of Earthquake Prediction, Tokyo, Japan.
Hashash YMA and Park D (2001) Non-linear one-dimensional seismic ground motion propagation in the
Mississippi embayment. Engineering Geology 62(1–3): 185–206.
Hatanaka M, Sugimoto M and Yoshio S (1985) Liquefaction resistance of two alluvial volcanic soils sampled by
in-situ freezing. Soils and Foundations 25(3): 49–63.
Hynes ME and Olsen RS (1999) Influence of confining stress on liquefaction resistance. Proceedings of the International
Workshop on the Physics and Mechanics of Soil Liquefaction. Balkema, Rotterdam, the Netherlands, pp. 145–152.
Idriss IM (1999) An update to the Seed–Idriss simplified procedure for evaluating liquefaction potential.
Proceedings of the TRB Workshop on New Approaches to Liquefaction. Federal Highway Administration,
Washington, DC, USA.
Idriss IM and Boulanger RW (2004) Semi-empirical procedures for evaluating liquefaction potential during
earthquakes. Proceedings of the 11th International Conference on Soil Dynamics and Earthquake Engineering
(ICSDEE) and the 3rd International Conference on Earthquake Geotechnical Engineering (ICEGE) (Doolin D,
Kammerer A, Nogami T, Seed RB and Towhata I (eds)). Stallion Press, Singapore, vol. 1, pp. 32–56.
Idriss IM and Boulanger RW (2006) Semi-empirical procedures for evaluating liquefaction potential during
earthquakes. Journal of Soil Dynamics and Earthquake Engineering 26: 115–130.
Idriss IM and Boulanger RW (2008) Soil Liquefaction During Earthquakes. Earthquake Engineering Research
Institute, Oakland, CA, USA.
Ishihara K (1985) Stability of natural deposits during earthquakes. 11th International Conference on Soil
Mechanics and Foundation Engineering, San Francisco, CA, USA, vol. 2, pp. 321–376.
Ishihara K (1993) Liquefaction and flow failure during earthquakes. Géotechnique 43(3): 351–451.
Ishihara K (1996) Soil Behaviour in Earthquake Geotechnics. Clarendon Press, Wotton-under-Edge, UK.
Ishihara K and Yoshimine M (1992) Evaluation of settlements in sand deposits following earthquakes. Soils and
Foundations 32(1): 173–188.
Itasca (1998) FLAC Manual Vers. 4.0. Itasca Consulting Group, Minneapolis, MN, USA.
Iwasaki T, Tokida K, Tatsuoka F, Watanabe S, Yasuda S and Sato H (1982) Microzonation for soil liquefaction
potential using simplified methods. Proceedings of the 3rd International Conference on Microzonation, Seattle,
WA, USA, vol. 3, pp. 1319–1330.
Japanese Geotechnical Society (1998) Remedial Measures Against Soil Liquefaction. Balkema, Rotterdam,
the Netherlands.
JRA (Japan Road Association) (2002) Specifications for highway bridges. Part V: seismic design. JRA, Tokyo,
Japan. (In Japanese.)
Kayen RE, Mitchell JK, Seed RB, Lodge A, Nishio S and Coutinho R (1992) Evaluation of SPT-, CPT- and shear
wave-based methods for liquefaction potential assessment using Loma Prieta data. Proceedings of the 4th
Japan-U.S. Workshop on Earthquake-Resistant Design of Lifeline Facilities and Countermeasures for Soil
Liquefaction, New York, NY, USA, vol. 1, pp. 177–204.

211

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Kottke AR and Rathje EM (2008) Technical Manual for Strata. University of California, Berkeley, CA, USA,
PEER Report 2008/10.
Lee DH, Ku CS and Yuan H (2004) A study of liquefaction risk potential at Yuanlin, Taiwan. Engineering Geology
71(1–2): 97–117.
Lee WF and Chen CH (2013) A case study on silty sand liquefaction – 2010 Hsin Hwa liquefaction in Taiwan.
Proceedings of the International Conference on Earthquake Geotechnical Engineering, Istanbul, Turkey.
Lees JJ, Ballagh RH, Orense RP, van Ballegooy S (2015) CPT-based analysis of liquefaction and re-liquefaction
following the Canterbury earthquake sequence. Soil Dynamics and Earthquake Engineering 79(B): 304–314.
Liao SSC and Whitman RV (1986) Catalogue of Liquefaction and Non-liquefaction Occurrences During
Earthquakes. Department. of Civil Engineering, Massachusetts Institute of Technology, Cambridge, MA,
UA, Research Report.
Lodge AL (1994) Shear Wave Velocity Measurements For Subsurface Characterization. PhD thesis, University of
California, Berkeley, CA, USA.
Maurer BW, Green RA, Cubrinovski M and Bradley BA (2015) Evaluation of the liquefaction potential index for
assessing liquefaction hazard in Christchurch, New Zealand. Journal of Geotechnical and Geoenvironmental
Engineering, ASCE 140(7): 10.1061/(ASCE)GT.1943-5606.0001117.
Mazzoni S, McKenna F and Fenves GL (2011) OpenSees Getting Started Manual. http://opensees.berkeley.edu/
wiki/index.php/Getting_Started (accessed 15/08/2018).
Mori K and Sakai K (2016) The GP sampler: a new innovation in core sampling. Proceedings of the 5th
International Conference on Geotechnical and Geophysical Site Characterisation, Gold Coast, Queensland,
Australia, pp. 99–124.
Moss RES, Seed RB, Kayen RE, Stewart JP, Der Kiureghian A and Cetin KO (2006) CPT-based probabilistic and
deterministic assessment of in situ seismic soil liquefaction potential. Journal of Geotechnical and
Geoenvironmental Engineering, ASCE 132(8): 1032–1051.
NCEER (National Center for Earthquake Engineering Research) (1997) Proceedings of the NCEER Workshop on
Evaluation of Liquefaction Resistance of Soils (Youd TL and Idriss IM (eds)).Taipei, Taiwan, pp. 41–88,
Technical Report NCEER 97-022.
Newmark NM (1965) Effects of earthquakes on dam and embankments. Géotechnique 5(2): 137–160.
OCDI (Overseas Coastal Area Development Institute) (2009) Technical standards and commentaries for port and
harbour facilities in Japan. OCDI, Tokyo, Japan.
Orense RP, Hickman NA, Hill BR and Pender MJ (2014) Spatial evaluation of liquefaction potential in
Christchurch following the 2010/2011 Canterbury earthquakes. International Journal of Geotechnical
Engineering 8(4): 420–425.
Plaxis (2011) PLAXIS 2D Reference Manual. Plaxis, Delft, the Netherlands.
Robertson PK and Wride CE (1997) Cyclic liquefaction and its evaluation based on the SPT and CPT. Proceedings
of the NCEER Workshop on Evaluation of Liquefaction Resistance of Soils (Youd TL and Idriss IM (eds)).
Taipei, Taiwan, Technical Report NCEER 97-022.
Robertson PK and Wride CE (1998) Evaluating cyclic liquefaction potential using the cone penetration test.
Canadian Geotechnical Journal 35(3): 442–459.
Robertson PK, Woeller DJ and Finn WD (1992) Seismic cone penetration test for evaluating liquefaction potential
under cyclic loading. Canadian Geotechnical Journal 29: 686–695.
Schnabel PB, Lysmer J and Seed HB (1972) SHAKE: A Computer Program for Earthquake Response Analysis of
Horizontally Layered Sites. Earthquake Engineering Research Center, University of California, Berkeley, CA,
USA, Report UCB/EERC-72/12.
Seed HB (1979) Soil liquefaction and cyclic mobility evaluation for level ground during earthquakes. Journal of
the Geotechnical Engineering Division, ASCE 105(2): 201–255.

212

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Liquefaction: practical aspects

Seed RB (2010) Technical Review and COMMENTS: 2008 EERI Monograph ‘Soil Liquefaction During
Earthquakes’ (by I.M. Idriss and R.W. Boulanger)., University of California, Berkeley, CA, USA, Geotechnical
Report UCB/GT 2010/01.
Seed RB and Harder LF (1990) SPT-based analysis of cyclic pore pressure generation and undrained residual
strength. Proceedings of the H Bolton Seed Memorial Symposium. BiTech, Vancouver, Canada, pp. 351–376.
Seed HB and Idriss IM (1971) Simplified procedure for evaluating soil liquefaction potential. Journal of Soil
Mechanics and Foundation Division, ASCE 97(9): 1249–1273.
Seed HB and Idriss IM (1982) Ground Motions and Soil Liquefaction during Earthquakes, Earthquake
Engineering Research Institute, Oakland, CA, USA.
Seed HB, Tokimatsu K, Harder LF, Chung RM (1984) The Influence of SPT Procedures in Soil Liquefaction
Resistance Evaluations. Earthquake Engineering Research Center, University of California, Berkeley, CA,
USA, Report UCB/EERC-84/15.
Seed HB, Tokimatsu K, Harder LF and Chung RM (1985) Influence of SPT procedures on soil liquefaction
resistance evaluation. Journal of the Geotechnical Engineering Division, ASCE 111(GT12): 1425–1445.
Seed RB, Cetin KO, Moss RES et al. (2003) Recent advances in soil liquefaction engineering: a unified and consistent
framework. 26th Annual Geotechnical Spring Seminar, Long Beach, CA, USA, keynote presentation.
Sego DC, Hofmann BA, Robertson PK and Wride CE (1999) Undisturbed sampling of loose sand using in-situ
ground freezing. Proceedings of the International Workshop on the Physics and Mechanics of Soil Liquefaction,
Baltimore, MA, USA, pp. 179–190.
Shibata T and Teparaksa W (1988) Evaluation of liquefaction potentials of soils using cone penetration tests. Soils
and Foundations 28(2): 49–60.
Suzuki Y, Tokimatsu K, Taya Y and Kubota Y (1995) Correlation between CPT data and dynamic properties of in
situ frozen samples. Proceedings of the 3rd International Conference on Recent Advances in Geotechnical
Earthquake Engineering and Soil Dynamics, St Louis, MO, USA, vol. I.
Suzuki Y, Koyamada K and Tokimatsu K (1997) Prediction of liquefaction resistance based on CPT tip resistance
and sleeve friction. Proceedings of the 14th International Conference on Soil Mechanics and Foundation and
Engineering, Hamburg, Germany, vol. 1, pp. 603–606.
Tokimatsu K and Asaka Y (1998) Effects of liquefaction-induced ground displacements on pile performance in the
1995 Hyogoken–Nambu earthquake. Soils and Foundations, pp. 163–177 (special issue).
Tokimatsu K and Seed HB (1987) Evaluation of settlements in sand due to earthquake shaking. Journal of
Geotechnical Engineering, ASCE 113(8): 861–878.
Tokimatsu K and Uchida A (1990) Correlation between liquefaction resistance and shear wave velocity. Soils and
Foundations 30(2): 33–42.
Tonkin & Taylor (2013) Liquefaction Vulnerability Study. Earthquake Commission, Wellington, New Zealand.
Toprak S and Holzer TL (2003) Liquefaction potential index: field assessment. Journal of Geotechnical and
Geoenvironmental Engineering, ASCE 129(4): 315–322.
Towhata I (2008) Geotechnical Earthquake Engineering. Springer, Heidelberg, Germany.
Tsuchida H (1971) Estimation of liquefaction potential of sandy soils. Proceedings of the 3rd Joint Meeting of the
US–Japan Panel on Wind and Seismic Effects, Tokyo, Japan.
Wakamatsu K (1992) Evaluation of liquefaction susceptibility based on detailed geomorphological classification.
Proceedings of the Annual Meeting of the Architectural Institute of Japan, Tokyo, Japan, vol. B, pp. 1443–
1444. (In Japanese.)
Wu J (2002) Liquefaction Triggering and Post-Liquefaction Deformations of Monterey 0/30 Sand Under
Uni-Directional Cyclic Simple Shear Loading. PhD thesis, University of California, Berkeley, CA, USA.
Yoshimi Y, Hatanaka M and Oh-oka H (1978) Undisturbed sampling of saturated sands by freezing. Soils and
Foundations 18(3): 59–73.

213

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Youd TL and Perkins DM (1978) Mapping liquefaction-induced ground failures. Journal of the Geotechnical
Engineering Division, ASCE 104(GT4): 433–446.
Youd TL, Idriss IM, Andrus RD et al. (2001) Liquefaction resistance of soils: summary report from the 1996
NCEER and 1998 NCEER/NSF workshops on evaluation of liquefaction resistance of soils. Journal of the
Geotechnical Engineering Division, ASCE 127(4): 297–313.
Youd TL, Hansen CM and Bartlett SF (2002) Revised multi-linear regression equations for prediction of lateral
spread displacement. Journal of Geotechnical and Environmental Engineering, ASCE 128(12): 1007–1017.
Zhang G, Robertson PK and Brachman R (2002) Estimating liquefaction-induced ground settlements from the
CPT. Canadian Geotechnical Journal 39: 1168–1180.
Zhang G, Robertson PK and Brachman R (2004) Estimating liquefaction-induced lateral displacements using the
standard penetration test or cone penetration test, Journal of Geotechnical and Geoenvironmental Engineering,
ASCE 130(8): 861–871.

214

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

ISBN 978-0-7277-6166-8
https://doi.org/10.1680/sdof.61668.215
ICE Publishing: All rights reserved

Chapter 8
Analysis and design of shallow
foundations
8.1. Introduction
Foundation failures due to seismic shaking have been observed during several major earthquakes in the past.
Such failures have occurred in soils susceptible to liquefaction, as well as in soils that did not liquefy. The
recent Canterbury earthquakes in New Zealand caused signiﬁcant bearing capacity failure not only of high-
rise commercial buildings but of many residential structures as well. Multi-storey buildings punched into,
tilted excessively and slid laterally on softened ground.

The estimation of the magnitude of settlement that a structure will undergo during seismic shaking is therefore
an important consideration in performance-based design. Traditionally, most estimations of building set-
tlements induced by soil liquefaction have been based on empirical equations or rules of thumb that were
developed to estimate post-liquefaction consolidation settlement for free-ﬁeld conditions; however, this
approach cannot possibly capture shear-induced and localised volumetric-induced deformations in the soil
underneath the shallow foundations, which, in turn, are affected by many factors, including building
geometry and the properties and conditions of the foundation ground. Moreover, the development of excess
pore water pressure that could induce changes in ground motion has not been considered directly in current
procedures.

8.2. Seismic bearing capacity of shallow foundations

One of the ways of transferring load from a structure to the ground is through the use of shallow foundations,
in which the loads are spread laterally at the contact between the foundation element and the ground. For
shallow foundations in soil, such load transfer occurs predominantly through the foundation base, and only a
small percentage of the load is transferred through the side of the foundation.

As the load on a foundation is increased, it ﬁrst experiences some settlement; once the load exceeds a certain
threshold value, some portion of the soil enters the plastic range. Further increase in the load would result in
enlargement of the plastic zones within the soil mass to the point that the free boundary is reached (say the
ground surface), when large settlement is possible without any increase in load; at this point, the soil would
have undergone shear failure (Figure 8.1). The bearing capacity of the foundation is deﬁned as the maximum
value of the load applied that will not cause shear failure in the soil. The theoretical maximum load
(or pressure) that can be supported without failure is called the ultimate bearing capacity; the allowable
bearing capacity is the ultimate bearing capacity divided by a factor of safety.

The bearing capacity of shallow foundations has been well established for the static case. Starting from a
postulated failure mechanism (circular failure surface, two-block failure surface, etc.), the load capacity is

215

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 8.1 General shear foundation failure, showing the formation of shear surfaces and changes in the load–
settlement relation

Load

Initial ground Final ground

surface surface

Shear surfaces

Settlement Load

evaluated from which the critical case corresponding to that mechanism is located; other mechanisms are
also examined. A typical bearing capacity equation is (Brinch Hansen, 1970; Meyerhof, 1951; Terzaghi,
1943)

1
qu = cNc + qNq + BNg (8:1)
2

where qu is the ultimate bearing capacity, c is the cohesion, q is the surcharge load at the level of the base
of the footing, g is the unit weight of the soil below the base of the footing, B is the footing width, and Nc,
Nq and Ng are known as bearing capacity factors. Similar equations have been derived for various shapes
of the foundation element.

Many spread footings are subjected to seismic loads, in addition to the static loads. For purposes of foundation
design, these loads are nearly always expressed in terms of equivalent static loads that are primarily horizontal, so
they produce shear loads on the foundations and thus require a shear load capacity evaluation. In addition, seismic
loads on the superstructure can produce additional normal loads (either downward or upward) on the foundations
that are superimposed on the static normal loads. When these loads are expressed as equivalent static loads, the
methods of evaluating the load capacity of foundations are essentially the same as for static loads. Based on this
pseudo-static concept, many analytical and numerical solutions are available, such as the limit equilibrium method,
limit analysis, methods of characteristics and ﬁnite-element analysis, to compute the seismic bearing capacity
factors required for the design of a foundation. Another approach adopted by many researchers is to determine the
reduction in the ultimate bearing capacity by incorporating pseudo-static seismic forces (e.g. Budhu and Al-Karni,
1993; Dormieux and Pecker, 1995; Sarma and Iossifelis, 1990).

However, in addition to the reduction in bearing capacity, foundations may also experience an increase in
settlement due to seismic loading. Thus, two sources of loading must be taken into account in the design of

216

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Analysis and design of shallow foundations

shallow foundations during earthquakes: ‘inertial’ loading caused by the lateral forces imposed on the
superstructure, and ‘kinematic’ loading caused by the ground movements developed during the earthquake.
In fact, Part 5 of Eurocode 8 (CEN, 2004) states that foundations shall be designed for the following two
loading conditions

■ inertia forces on the superstructure transmitted on the foundations in the form of axial, horizontal
forces and moment
■ soil deformations arising from the passage of seismic waves.

When considering the above loading conditions, structures are designed such that (a) there must be an
adequate factor of safety against bearing capacity failures and (b) settlement due to the building loads must
not exceed an acceptable value. Based on observations of building failures during earthquakes, the number of
structures damaged by earthquake-induced settlements is far more than those that have earthquake-induced
bearing capacity failures; thus, the settlement criterion generally governs.

The most common cause of seismic bearing capacity failure is liquefaction of the underlying soil. Liqueﬁable
soils are categorised by many seismic codes as extreme ground conditions. Liquefaction-induced shear
strength degradation of the foundation subsoil may result in post-shaking static bearing capacity failure, but
it is the excessive seismic settlements that accumulate that pose a hazard to many structures. Thus, while
estimation of the degraded bearing capacity of liqueﬁed subsoil may be important, the accurate estimation of
the associated dynamic settlements is necessary to ensure a viable performance-based design of shallow
footings.

Most estimations of building settlements induced by soil liquefaction are based on empirical correlations or
rules of thumb that were developed to estimate post-liquefaction consolidation settlement for free-ﬁeld
conditions; however, this practice ignores deviatoric deformations (settlements due to the cyclic inertial
forces acting on the structures within the liqueﬁable soil under a building foundation) as well as volumetric
deformations due to localised drainage during shaking. It follows, therefore, that the magnitude of building
settlement is not only a function of the soil properties and earthquake loads but also of the structure geometry
as well.

8.3. Case studies: observed settlements of buildings during earthquakes

Failures of buildings on shallow foundation have been observed during several historical and recent
earthquakes. Such failures have occurred mostly in soils prone to liquefaction. Some of the well-known
documented cases of building damage due to liquefaction are discussed below.

8.3.1 1964 Niigata earthquake

The moment magnitude (Mw) 7.5 earthquake that shook Niigata City (Japan) in 1964 brought soil lique-
faction to the forefront of the geotechnical engineering profession. Most of the lower part of the city is built on
recent deltaic deposits from the Shinano and Agano rivers, mainly consisting of unconsolidated sand. One of
the peculiar phenomena observed during the earthquake was the settlement and tilting of many reinforced
concrete buildings in Kawagishi-cho, which can be attributed to the liquefaction of the foundation ground.
It was reported that approximately 340 reinforced concrete buildings in the city were damaged as a result of
extensive liquefaction of the sandy ground (Kishida, 1966; Ohsaki, 1966; Seed and Idriss, 1967). Figure 8.2
illustrates some of the damage observed in the area. Many of the three- to four-storey structures settled more
than 1 m, often accompanied by severe tilting. The maximum settlement was reported to be 3.8 m (Yoshimi
and Tokimatsu, 1977).

217

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 8.2 Failure of reinforced concrete buildings on shallow foundation at Kawagishi-cho, Niigata, due to soil
liquefaction during the Niigata earthquake (Photos courtesy of Hiroshi Takeuchi and the Department of Civil
Engineering, University of Tokyo)

Following an investigation of the degree of settlement of the buildings, Yoshimi and Tokimatsu (1977)
considered the depth of the liqueﬁable layer to dominate the settlement estimation. For a given depth of
free-ﬁeld liquefaction, the settlement of the structure decreased as the width of the structure increased.
This relationship was independent of the number of storeys, the presence of a basement or the presence of
short friction piles.

8.3.2 1990 Luzon earthquake

The Mw 7.7 earthquake that rocked the north central region of Luzon Island (Philippines) on 16 July 1990
caused widespread damage to buildings and other structures. Although located about 100 km from the
epicentre, Dagupan City suffered the most damage due to the liquefaction of a loose saturated sand deposit.
Most of the city’s land areas were swampy regions that were reclaimed by ﬁlling as the business centre
developed; hence, the poorly compacted nature of such deposits accounts for the area’s high susceptibility to
liquefaction (Orense et al., 1991).

Field testing after the earthquake indicated that liquefaction must have developed through the sandy deposit
from a depth of 0–3 m down to a depth of 10 m. Investigations performed by Adachi et al. (1992) revealed
settlements of 30 damaged buildings ranging from 0.25 m to as much as 2.5 m. Most of these buildings were
on shallow foundations bearing on uniform ﬁne clean sand. Some of the damaged buildings are shown in
Figure 8.3. In addition, Tokimatsu et al. (1991) investigated 120 reinforced concrete buildings, and noted that
most of them had two to four storeys and were founded on shallow footings with no piles, and that some of the
most signiﬁcant settlements were observed in corner buildings, in buildings without adjacent structures on
one or both sides, in buildings surrounded by light-weight structures, and in those parts of the area where
there was greater separation between adjacent buildings. All of these observations pointed to the importance

218

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Analysis and design of shallow foundations

Figure 8.3 Examples of buildings in Dagupan City that sank and tilted due to soil liquefaction following the 1990
Luzon earthquake

of the conﬁning effect of the whole building and adjacent structures in reducing the level of settlement.
This factor appeared to be more important than the width and other characteristics of the individual footings.
The damage to the superstructure was greater for buildings that had individual shallow footings without tie
beams or with beams of low rigidity, thus allowing for signiﬁcant differential settlements.

8.3.3 2010–2011 Canterbury earthquake sequence

The city of Christchurch (New Zealand) and some surrounding towns experienced widespread damage due
to soil liquefaction induced by seismic shaking during the 2010–2011 Canterbury earthquake sequence,
beginning on 4 September 2010 with the Mw 7.1 Darﬁeld earthquake. Other notable earthquakes in this
sequence were the 22 February 2011 Mw 6.2 Christchurch earthquake, and the two earthquake events on
13 June 2011 and 23 December 2011. Each resulted in widespread liquefaction-induced damage, with the
most severe damage occurring during the Christchurch earthquake (Orense et al., 2012). Ground subsidence
resulting from liquefaction ejecta, liquefaction-induced differential settlement and lateral spreading were the
principal ground deformation modes that damaged residential dwellings in the Canterbury region. Typical
damage observed in residential dwellings is illustrated in Figure 8.4.

Following the 2011 Christchurch earthquake, visual assessment of foundation damage was carried out on
60 000 residential houses in Christchurch by geotechnical engineers as part of detailed land damage
assessments for the New Zealand Earthquake Commission. Visually observed liquefaction-related damage to
residential house foundations was assessed based on criteria reﬂecting the type of damage and its severity.
The visual assessments were categorised into one of three foundation damage categories comprising:
(a) either none to minor (building differential settlement (BDS) < 20 mm); (b) moderate (20 mm < BDS
< 50 mm); or (c) major (BDS > 50 mm). Chapman et al. (2015) showed that the 16 000 houses that were
assessed as having experienced major differential settlement (BDS > 50 mm) were likely to have experienced
moderate-to-severe mapped land damage and more than 200 mm of liquefaction-related subsidence. That is,
as liquefaction-related subsidence increases, the proportion of buildings with measured BDS > 50 mm
increases; conversely, the proportion of buildings with measured BDS < 25 mm decreases. This indicates the
good correlation between severity of liquefaction and foundation damage.

219

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 8.4 Liquefaction-induced damage to residential houses in Christchurch due to the 2010–2011 Canterbury
earthquake sequence

In response to the Canterbury earthquake sequence, the Ministry of Business, Innovation and Employment
(MBIE) grouped the residential lands into four different zones (MBIE, 2015): the Green zone, where repair
or rebuild process could begin; the Red zone, where land repair would be prolonged and uneconomical;
the Orange zone, where further investigation and assessment were required to reclassify it as a Red or
Green zone; and the White zone, which were still being mapped. Based on the expected future liquefaction
performance, residential lands in the Green zone were then assigned into three foundation technical categories:
TC1, TC2 and TC3. These categories are described in Table 8.1, together with the minimum recommended
vertical settlements to be used in foundation design for each technical category. Per the MBIE guidelines,
the indicator criteria for foundation damage not requiring structural repair is settlement to be <50 mm and
the ﬂoor slope to be <1/200 between any two points >2 m apart.

Table 8.1 Technical categories for rebuilding Christchurch and vertical settlement limits

Technical category Description Settlement limita: mm

TC1 Liquefaction damage is unlikely in future large earthquakes. 15

Standard residential foundation assessment and construction
is appropriate.
TC2 Liquefaction damage is possible in future large earthquakes. 50
Standard enhanced foundation repair and rebuild options in
accordance with MBIE guidance are suitable to mitigate against
this possibility.
TC3 Liquefaction damage is possible in future large earthquakes. >100
Individual engineering assessment is required to select the
appropriate foundation repair or rebuild option.

MBIE (2015)
a
Under a design serviceability limit state event

220

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Analysis and design of shallow foundations

8.3.4 2011 Tōhoku earthquake

A devastating earthquake hit the Tōhoku and Kanto regions of Japan on 11 March 2011, causing extensive
damage to life and property as a result of a large-scale tsunami and damage to nuclear power plants. The scale
of the liquefaction damage caused by the Mw 9.0 earthquake was the largest recorded anywhere in the world,
extending about 500 km along the coast from southern Iwate Prefecture to Kanagawa Prefecture (Kazama and
Noda 2012). In the Tōhoku region, typical damage seen everywhere included sand boils, ﬂoating manholes,
tilting apartment buildings, damaged housing, the collapse of levees and other manifestations of liquefaction.
However, it is considered that, in most cases, the traces of liquefaction along the coastal areas had been
washed away by the tsunami. Even in these cases, however, liquefaction was induced by the subsequent
aftershocks, so it is inferred that liquefaction was also caused by the main earthquake.

Overall, liquefaction damage was confirmed in 96 cities, towns and villages, and was concentrated along the
coast of Tokyo Bay and the basin of the Tonegawa River (MLIT, 2011). As a result of the earthquake, about
27 000 wooden houses were damaged due to liquefaction. Whereas design codes for many structures
(buildings, highway bridges, ports, etc.) have considered soil liquefaction since 1974, liquefaction has not
been accounted for in the design of wooden houses; this is the main reason why such a large number of houses
were damaged. Many houses settled and tilted, although they suffered no damage to walls and windows; some
damage to houses is shown in Figure 8.5. In the significantly tilted houses, inhabitants felt dizzy and sick, and
found it difficult to live in their houses after the earthquake. In May 2011, the Japanese government
announced a new standard for the evaluation of damage to houses based on two factors, settlement and
inclination, as shown in Table 8.2. A new class of ‘large-scale half-collapsed house’ was also introduced, and
houses tilted at angles of more than 50/1000, of 50/1000 to 16.7/1000 and of 16.7/1000 to 10/1000 were
judged to be totally collapsed, large-scale half-collapsed and half-collapsed houses, respectively, under the
new standard.

Figure 8.5 Tilted and sunk building structures due to the 2011 Tōhoku earthquake sequence (Courtesy of Japan
Home Shield)

221

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Table 8.2 New Japanese standard for evaluating level of damage to residential houses following the Tōhoku earthquake

Evaluation method Grade of damage

Total collapse Large-scale half-collapse Half-collapse Partial collapse

Inclination >50/1000 16.7/1000 to 50/1000 10/1000 to 16.7/1000 <10/1000

Settlement a
1 m to the ﬂoor level Up to the ﬂoor level 25 cm to the top of footing

Modiﬁed from Yasuda (2014)

a
Settlement of a section of a house relative to the ground surface

8.3.5 Relation between settlement and depth of the liqueﬁable layer

Based on actual case histories, the importance of the width of the building in determining the average
foundation settlement on liquefying sand is illustrated by Figure 8.6, which contains data from both the
Niigata and Luzon earthquakes. The data deﬁne a consistent band, with a clear trend for the normalised

Figure 8.6 Normalised foundation settlement versus the normalised foundation width based on available case
histories (Liu and Dobry, 1997; reproduced with permission from the publisher, ASCE)

0.4

1964 Niigata earthquake

Average foundation settlement/thickness of

(Yoshimi and Tokimatsu, 1977)

1990 Luzon Philippine earthquake
0.3 Dagupan City (Adachi et al., 1992)
liquefied soil

0.2

0.1

0
0 1 2 3 4
Building width/thickness of liquefied soil

222

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Analysis and design of shallow foundations

settlement St/DL to decrease rapidly as the normalised building width B/DL increases, with DL being the
estimated thickness of the liqueﬁed soil. Three signiﬁcant observations that can be drawn from the graph

■ building settlement is proportional to the thickness of the liqueﬁed layer

■ liquefaction-induced building settlement is typically greater than the settlement of the ground surface
in the free ﬁeld
■ foundation settlement is roughly inversely proportional to the foundation width.

8.4. Case studies: settlements of footings in physical model tests

Several researchers have used shaking table and centrifuge tests to study the seismic performance of rigid,
shallow model foundations situated atop deep, uniform deposits of saturated, loose-to-medium dense, clean
sand. These evaluation techniques are better than case history studies because the mechanisms involved in
liquefaction-induced foundation settlements can be studied in detail.

8.4.1 Shaking table tests

Yoshimi and Tokimatsu (1977) conducted shaking table tests on models of rigid structures placed on
saturated sand in order to study the pore pressure development in the sand near the structure and to study the
factors that inﬂuenced the settlement of the structure. They used a box 140 cm long, 19.75 cm wide and 40 cm
deep, with transparent walls, as illustrated in Figure 8.7(a). The bottom of the box was roughened by a layer of
coarse sand to prevent slippage, while rubber cushions were placed on both sides to allow shear deformation
of the sand during shaking. Their tests showed that (a) the excess pore pressure that developed below
the centre of the model structure was smaller than that away from the structure (Figure 8.7(b)) and (b) the
settlement of the structure decreased as the width of the structure increased for a given depth of liquefaction
(Figure 8.7(c)).

Small-scale shaking table tests in 1-G condition, however, cannot simulate every detail of a liquefaction
event in the field. Of particular importance is that because the models are limited in depth, these tests do not
realistically simulate the stress conditions in-situ. Since the stress-strain dilatancy behaviour of sand is
affected by the magnitude of confining pressure, liquefied soil behaviour in a shake table model may differ
significantly from the field response. Secondly, the shorter drainage paths in scale model result in more rapid
dissipation of excess pore water pressure than that in the field. Consequently, liquefied scale model tends to
re-solidify more rapidly, resulting in smaller settlements, when the model is subjected to the same magnitude
of earthquake shaking as in the field.

To offset the effect of low confining pressure, shaking table tests at real scale are performed. For example,
Yasuda and Berill (2000) reported a test conducted using a large-scale shaking table to investigate the
behaviour of pile foundations and raft foundations during shaking. The test set-up is shown in Figure 8.8(a).
The laminar box employed was 12 m long, 3.5 m wide and 6 m high. On the right side of the figure
is the model raft foundation, consisting of a concrete block 0.9 m wide, 1.2 m long and 1.86 t in weight.
Figure 8.8(b) illustrates the time history of the footing settlement and the excess pore water pressure. The
foundation settled 6.3 cm during shaking; since the ground surface subsided by 1.6 cm immediately after
shaking, the relative settlement of the raft foundation during shaking was 4.7 cm. In addition, water and sand
started to erupt from the ground surface about 1 min after the shaking stopped; then the footing began to settle
again in a gradual trend until a final settlement of 51 cm. At this stage, the ground subsidence was 9.5 cm,
indicating a final relative settlement of 41.5 cm. The tests showed that foundation settlement can occur not
only during shaking but also after shaking, due to the migration of excess pore water pressure – something
that was not captured by small-scale tests.

223

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 8.7 Small-scale shaking table test: (a) experimental set-up; (b) time histories of excess pore water pressures,
table motion and settlement; (c) effect of the width ratio on settlement (Yoshimi and Tokimatsu, 1977; reprinted
with permission from the author)

(Dimensions in cm)
B 19.7

Pore pressure
gauges
P1 P2 P3

40
D
10 15 15
8.5 P4 P5 P6

130 19.75

(a)

5 20
P1 0 0
Excess pore pressure ratio, u/σv0: %
2

10 Test series C
structure acceleration Excess pore pressure, u: g/cm

50
P2 0
0 Dr = 50%
10 q = 0.2 t/m2
P3 100 20
0 0 Vibration stopped
10 as liquefaction first
Settlement ratio, S/D: %

P4 20 4nL developed
0 0
10 50 Vibration continued
P5 15 beyond liquefaction to
0 0 the number of cycles
15 80 indicated
P6
q=0
0 0 10
Table

a: g

100
0
100
0
5
Settlement

1 5nL
S: mm
of

Time:s
4nL 6nL
0 1 2 3

0
P1 P2 P3 0 1 2 3
P4 P5 P6
Width ratio, B/D

(b) (c)

224

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Analysis and design of shallow foundations

Figure 8.8 Large-scale shaking table test: (a) experimental set-up; (b) time history of the foundation settlement and
the excess pore water pressure (Courtesy of Kiso-jiban Consultants Ltd.)

200 2050 1200 1600 1200 1500 1200 1050 900 900 200

470 280 200 650 180 650 200

Unit: mm
A B D21 D31 C
D11 Raft foundation
D5 D6
D8 D9 D7 D4
D12
PHC piles (fixed) D D32 Steel piles (fixed)
PHC piles (pin) 22

200 2050 1200 1600 1200 1500 1200 1050 900 900 200
D5 D2D6D3 D7
D1 D8 D4
D9
5080

Strain gauge
Shear displacement
transducer (type A)
Piezometer
Accelerometer
Displacement transducer
200 250

×3

(a)

1.0 0
Excess pore water pressure ratio

Excess pore water

Settlement of foundation: mm
Eruption of water
pressure ratio
0.8 Settlement of 10
foundation

0.6 20
Submergence of footing

0.4 30
End of settlement

0.2 40

0 50
0 50 100 150 200 250 300 350 400 450 500

Time: s

(b)

225

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

8.4.2 Centrifuge tests

Similar to the large-scale shaking table, geotechnical centrifuge equipment addresses the shortcomings
of 1g small-scale tests and allows for the correct simulation of gravity-induced stress conditions in scale
models of soil structures. Conceptually, the technique consists of increasing the confining environment
in the model soil, so that the confining stress is identical in both the model and the prototype at
homologous points. If all the pertinent scaling laws are satisfied, scale models in a centrifuge can more
accurately simulate the behaviour of full-scale earth structures. Thus, the technique allows soil lique-
faction tests to be performed at a conveniently reduced scale and provides data applicable to full-scale
problems.

Liu and Dobry (1997) performed centrifuge model experiments to investigate the mechanism of liquefaction-
induced settlement of a shallow foundation, as well as the effectiveness of sand densiﬁcation by vibro-
compaction under the footing. The tests, conducted under 16g acceleration, modelled a uniform clean sand of
Dr = 50% and of grain size ranging from very ﬁne to coarse. The test set-up using the Rensselaer Polytechnic
Institute (RPI) centrifuge apparatus for non-compacted ground is shown in Figure 8.9(a), while the time
histories are illustrated in Figure 8.9(b). Whereas all excess pore pressures measured away from the footing

Figure 8.9 RPI centrifuge tests: (a) experimental set-up; (b) time histories for acceleration, excess pore water
pressure and foundation settlement (Liu and Dobry, 1997; reproduced with permission from the publisher, ASCE)

Sf 80 g
q = 100 kPa
B = 57 mm
Ss af Key:
PF1 PC1 PE as Accelerometer
PF2 PC2 Pore pressure
PF3 PC3 Dr ≈ 52% 157 mm
transducer
ai Saturated sand
LVDT
458 mm
16g
(a)

60
0.5 af
30
Excess pore pressure: kPa

0.0 PF1 PC1 PE

–0.5 0
Acceleration: g

0.5 as 60 0 20 40
0.0 30 PF2 PC2 0.0
Settlement: m

–0.5 0 Sf
0.5 0.5
al 60
PF3 PC3 0.0
0.0 30
0.5 Ss
–0.5 0
0 5 10 0 20 40 0 20 40 0 20 40
Time: s
(b)

226

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Analysis and design of shallow foundations

reached initial liquefaction (i.e. the free field liquefied), those under the footing did not, and showed slower
pore pressure response due to the dilative behaviour of the saturated granular soil cause by the substantial
static driving shear stress that exists there due to the footing load. The excess pore pressure under the footing
at a shallow depth (PC1) was close to the free-field value of PF1 by the end of shaking; this pore pressure at
PC1 (as well as at PC2) continued to increase for a while after the shaking had stopped, suggesting water flow
took place towards PC1 and PC2, both from below and horizontally from the free field. This indicates that the
generation of excess pore pressures during shaking did not take place in an undrained condition, and that the
fast redistribution of excess pore water upwards and towards the foundation took place in a three-dimensional
pattern at all times. The footing settled a total of Sf = 0.558 m, which is more than twice the free-field
settlement Ss. It was also observed that Sf developed almost linearly with time, and essentially stopped at the
end of shaking.

Liu and Dobry (1997) also compared their centrifuge results with the empirical chart shown in Figure 8.6
representing observations in Niigata (Japan) in 1964, and Dagupan (Philippines) in 1990. They noted that the
centrifuge data points without soil compaction were generally consistent with the ﬁeld data band shown in the
ﬁgure, if building width is used for the comparison plot.

A series of dynamic centrifuge tests were undertaken by Hausler (2002) to study the effect of depth and the
lateral extent of ground improvement on the settlement and acceleration of a structure on an embedded,
shallow foundation. By examining the occurrence and timing of liquefaction in the free ﬁeld and the sequence
and degree of soil deformation under and near the structures and improved zones, Hausler (2002) reported
ﬁve settlement mechanisms

■ lateral spreading and high vertical strain of the unimproved, but not liquefied, soil under the improved
zone or structure when the adjacent free field soil liquefies
■ rapid settlement due to loss in strength of the soil below the structure or improved zone when that soil
liquefies
■ rapid settlement due to dynamic compaction in the soil below the structure or improved zone without
full liquefaction of that soil
■ permanent dilation resulting in negative (expansive) volumetric strains
■ long-term settlement due to post-liquefaction dissipation of excess pore water pressure in the soils
below the structure or improved zone.

Dashti (2009) performed a series of centrifuge experiments involving buildings with rigid mat foundations
situated atop a relatively thin deposit of liquefiable clean sand to identify the mechanisms involved in
liquefaction-induced building settlement. These experiments are described in more detail by Dashti (2009)
and Dashti et al. (2010a, 2010b). The experiments were conducted at a centrifuge acceleration of 55g.
The parameters investigated included the thickness and the relative density of the liquefiable layer and the
structural properties of the models, as well as the influence of ground motion characteristics. The plan view
and cross-section of the model used in experiment T3-30 (involving a liquefiable soil layer with a prototype
thickness of 3 m and nominal relative density Dr = 30%) are shown in Figure 8.10. Two other tests were also
conducted, namely T3-50 (with a liquefiable soil layer with a prototype thickness of 3 m and nominal relative
density Dr = 50%) and T3-50-SILT (similar to T3-50 but the 2 m-thick Monterey Sand placed on top of
liquefiable Nevada Sand was replaced by a 0.8 m-thick layer of silica flour underlying a 1.2 m-thick layer of
Monterey Sand).

227

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 8.10 Centrifuge model in experiment T3-30: (a) plan view; (b) cross-sectional view (dimensions are in metres
at the prototype scale) (Dashti et al., 2010a; reproduced with permission from the publisher, ASCE)

12
6 6

18 B 18 A 16 C 15
9 9
18

17 17
12

Shaking direction
(a)

Pore pressure transducer

Linear potentiometer
Accelerometer

Building B Building A Building C Monterey

80 KPa 80 KPa 130 KPa sand

2m 1.1 m
3m

Settlement Loose
Plates Nevada Sand
21 m
Dense Nevada
Sand

(b)

Representative excess pore water pressure– and settlement–time histories recorded in the free field when a
moderate Port Island event (peak ground acceleration scaled to 0.15g) in T3-30, T3-50-SILT and T3-50 are
shown in Figure 8.11(a). It is observed that the free-field settlements were initially quite similar for all three
cases. The re-stiffening of the silt layer in T3-50-SILT after strong shaking likely led to the slowing down of
free-field surface settlements and caused long-term heave due to water flowing from under the structures
towards the free field (Dashti et al., 2010a). Settlements continued at a higher rate in T3-50 (with no silt layer)
throughout shaking, but slowed down more rapidly compared with those measured during T3-30. The greater
tendency for horizontal flow towards the free field in T3-30 continued to supply excess pore water pressures
that dissipated upwards vertically. As a result, the volumetric straining in the free field continued for a longer
period of time. Dashti et al. (2010a) asserted that the settlements in the free field occurred predominantly
during strong shaking, which suggested that partial drainage occurred during strong shaking; therefore, the
assumption of a globally undrained loading was not valid in these experiments.

228

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Analysis and design of shallow foundations

Figure 8.11 Time histories for centrifuge experiments during the moderate Port Island event: (a) excess pore water
pressures at the mid-depth of the loose liquefiable layer and ground surface settlements recorded in the free field;
(b) excess pore water pressures at the mid-depth of the liquefiable layer under structure A and average building
vertical displacements (Dashti et al., 2010a; reproduced with permission from the publisher, ASCE)

Acceleration: g
–25 0.25
Input acceleration
0 0
Vertical displacement: mm

25 –0.25

Acceleration: g
Baseline structure – moderate Port Island event
–50 0.5

Vertical displacement: mm
50 Input acceleration
Displ. T3-50-SILT 0 0
75 50 T3-50 –0.5
100
100 Displ. T3-50 T3-50-SILT
150
125 200 T3-30
Displ. T3-30 250
150 300
50 ru = 1.0 100
pressure: kPa

ru = 1.0
Excess pore

T3-30 pressure: kPa

Excess pore

30 60
T3-50-SILT T3-30
10 20
T3-50 T3-50-SILT
T3-50
–10 –20
0 10 20 30 40 50 60 70 0 10 20 30 40
Time:s Time:s
(a) (b)

The corresponding excess pore water pressure– and settlement–time histories recorded under the baseline
structure A during this moderate Port Island event are shown in Figure 8.11(b). It can be seen that the
structures began to settle after one significant loading cycle. Building settlement was significantly larger in
T3-30 compared with the other two cases. The rates of settlement reduced dramatically and almost stopped
after the end of strong shaking (t ≈ 10–12 s) in T3-50-SILT and T3-50, while they continued at a rapidly
decreasing rate in T3-30 beyond the end of strong shaking. Dashti et al. (2010a) noted that more significant
excess pore water pressure generation and strength loss under structures within the looser soil in T3-30
(Dr = 30%) amplified key liquefaction-induced displacement mechanisms during and after strong shaking. In
addition to the higher resistance to pore water pressure generation and the smaller void space available for
volumetric densification, the greater stiffness and dilative tendency of the Dr = 50% sand likely arrested shear
strains under buildings sooner. However, Dashti et al. (2010a) noted that these observations may not apply to
buildings with larger height/width ratios and larger building inertial loads, because they may respond more
vigorously to amplified ground oscillations resulting from an increase in the relative density of the soil.

Thus, these centrifuge tests by Dashti (2009) highlighted the fact that several additional mechanisms of a localised
nature can contribute to building settlement, which can be attributed to previously unaccounted-for volumetric
strains due to localised partial drainage and diminishing effective stresses. The centrifuge test results also stressed
the importance of deviatoric strains, due to partial bearing failure or ratcheting of the structure.

229

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

To further examine the volumetric mechanism in deposits with thin liquefiable layers, Adamidis and
Madabhushi (2015) performed three centrifuge tests to examine how changes in the foundation width (B) to
liquefiable layer depth (DL) ratio affect the produced settlement mechanisms. They identified the deformation
mechanisms produced in each case through displacement vector fields corresponding to the liquefiable layer.
When B/DL = 0.4, an extended failure mechanism was mobilised, and the structure settled as the soil
underneath it lost lateral support from the fully liquefied soil of the free field. In the case when B/DL = 1, the
settlement of the structure was primarily caused by cyclic loading imposed by soil–structure interaction,
which gradually pushed out the soil from underneath the foundation. Finally, when B/DL = 2, an exacerbated
version of the second case response was observed, with displacements being even more localised. Adamidis
and Madabhushi (2015) also reported that transient hydraulic gradients and localised drainage played
important role in all three tests. They concluded that the volumetric mechanisms affecting the whole layer are
not prevalent for any of the cases investigated and, consequently, methodologies that rely on such mecha-
nisms were not appropriate. For small values of B/DL, parameters relevant to bearing capacity failure should
be considered, whereas for values of B/DL above 1, the focus should be on localised soil–structure interaction.

8.4.3 Main observations from model tests

Based on shaking table and centrifuge tests (on a shallow foundation constructed on top of a uniform loose to
medium-dense clean sand deposits) that have been carried out by Yoshimi and Tokimatsu (1977), Liu and
Dobry (1997) and Hausler (2002), one important observation was that buildings settled roughly linearly with
time during the shaking process, and generally settlement was greater than for the free ﬁeld ground.
Therefore, the inertial forces of the building were considered to be a signiﬁcant factor. Physical model tests
also showed that most of the building settlement occurred during strong shaking, with a smaller contribution
resulting from post-shaking soil reconsolidation due to excess pore water dissipation (e.g. Hausler, 2002).

While case histories from previous earthquakes show that the depth of the liquefiable layer is the dominating
factor in defining the amount of building settlements, the tests by Dashti (2009) and Adamidis and Madabhushi
(2015) challenged its validity for thin liquefiable layers. To emphasise this, Dashti et al. (2010a) incorporated the
results of their centrifuge tests in the normalised settlement versus normalised width plot shown in Figure 8.6,
and the revised plot is depicted in Figure 8.12. It is apparent than the results when the liquefiable layer is
relatively thin were not consistent with other experimental results and case history observations.

While the role of the thickness of liqueﬁable layer has been more or less clariﬁed, further research needs to be
done to investigate the effect on building settlement of the relative density of the sand, the type of seismic
motion, the aspect ratio and the bearing pressure of the building.

8.5. Factors affecting settlements of shallow foundations in liqueﬁed

ground
It is common for geotechnical engineers to assess liquefaction-induced settlement of buildings based on free-
field settlement due to liquefaction (i.e. only volumetric deformation is considered); however, as pointed out
from the case histories and model tests discussed above, this approach is deficient and can lead to inadequate
designs. The actual mechanism of settlement associated with liquefaction is much more complex. The effect
of structural inertia forces and shear deformation of the liquefied ground on the magnitude of settlement can
be very substantial in many cases. In addition to the settlement associated with volumetric strain, buildings
can experience settlement associated with shear deformation or deviatoric strain. This effect is more
prominent for tall buildings with high contact pressures, as such structures generate high shear stresses in
liquefied soil during seismic shaking.

230

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Analysis and design of shallow foundations

Figure 8.12 Comparison of the normalised foundation settlements obtained from centrifuge experiments with
available case histories and physical model tests (Dashti et al., 2010a; reproduced with permission from the
publisher, ASCE)

Upper bound 1964 Niigata and 1990 Luzon earthquakes

0.30 Lower bound 1964 Niigata and 1990 Luzon earthquakes
Yoshimi-Tokimatsu shaking table tests, 1977
Liu and Dobry centrifuge tests, 1997 – without compaction
Average foundation settlement/thickness of

Hausler centrifuge tests, 2002 – without compaction

0.25 T3-30 results
T3-50-SILT results
T6-30 results
0.20
liquefiable layer

0.15 Building A
Building C
Building C Building B
0.10
Building A

0.05 Building B
Building B

0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Building width/thickness of liquefiable layer

Based on centrifuge tests and numerical studies, Bray and Dashti (2010) and Dashti and Bray (2013)
observed that building settlement is due primarily to two mechanisms: (a) volumetric-induced settlement
mechanisms that occur throughout the deposit and (b) deviatoric soil deformations occurring near the
structure. The effects of each of these settlement mechanisms and their relative contribution to the total
building movement depend on the soil and structural properties and on the ground motion characteristics.

8.5.1 Volumetric-induced settlement mechanisms

The primary volumetric-induced settlement mechanisms are discussed below (Dashti and Bray, 2013).

8.5.1.1 Localised volumetric strains during partially drained cyclic loading

Results of case studies and centrifuge tests showed that saturated sand does not respond in a fully undrained
manner during earthquake loading, contrary to the usual assumption made. Once a hydraulic gradient is
formed, local pore water migrations start to occur. The potential for drainage is further increased with the
increase in the hydraulic conductivity of the soil associated with the liquefaction process. As a result, cyclic
shear-induced localised volumetric strains caused by migration and redistribution of pore water pressure are
expected to occur during strong shaking (Figure 8.13(a)). The contribution of this volumetric strain, referred
to as ep-DR, to the total deformation is controlled by the magnitude of three-dimensional transient hydraulic
gradients.

231

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 8.13 Liquefaction-induced displacement mechanisms: (a) volumetric strains caused by water ﬂow in
response to transient gradients; (b) partial bearing capacity failure due to soil softening; (c) soil–structure interaction
(SSI)-induced building ratcheting during earthquake loading (Dashti and Bray, 2013; reproduced with permission
from the publisher, ASCE)

Rotational failure Punching settlement

Flow

(a) (b)

Tension Tension
Compression Compression
(c)

8.5.1.2 Settlements due to sedimentation

As explained earlier, during the dissipation of the high excess pore water pressure, the settling particles
accumulate at the bottom to form a solidiﬁed zone, which increases in thickness with time and, at the same
time, consolidates under its own weight. The magnitude of sedimentation-induced volumetric strains,
referred to as ep-SED, depends on the extent of the soil skeleton breakdown beyond the point of initial
liquefaction.

8.5.1.3 Settlement due to consolidation

Consolidation primarily occurs when a continuum soil skeleton is formed (after sedimentation). As the excess
pore water pressures dissipate and the effective stress of the soil increases, permanent volumetric strains,
called ep-CON, occur as the solidiﬁed soil structure consolidates. Consolidation-induced settlements start
locally immediately after pore water pressures are generated at a given location, and achieve their highest rate
when the difference between the rate of outward ﬂow and the rate of seismically induced pore pressure
generation is greatest.

8.5.1.4 Soil expansion due to reduction of the effective stress

Although not a dominant settlement mechanism, local expansive volumetric strains (ep-EXP) may occur due to
increased pore water pressures and reduced effective stresses during shaking before the soil structure begins
to break down near liquefaction.

232

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Analysis and design of shallow foundations

8.5.1.5 Settlement due to the compressibility of pore air in partially saturated soil or to
compliance error
When the soil is partially saturated (as in the centrifuge tests performed), some volume change could
potentially occur during earthquake shaking due to the compressibility of pore air (ep-ECOMP). However, this
mechanism is not likely to contribute signiﬁcantly to liquefaction-induced settlements at building sites with
shallow water tables in the ﬁeld.

8.5.2 Deviatoric soil deformations

Bray and Dashti (2010) and Dashti and Bray (2013) also indicated that deviatoric (shear-induced) soil
deformations near a structure can be critical, especially when the shaking level is high. These deformations
depend on static driving shear stresses caused by the foundation bearing loads (a = tstatic/s′v0) and the
SSI-induced cyclic loads, as well as the soil properties. The primary deviatoric-induced settlement mecha-
nisms are explained below (Dashti and Bray, 2013).

8.5.2.1 Partial bearing capacity failure due to strength loss

Due to the loss in stiffness and strength loss in the foundation ground, punching settlement of structures or
building tilt can occur under the static load of structures, as shown in Figure 8.13(b)); this deviatoric strain is
referred to as (eq-BC). In addition, the rapid upward migration of water from the softened portion may further
reduce the strength of surﬁcial soil, generating seepage-induced liquefaction; thus, additional deformation or
bearing capacity loss can occur.

8.5.2.2 Settlements due to SSI-induced cyclic loading of the foundation

Shear-induced structural settlements (i.e. ratcheting) may occur due to SSI-induced cyclic loading near the
edges of the foundation (i.e. differential horizontal inertial forces between the foundation and the ground as
well as vertical loading pulses due to structural rocking); this is called ϵq-SSI. As shown in Figure 8.13(c), this
mechanism results in cumulative settlement of structures, and may occur with or without liquefaction.

8.5.3 Ground loss due to ejecta

During liquefaction, the pore water pressure increases and allows the saturated sediment to behave tempo-
rarily as a liquid. Pore water pressure is then relieved through the surface ejection of water and the liquefied
material (known as ejecta), facilitated by the formation of cracks that form in the overlying cap of unsaturated
sediment. The resulting surficial features come in the form of cone-shaped and oblong deposits known as sand
blows or sand boils that can be as high as a metre and tens of metres in width. If the foundation ground
underneath a structure liquefies, the soil ejecta can be carried to the ground surface, or they can be carried to
accessible buried voids (basements, buried culverts and sewers, etc.). Figure 8.14 illustrates the mechanism of
ground loss underneath a structure. Such ground loss is usually localised, and could result in differential
settlement of the structure.

8.6. Estimation of total and differential settlements

Estimation of local differential settlements as 50% of the total settlement is a commonly used rule of thumb.
That is, when estimating liquefaction-induced differential settlements of structures, the post-liquefaction-
induced reconsolidation settlement is calculated based on in situ penetration resistance data (e.g. the cone
penetration test or the standard penetration test), and then it is assumed that differential settlement between
foundation elements may be up to half of the total settlement. Hover, such ‘one-half of the total settlement’
rule of thumb has no basis whatsoever.

233

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 8.14 Ground loss underneath a structure due to ejecta

In estimating foundation settlement, many practitioners tend to make use of the widely known empirical
relationship between the normalised footing width and the normalised foundation settlement, as
shown in Figure 8.6. However, as the centrifuge tests showed, there are many mechanisms involved in
liquefaction-induced foundation settlements, and such normalisation does not work for cases involving
relatively thin layers of liqueﬁable soil. Therefore, this type of normalisation should not be used for this
situation.

Ishii and Tokimatsu (1988) suggested that the total footing settlement during earthquake shaking S is the sum
of the settlement due to volumetric strain caused by the shaking Sv and the immediate settlement due to
change in soil modulus Se, so

S = Sv + Se (8:2)

1 1
Se = qBIp − (8:3)
E2 E1

where q is the contact pressure of the structure; B is the width of the structure; Ip is a coefﬁcient relating to
the dimension of the structure, thickness of the soil layer and Poisson’s ratio of soil; and E1 and E2 are the
Young’s moduli of soil before and during earthquake shaking, respectively. For liquefaction-induced
settlements, Sv is the post-liquefaction-induced reconsolidation settlement (calculated using the methods
discussed in Chapter 7); Se can also be obtained using Equation 8.3, but E2 should be evaluated con-
sidering the change in the effective stress due to excess pore water pressure generation as well as the shear
strain level developed in the soil. However, when the severity of liquefaction is high, the settlement of the
structure is affected by the shear deformation of the ground, and E2 is very difﬁcult to determine accu-

234

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Analysis and design of shallow foundations

rately; thus, Equation 8.2 cannot be used. Instead, Ishii and Tokimatsu (1988) proposed calculating the
settlement as

S = Sv rb (8:4)
where rb is a scaling factor that takes into account shear deformation and can be obtained from Figure 8.15,
which is based on the observations following the Niigata earthquake. They noted that structures with a large
width (compared with the thickness of the liqueﬁed layer) underwent reduced liquefaction-induced settle-
ment; that is, appreciable settlement occurred when the width ratio was <2 whereas the settlement was small
when the width ratio was >3. Thus, they deﬁned the parameter rb as equal to the settlement ratio normalised by
the settlement ratio at a width ratio equal to 3.

For light structures (one- to three-storey buildings) founded on a non-liquefiable cohesive crust over
liquefied soil, Naesgaard et al. (1998), based on limited total-stress dynamic numerical analyses, proposed
a chart to estimate settlement associated with the shear strain deformations of the ground. The chart,
showing shear strain-induced settlement versus the post-liquefaction bearing capacity safety factor, is
provided in Figure 8.16. Note that this settlement needs to be added to the post-liquefaction free-field
consolidation settlement to get the total footing settlement.

Some empirical methods based on centrifuge tests were also proposed. For example, Kawasaki et al. (1998)
conducted 30 centrifuge tests on footings for a power transmission tower, and proposed estimating the footing
settlement S as

S = S0 C1 C2 C3 C4 C5 D1 D2 (8:5)

where the settlement under particular standard conditions S0 is modified by the following factors: C1 for the
thickness of the liquefiable layer, C2 for the thickness of the non-liquefiable layer, C3 for the density of the

Figure 8.15 Estimation of the scaling factor as a function of the width ratio (Ishii and Tokimatsu, 1988)

0.3 15

. Pile foundation
Shallow foundation
Scaling factor for

.
Settlement ratio S/D

0.2 10
the width rb

1–2 storey
3 storey .
4 storey .
. .
5–6 storey
.
0.1 . 5
With basement
.. . .
..
..
. . .
1
0
0
0 1 2 3 4 5
Width ratio, B/D

235

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 8.16 Correlation between post-liquefaction shear-induced settlement and post-liquefaction bearing
capacity factor of safety (Naesgaard et al., 1998; reproduced with permission from the publisher, ASCE)

Post-liquefaction bearing capacity factor of safety, Fs

0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
0
Post-liquefaction shear induced

200
footing settlement: mm

400

600

800 Shear induced settlement (mm) = 816 × Fs–1.9

1000

1200

ground, C4 for the grain size, C5 for the degree of liquefaction (in terms of FL), D1 for the width of the footing
and D2 for the load intensity. Each of these factors is shown in Figure 8.17. In the tests, the conditions
set as standard were C1 = 13.2 m, C2 = 0 m, C3 = 40%, C4 with D50 = 0.161 mm (Toyoura sand), C5 with
FL = 1, D1 = 3.2 m and D2 = 28.4 kPa.

While this approach attempts to incorporate various factors affecting settlement, it has not been validated
using case histories.

Alternatively, numerical methods can also be used to design foundation systems for buildings founded on
liquefiable ground. A simplified finite-element-method approach for the evaluation of liquefaction-induced
displacements has been proposed by Yasuda et al. (1999). Essentially, it involves two stages of analyses:
(a) deformations are first calculated for the static loads applied to the foundation, using the initial (pre-
liquefaction) shear moduli of the subsoil layers; and (b) the shear moduli are consequently reduced, in order to
account for the effects of liquefaction, and the deformation is calculated considering the stresses obtained in
the first stage. The liquefaction-induced deformations are then calculated as the difference between the two
stages. Such an approach has been implemented successfully in a number of case studies (e.g. Yasuda, 2005;
Yasuda et al., 2001); however, it does not provide immediate insight to all factors (as mentioned above) that
control the liquefaction performance of footings.

With better constitutive models to simulate soil liquefaction, numerical (e.g. finite-element or finite-
difference) models can describe complex soil profiles and SSI effects during soil liquefaction. These have
been used in many routine geotechnical projects to analyse and design foundations under various conditions.
Numerical simulations have also been used to evaluate the basic relationships between the magnitude of
building settlement and various factors, such as the footing width and the thickness of the liquefiable layer
(e.g. Dashti and Bray, 2013; Elgamal et al., 2005; Popescu and Prevost, 1993; Shahir and Pak, 2010).

Bray and Macedo (2017) conducted hundreds of nonlinear effective stress fully coupled soil-structure
interaction dynamic analyses to identify the key parameters controlling liquefaction-induced settlement for
buildings with shallow foundations and used the results to developed a simpliﬁed procedure for estimating
liquefaction-induced building settlement. For further details, readers are recommended to refer to their paper.

236

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Figure 8.17 Factors used to modify settlements for standard conditions (data courtesy of S. Yasuda)

2.0 2.0 2.0 2.0

C1 C2 C3 C4
1.5 1.5 1.5 1.5

1.0 1.0 1.0 1.0

Factor
0.5 0.5 0.5 0.5

0.0 0.0 0.0 0.0

0 2 4 6 8 10 12 0 2 4 6 0 10 20 30 40 50 60 0 0.1 0.2 0.3 0.4 0.5
Liquefiable layer thickness Unliquefiable layer thickness: m Relative density: % D50: mm
–½ footing width: m

2.0 2.0 2.0

C5 D1 D2
1.5 1.5 1.5

1.0 1.0 1.0

Factor
0.5 0.5 0.5

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
0.0 0.0 0.0
0 0.5 1 1.5 2 0 1 2 3 4 5 –20 –10 0 10 20 30 40 50 60
Factor of safety, FL Footing width: m Additional pressure: kPa
Analysis and design of shallow foundations

237
Seismic Design of Foundations

8.7. Seismic design of mat foundations

Mat foundations are generally used in the following cases: (a) where individual footings would cover a large
proportion of the building footprint; (b) where the subsurface foundation soils are variable or contain pockets
of soft or liquefiable soils; or (c) if resistance to hydrostatic uplift is necessary (e.g. basement floors). They are
used to distribute concentrated column and wall loads across the entire building area to lower the contact
pressure (compared with conventional spread footings) and to equalise differential settlements for variable
ground conditions. Raft foundations are a special case of mat foundations that are sufficiently stiff and strong
to distribute the entire superstructure load uniformly across the base and to behave essentially as a rigid unit.

In the design of mat and raft foundations, the bearing capacity, settlement and uplift pressure are routinely
considered. Generally, bearing failure is unlikely unless the mat or raft is heavily loaded close to the edge
such that a bearing failure underneath the edge would occur, leading to structural failure. However, care is
required for cases where mats are founded on liqueﬁable soils; during earthquake-induced liquefaction, the
seismic behaviour of the soil supporting the mat foundation will alter the overall performance of the foun-
dation. This often results in the tilting of the structure as a whole with the superstructure remaining essentially
intact. Thus, mat and rigid raft foundations must be adequately designed against differential settlements.

8.8. Summary
Shallow foundations, such as footings and rafts, are typically designed to take into account the bearing
capacity and settlement criteria. However, when considering building failures during earthquakes, the
number of structures damaged by earthquake-induced settlements is far higher than those that have
earthquake-induced bearing capacity failures; thus, the settlement criterion generally governs.

When building on top of liquefiable ground, it is important to clarify that liquefaction-induced settlements
of footings cannot be estimated using empirical methodologies developed for free-field conditions, as
the controlling mechanisms for the two events are different. That is, free-field settlements consider only
volumetric-induced strains, and consequently they take place during the dissipation of earthquake-induced
pore pressures, mostly after the shaking. In contrast, footing settlements include not only volumetric-induced
strains but also deviatoric-induced deformations.

However, while the settlement-generating mechanisms have been conceptually identified through a number
of experimental studies and field observations, their case-specific significance remains unclear. The contri-
bution of each mechanism to the overall settlement depends on many factors, such as the characteristics of the
earthquake motion, the liquefiable soil and the structure. Thus, well-calibrated analytical tools and design
procedures that identify, evaluate and mitigate the most critical mechanisms of liquefaction-induced settle-
ment are needed.

REFERENCES
Adachi T, Iwai S, Yasui M and Sato Y (1992) Settlement of inclination of reinforced concrete buildings in Dagupan
city due to liquefaction during 1990 Philippine earthquake. Proceedings of the 10th World Conference on
Earthquake Engineering, Madrid, Spain, pp. 147–152.
Adamidis O and Madabhushi GSP (2015) Deformation mechanisms under shallow foundations during earthquake-
induced liquefaction. Proceedings of the 6th International Conference on Earthquake Geotechnical Engineer-
ing, Christchurch, New Zealand.
Bray JD and Dashti S (2010) Liquefaction-induced movements of buildings with shallow foundations. 5th Inter-
national Conference on Recent Advances in Geotechnical Earthquake Engineering And Soil Dynamics, San
Diego, CA, USA, paper OSP-2.

238

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Analysis and design of shallow foundations

Bray, J.D. and Macedo, J. (2017) 6th Ishihara Lecture: Simplified Procedure for Estimating Liquefaction-Induced
Building Settlement, Soil Dynamics and Earthquake Engineering, 102: 215–231
Brinch Hansen JA (1970) Revised and Extended Formula for Bearing Capacity. Danish Geotechnical Institute,
Copenhagen, Denmark.
Budhu M and Al-Karni A (1993) Seismic bearing capacity of soils. Géotechnique 43(1): 181–187.
CEN (Comité Européen de Normalisation) (2004) EN 1998-5. Design of structures for earthquake resistance.
Part 5: Foundations, retaining structures and geotechnical aspects CEN, Brussels, Belgium.
Chapman L, van Ballegooy SGA and Lacrosse V (2015) Correlation of differential building settlement with
predicted CPT-based liquefaction vulnerability parameters. Proceedings of the 6th International Conference in
Earthquake Geotechnical Engineering, Christchurch, New Zealand.
Dashti S (2009) Toward Evaluating Building Performance on Softened Ground. PhD thesis, University of
California, Berkeley, CA, USA.
Dashti S and Bray JD (2013) Numerical simulation of building response on liquefiable sand. Journal of
Geotechnical and Geoenvironmental Engineering, ASCE 139(8): 1235–1249.
Dashti S, Bray JD, Pestana JM, Riemer MR and Wilson D (2010a) Mechanisms of seismically-induced settlement
of buildings with shallow foundations on liquefiable soil. Journal of Geotechnical and Geoenvironmental
Engineering, ASCE 136(1): 151–164.
Dashti S, Bray JD, Pestana JM, Riemer MR and Wilson D (2010b) Centrifuge testing to evaluate and mitigate
liquefaction-induced building settlement mechanisms. Journal of Geotechnical and Geoenvironmental
Engineering, ASCE 136(7): 918–929.
Dormieux L, and Pecker A (1995) Seismic bearing capacity of foundation on cohesionless soil. Journal of
Geotechnical Engineering 121(3): 300–303.
Elgamal A, Lu J and Yang Z (2005) Liquefaction-induced settlement of shallow foundations and remediation:
3D numerical simulation. Journal of Earthquake Engineering 9: 17–45 (special issue 1).
Hausler EA (2002) Influence of Ground Improvement on Settlement and Liquefaction: A Study Based on Field
Case History Evidence and Dynamic Geotechnical Centrifuge Tests. PhD thesis, University of California,
Berkeley, CA, USA.
Ishii Y and Tokimatsu K (1988) Simplified procedures for the evaluation of settlements of structures during
earthquakes. Proceedings of 9th World Conference on Earthquake Engineering, Tokyo-Kyoto, Japan, vol. 3,
pp. 95–100.
JGS (Japanese Geotechnical Society) (2004) Photographs and Motion Picture of Niigata City Immediately after
the Earthquake in 1964. JGS, Tokyo, Japan. (CD-ROM.)
Kawasaki K, Sakai T, Yasuda S and Satoh M (1998) Earthquake-induced settlement of an isolated footing for
power transmission tower. Centrifuge 98: 271–276.
Kazama M and Noda T (2012) Damage statistics (summary of the 2011 off the Pacific Coast of Tōhoku earthquake
damage). Soils and Foundations 52(5): 780–792.
Kishida H (1966) Damage to reinforced concrete buildings in Niigata city with special reference to foundation
engineering. Soils Foundations 6(1): 71–88.
Liu L and Dobry R (1997) Seismic response of shallow foundation on liquefiable sand. Journal of Geotechnical
and Geoenvironmental Engineering, ASCE 123(6): 557–567.
MBIE (Ministry of Business, Innovation and Employment) (2015) Guidance on Repairing and Rebuilding Houses
Affected by the Canterbury Earthquakes. MBIE, Wellington, New Zealand.
Meyerhof GG (1951) The ultimate bearing capacity of foundations. Géotechnique 2: 301–332.
MLIT (Ministry of Land, Infrastructure, Transport and Tourism) (2011) Liquefaction Countermeasure Technology
Investigation Committee Investigation Results. MLIT, Tokyo, Japan. See http://www.mlit.go.jp/common/
000169750.pdf (accessed 16/08/2018). (In Japanese.)

239

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Naesgaard E, Byrne PM and Ven Huizen G (1998) Behaviour of light structures founded on soil crust over
liquefied ground. Proceedings of Geotechnical Earthquake Engineering and Soil Dynamics III. American
Society of Civil Engineers, Reston, VA, USA.
Ohsaki Y (1966) Niigata earthquake, 1964 building damage and soil condition. Soils and Foundations 6(2): 14–37.
Orense R, Towhata I and Ishihara K (1991) Soil liquefaction in Dagupan City during the 1990 Luzon, Philippines
earthquake. Proceedings of the 26th Japan National Conference on Soil Mechanics and Foundation
Engineering, Nagano, Japan, pp. 871–874.
Orense R, Pender M and Wotherspoon L (2012) Analysis of soil liquefaction during the recent Canterbury
(New Zealand) earthquakes. Geotechnical Engineering Journal of the SEAGS & AGSSEA 43(2): 8–17.
Popescu R and Prevost JH (1993) Centrifuge validation of a numerical model for dynamic soil liquefaction.
Soil Dynamics and Earthquake Engineering 12(2): 73–90.
Sarma SK and Iossifelis IS (1990) Seismic bearing capacity factors of shallow strip footings. Géotechnique 40(2):
265–273.
Seed HB and Idriss IM (1967) Analysis of soil liquefaction: Niigata earthquake. Journal of Soil Mechanics and
Foundations 93(3): 83–108.
Shahir H and Pak A (2010) Estimating liquefaction-induced settlement of shallow foundations by numerical
approach. Computers and Geotechnics 37(3): 267–279.
Terzaghi K (1943) Theoretical Soil Mechanics Wiley, New York, NY, USA.
Tokimatsu K, Midorikawa S, Tamura S, Kuwayama S and Abe A (1991) Preliminary report on the geotechnical
aspects of the Philippine earthquake of July 16, 1990. Proceedings of the 2nd International Conference on
Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics, St Louis, MO, USA, pp. 357–364.
Yasuda S (2005) Recent several studies and codes on performance-based design for liquefaction in Japan.
Proceedings of Geotechnical Earthquake Engineering Satellite Conference, Osaka, Japan.
Yasuda S (2014) New liquefaction countermeasures for wooden houses. Proceedings of the, Soil Liquefaction
During Recent Large-scale Earthquakes (Orense RP, Towhata I and Chouw N (eds)), CRC Press, Boca Ratan,
FL, USA, pp. 167–180.
Yasuda S and Berrill JB (2000) Observations of the earthquake response of foundations in soil profiles containing
saturated sands. Proceedings of GeoEng 2000 – An International Conference on Geotechnical and Geological
Engineering, Melbourne, Australia, pp. 1441–1470.
Yasuda S, Yoshida N, Adachi K, Kiku H, Gose S and Masuda T (1999) Simplified practical method for evaluating
liquefaction-induced flow. Journal of Geotechnical Engineering, JSCE 638(III-49): 71–89.
Yasuda S, Yoshida N, Kiku H, Abo H and Uda M (2001) Analyses of liquefaction-induced deformation of grounds
and structures by a simple method. Proceedings of the 4th International Conference on Recent Advances in
Geotechnical Earthquake Engineering and Soil Dynamics, San Diego, CA, USA, paper 4.34.
Yoshimi Y and Tokimatsu K (1977) Settlement of buildings on saturated sand during earthquakes. Soils and
Foundations 17(1): 23–38.

240

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

ISBN 978-0-7277-6166-8
https://doi.org/10.1680/sdof.61668.241
ICE Publishing: All rights reserved

Chapter 9
Pile foundations
9.1. Introduction
This chapter

■ summarises the plausible failure mechanisms of pile-supported structures during earthquakes in

different types of ground
■ highlights the different design calculations that engineers need to perform while designing pile
foundations
■ provides illustrative examples to show the application of the methodology.

9.2. Why piled foundation are used

Piles are a particular type of deep foundation generally chosen for use at a construction site where the soils at
shallow depths are weak and therefore have very low bearing capacity. It is not an overstatement to say that
piles are used in most heavily loaded structures such as multi-storey buildings, bridges, ﬂyovers and oil
platforms. Piles are also used in seismic-prone zones having loose to medium-dense sandy soil or soft clay.
As noted in Chapters 6 and 7, if loose to medium-dense sands are saturated due to a shallow water table, they
lose their strength and stiffness during earthquake shaking – a process is known as liquefaction. In other
words, these soil deposits momentarily become a ﬂuid-like material, affecting the foundations.

9.3. Pile foundations in seismic areas

Engineered pile-supported structures still collapse after most major earthquakes despite the large factors of
safety employed in their design. Post-earthquake surveys often show that the superstructure (i.e. the part of
the structure above the ground) is intact/undamaged but has tilted or rotated as a whole – rendering it useless.
This suggests that the foundations may have been damaged. Figure 9.1 shows a pile-supported building
where the building has tilted. Note that these pile-supported buildings tilt irrespective of the type of ground
(i.e. level ground where no lateral spreading was observed and laterally spreading ground), see Bhattacharya
and Madabhushi (2008).

9.3.1 Observed pile damage from detailed ﬁeld investigations

As earthquakes are very rapid events and as much of the damage to piles occurs beneath the ground, it is hard
to ascertain the failure mechanism unless deep excavations are carried out. Twenty years after the 1964
Niigata earthquake (Japan) and also following the 1995 Kobe earthquake (Japan), investigations were carried
out to ﬁnd the failure pattern of the piles (BTL Committee, 2000; Yoshida and Hamada, 1990). Piles were
excavated or extracted from the subsoil, borehole cameras were used to take photographs and pile integrity
tests were carried out. These studies hinted at the location of cracks and damage patterns for the piles. Of
particular interest is the formation of plastic hinges in the piles. This indicates that the stresses in the pile
during the earthquakes exceeded the yield stress of the pile material. As a result, the design of pile foundations
in seismically liqueﬁable areas continues to be of interest to the earthquake geotechnical engineering
community.

241

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 9.1 Collapse of a pile-supported building. Kandla Port Tower near the Arabian Sea and the damage
following the 2001 Bhuj earthquake (India) where lateral spreading was observed

The main observations from these studies were as follows.

■ The failure of foundations not only occurred in laterally spreading ground but was also observed in
level ground where no lateral spreading would be anticipated (see Figure 9.1).
■ The failure patterns of buildings on level ground and on laterally spreading ground were similar; that
is, the buildings tilted considerably. In most cases, the superstructure remained undamaged. Therefore,
considering building failure alone, it is difficult to ascertain the type of associated ground (i.e. whether
lateral spreading or level).
■ Figure 9.2 shows several examples of pile damage, together with the soil profiles and the damage
pattern. Cracks were observed near the bottom and top boundaries between the liquefied and
non-liquefied layers. Often, cracks were observed at the pile head.
■ Plastic hinges also formed not only at the boundaries of the liquefiable and non-liquefiable layers but
also at various depths.
■ Horizontal displacement of the piles was measured in some cases of pile failure and, in some,
the displacement of the piles was found to agree with the horizontal displacement of the ground.
■ There were few cases where plastic hinges formed at the middle of the liquefied layer (e.g. see
Figure 9.2(b)). These forms of damage were observed when the structure was very close to a quay
wall that moved significantly towards the sea.

242

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

■ Large-diameter piles performed well during the earthquakes. For example, while jetties in Kandla Port
that were supported on 0.5 m-diameter reinforced concrete piles suffered cracking during the 2001
Bhuj earthquake (India), the 1.0 m diameter reinforced concrete piles on newer jetties and concrete-
filled-steel tubular piles in the same area performed well. However, it was observed following the
2011 Tō hoku earthquake (Japan) that fewer large-diameter piles (1–2 m diameter and >50 m long)
performed very well (Bhattacharya et al., 2011), rather than many small diameter piles
■ It is clear that the same kinds of failures are being observed not only in developing countries
(e.g. India, see Figure 9.1) but also in first-world countries (e.g. Japan). As it can be argued that
construction in first-world countries maintains a higher degree of quality control in design and
construction, this is strong evidence that the correct failure mechanism(s) governing pile failure have
not been properly accounted for during design.
■ As plastic hinges form, this suggests that the bending moments or shear forces that are experienced
by the piles exceed the bending or shear-carrying capacity of the pile section. As earthquakes are
dynamic in nature and piles carry large vertical loads, the static and dynamic instability of pile-
supported structures needs to be studied.

Figure 9.2 (a) Building in Niigata city. (b) Building in Niigata city (Kawamura et al., 1984). (c) Building in Kobe city
(Fujii et al., 1998). (d) A Governmental building, Kobe city (Editorial Committee for the Report on the Hanshin-Awaji
Earthquake Disaster, 1998) [Photo courtesy: Prof Yoshida, Dr T. Tazoh (Late)]

(a) (b) SPT N value

Depth: m SPT N value 0 10 20 30
Depth: m GL – m

1700
0 0 0 10 20 30 40 50
GL ± 0 1 7
2 9
1 1 2000~3000 4500~6500 2500~3500
3 8
2 2 5 4 7
5 13
3 3
Sand 6 7
≈11 000

φ ≈ 10–15° 7
Depth: m

4 4 9
10
8 10
5 5 9 12
10 8
6 6 15 No. 1
No. 2 11 19
7 7 12 23
400

13 19
8 8
20 14 25
1000~2000
Unit: mm
No. 1 No. 2 No.3 Depth:
(c)
Pile cap Soil type SPT N value
0 Pile cap Pile cap m
0 50
0

–5 –8.50 to Reclaimed
–9.17 to –5
–7.30 m –8.90 m –9.39 m
fill
end of Large cracks large cracks
investigation with concrete with concrete –10
Depth: m

–10 failure failure

–13.00 m
Silt GL, ground level
–13.00 m –15 clay
pile joint
–15
pile joint SPT, standard penetration test
–20

–20 –20.80 m –25

end of Sand
investigation –22.90 m
end of
–30 and
–25 investigation clay
–35

–30 –40

243

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 9.2 (Continued)

(d) SPT N value

0 10 20 30 40 50 60
0
0.85
1
7

3.17
1.85
2
2.85 11
3 B1F
Depth: m
3
4 16

2.40
Pit
4.85
5 9
5.65 Base
6 Crack
9
7
11 Investigation hole
7.75
8
8.65 6 φ = 1100
9
3.6
9.70
10 10.15
20
Crack
11
11.60 38
12 12.20
68
12.80
13
15
13.85
14
41
14.95
15
54
16
69
17
45
17.80
18 18.25
48
18.90
19
94
19.85
20 20.20
40
20.75
21 21.15
21.65 13
22
22.55 20
23
23.60 300

244

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Aside. For standard buildings in non-seismic areas (e.g. London, UK), the main predominant load carried
by piles is vertical, although horizontal loads do arise due to eccentricity of loading or wind loading and so
on. In contrast, in seismic zones during earthquakes, owing to the inertia of a building (i.e. the motion of
the building), large horizontal loads are experienced by the piles. Pile are therefore designed as beams
supported by the ground – in engineering terms, ‘beams on non-linear Winkler foundations’, see Chapter 5
(Section 5.8). However, if a substantial part of a pile is unsupported, due to a liquefaction-type phenomenon
or any other process, the pile behaves as an unsupported column. As an approximation, if the pile is founded
in liquefiable soils, it is may be more appropriate to treat the pile as a long unsupported column in the
liquefiable zone. This will therefore be a stiffness-dominated design. On the other hand, for piles in non-
liquefiable areas where the soil significantly supports the pile even under seismically induced large strains,
beam approximations may be appropriate. It is safe to use a generalised beam–column formulation.

9.4. Failure mechanisms of pile-supported structures

From a practical point of view, failure mechanisms can be classiﬁed into two broad types

■ Piles pass through layered soil (a non-liquefiable base layer at a deeper depth, a liquefiable layer and a
non-liquefiable crust above the liquefiable layer) with or without lateral spreading. Figure 9.3 shows
four possibilities. Case I refers to where there is a large non-liquefiable crust that may potentially
slide following subsurface liquefaction. Cases II and III are typical of a river bridge pier where the
founding level of the pile is in either a non-liquefiable bearing layer or a liquefiable layer. Further
details on bridge issues are dealt in Chapter 10. Case IV is a case where the top soil is non-liquefiable
(a clay layer) and the pile rests in the liquefiable layer. It is shown later that the building in Figure 9.1 is
in this category.
■ Piles pass through layered soils where none of the layers may liquefy (Figure 9.4). In some cases, the
soil layers may have a sharp interface having high stiffness contrast, and large kinematic moments
may develop at the interface due to wave passage effects. The location of kinematic bending moments
may be different from the location of the maximum inertial moments. This design scenario is mentioned
in the Eurocode 8 (CEN, 2004) and discussed later in this chapter.

Figure 9.3 Different conﬁgurations of pile foundations in liqueﬁable soils

Case I Case II Case III Case IV

Non-liquefied crust Non-liquefied base layer

LIquefied soil Soil flow

245

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 9.4 Liqueﬁed natural gas (LNG) tanks supported on piles

Soft soil
Soft soil
Vs1 Location of maximum
kinematic bending
moment (interface of
Stiff soil soil layer having high Stiff soil
Vs2 stiffness contrast)

Hard soil Hard soil

9.4.1 Pile behaviour in liqueﬁable soils

Predicting the behaviour of pile foundations under earthquake loading is a complex problem as it involves
consideration of strong motions, the free-field site response, superstructure behaviour, dynamic character-
istics of the system as the earthquake progresses and soil–pile–superstructure interaction. Significant research
has been done in the area of soil–pile–superstructure interaction in liquefiable soil, and this section provides a
commentary on the loading on a pile as an earthquake progresses.

Figure 9.5(a) considers four critical stages of loading to the piles during the event. The loadings at these
discrete stages are as follows.

■ Stage I. Before, or just at the onset of seismic shaking, the piles are predominantly subjected to a static
axial load, denoted by Pgravity. The pile section is expected to have axial stresses. Pgravity (stage I)
represents the axial load on the piles in the normal condition, and this axial compressive load may
increase/decrease further due to the inertial effect of the superstructure (Vinertial), as shown in stage II.
Pstatic can be obtained from the weight of the superstructure (i.e. simply dividing the weight by the
number of piles). Vinertial and Hinertial can be estimated from the rocking motion of the superstructure,
which will be transferred to the piles by a push–pull action (i.e. upwards tensile force on one side and
downwards compressive force on other side).
■ Stage II. The pile will be subjected to inertial forces due to vibrations of the superstructure. The
inertial load (Hinertial) is dependent on the time period of the structure as well as on the strong motion
characteristics. As a simplification, the inertial load can be estimated based on response spectra method,
as explained in Chapter 5, see Section 5.5. The axial load will continue to act on the pile. There will be
additional vertical load on some piles due to inertia (Vinertial). However, the increase in the dynamic axial
load will be insignificant, as the length of the laterally unsupported pile is not high. The bending moment
in the pile can be estimated using the beam on a non-linear Winkler foundation model, see Section 5.8.
■ Stage III. In this stage the soil has liquefied, and the pile has lost its axial and lateral resistance from
the surrounding liquefied soil and acts like a long laterally unsupported column. The fundamental
frequency of the structure and the damping of the soil have changed significantly. The inertial force
on the pile will now be a function of a new fundamental period and damping of the structural system.
There will also be a substantial increase in the dynamic axial load on the pile. If the embedment of the
pile is loss in the non-liquefiable layer below the liquefiable layer, the structure may settle.

246

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Figure 9.5 (a) Stages of loading on a piled foundation. (b) Changes in the modal parameters of a pile-supported
structure

Pgravity Pgravity + Vinertial Pgravity + Vinertial Pgravity + Vinertial

Hinertial Hinertial Hinertial

Loose Liquefied Liquefied

sand sand sand

Stage I Stage II Stage III Stage IV

Before Shaking starts. Soil liquefies. On sloping ground
earthquake on Soil yet to Vertical inertial soil liquefies.
level ground liquefy forces act with Lateral spreading may
gravity. Piles may combine with behaviour
starts to buckle or in stage III
settle

(a)

Before liquefaction After liquefaction

Feq Feq_liq
h1 h2

Liquefiable layer
Liquefied Df_liq

Adyn
Stiff layer Stiff layer

Adyn_liq

Feq Equivalent lateral force Feq > Feq_liq

Adyn Axial dynamic load Effects due to liquefaction Adyn < Adyn_liq
Df Depth of fixity Df < Df_liq
(b)

247

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

■ Stage IV. In some areas, due to the lateral ﬂow of soils, piles will be subjected to passive soil
pressure. A static axial load will continue to act on the pile foundation. Depending on the duration
of the earthquake, there may be an inertial load acting on the foundation.

Apart from the loading, the modal parameters – the dynamic properties of the building (ﬁrst resonant frequency
and damping) – in stage II will be very different from stage III. In most cases, the time period of the building will
increase few fold during the transition from stage II to stage III. Damping of the structure will also increase, and
in some instances can go up to 20%. Figure 9.5(b) shows a simple model to capture the phenomenon, and is
described in more detail later in this chapter. The theory is developed in Lombardi and Bhattacharya (2014).

Figure 9.6 shows the time history of the various mechanisms/processes not only in the structure and but also in
the supporting soil during an earthquake. The time history shows the typical changes in frequency of the
structure–pile–soil system, the depth of liquefaction, the soil stiffness, the soil damping and ground movement
due to ground shaking (i.e. from the onset of the strong motion to the end of the earthquake and ﬁnally dis-
sipation of the excess pore water pressure). The diagram thus depicts the cause of an event (i.e. earthquake) and
its effects (i.e. ground deformation, excess pore water pressure) and the consequences on soil properties and
also on pile-supported structures. The values presented are indicative and not typical, and should not be used for
calculations. One of the aims of this chapter is to provide engineering procedures to calculate these values.

Figure 9.6 Time history of pile–soil interaction during an earthquake

f = f0

f = f0/2
f = f0/5

Natural frequency
of the structure

Depth of
liquefaction
ξ=0 ξ = 15%
ξ = 2%
Soil damping

k = 0.7k0
k = k0
k = 0.1 to 0.01k0
Soil stiffness

Excess pore water pressure

(at a particular depth)
∆ = ∆max
∆=0
Ground deformation

Ground acceleration

Time

248

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

9.4.1.1 Aside: dynamic ampliﬁcation of responses – tuning of a structure and earthquake

motion as the soil liquefies
Figure 9.7 shows the different stages of loading of a pile-supported structure during a seismic liquefaction-
induced event. With the onset of liquefaction, that is, with pore water pressure build-up (at full liquefaction,
the excess pore water pressures reach the overburden vertical effective stress), the soil loses its strength and
stiffness. This may take few seconds, and for the example shown in Figure 9.7 the time required is just over
5 s. In these 5 s, the time period of the structure may increase a few fold. For example, Bhattacharya (2006)
implication estimated that the time period of a building increased from 0.5 to about 5 s. While the structure
transits from 0.5 to 5 s, it may get tuned with some frequencies of the earthquake. Experimental studies
carried out by Lombardi (2014) and Rouholamin et al. (2017) suggest that there will be amplification of
bending moments. The amplification depends on the time taken to reach full liquefaction (i.e. the speed at
which the ground liquefies). Therefore, the design methods need to consider that a pile acts as an unsupported
column over the liquefied depth. A pile that has a high slenderness ratio will be prone to buckling instability,
which will also be amplified by imperfections in the pile, lateral loads and the dynamics of the earthquake.

Figure 9.7 Loading conditions acting on a typical pile-supported structure subjected to seismic induced
liquefaction. EPWPR (Excess Pore Water Pressure Ratio)

Before Before At full

earthquake liquefaction liquefaction

Axial Axial Axial

Inertia Inertia

Liquefiable
soil

Non-liquefiable
soil

(a) (b) (c)

Input motion Express pore water

Time to reach
fully liquefaction
acceleration: g pressure ratio

1.2
0.8
0.4
0
0 5 10 15 20 25 30 35 40 45 50
1.0
0.5
0
–0.5
–1.0
0 5 10 15 20 25 30 35 40 45 50
Time: s
(d)

249

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

9.4.1.2 Mathematical treatment of instability in piles

Static instability (i.e. Euler type) is critical for the pile foundation when the soil remains liqueﬁed and the
depth of liquefaction is high enough. On the other hand, dynamic instability is critical, when the liquefaction
starts and progresses in such a way that the fundamental frequency of the pile–soil system may come close to
the dominant earthquake frequency that may give a worse response for the pile. These are explained below.

Static instability. The pile in an elastic soil can be considered as a Euler–Bernoulli beam having a bending
stiffness of EI (where E is Young’s modulus and I is the moment of inertia) and resting against a linear
uniform elastic support of stiffness k. A constant static lateral force F is applied at the top of the beam (x = L,
where L is the pile length). The well-known equation of static equilibrium can be expressed, following
Figure 9.8(a) as

∂4 wð xÞ ∂2 wðxÞ
EI +P + kwð xÞ = 0 (9:1)
∂x4 ∂x2

where w(x) is the transverse deflection of the beam, x is the spatial coordinate along the length of the beam
and P is the axial load. It is assumed that the mechanical properties of the beam are constant along the
length. Equation 9.1 is a fourth-order partial differential equation, and requires four boundary conditions for
its solution. To comment on the pile head deflection, it is useful to find the deflection at the top of the beam D
in terms of k (soil support), P and EI. The Euler critical load of the cantilever beam is denoted by Pcr.

Let D0 be the deﬂection of the beam due to the lateral force F without the inﬂuence of the axial load and the
support stiffness k. From basic mechanics

FL3
D0 = (9:2)
3EI

Equation 9.2 is solved for the ﬁxed-free cantilever case, and is plotted in Figure 9.8 in terms of non-
dimensional parameters. These parameters are deﬁned as

P
non-dimensional axial load = (9:3)
Pcr
kL4
non-dimensional support stiffness h = (9:4)
EI
D
buckling amplification factor = (9:5)
D0

The above analysis shows that as soil liquefies progressively (i.e. k diminishes), the lateral pile head deflection
increases. It may be noted that the amplification is purely due to the axial load and clearly does not consider
the effect of dynamics of the problem. This above formulation can be applied during the lateral shaking phase
when the main shaking has ceased. This, however, may not be applicable when the inertial force acts.

Dynamic instability. Usually, the time period of vibration of a pile-supported structure is estimated based
on formulas that are derived from internationally calibrated data. This time period depends on the dimensions
of the superstructure without any consideration of the foundation (see Chapter 5, Section 5.2 for details).
However, during and after liquefaction, as the pile loses its lateral conﬁnement, it becomes an integral part of
the superstructure. The frequency of the structure may alter substantially, and in most cases will reduce.

250

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Figure 9.8 (a) Combined pile–soil model using a Euler–Bernoulli beam with axial and lateral force resting against a
distributed elastic support. (b) Variation of the buckling ampliﬁcation factor with respect to support stiffness and
axial load

F
w (L)

k
El

w (x)

(a)

20
Buckling amplification

0 0
1 20
0.8
0.6 0.4 40
0.2 0 60
η = kL4/El
(b) P/Pcr

251

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Reduction in the fundamental frequency of the structure during the earthquake will increase its flexibility,
and the structure may suffer more lateral deformation. The bending moment in the piles may increase
significantly if the altered natural frequency of the structure comes close to the driving frequency of the
earthquake. Figure 9.9 shows the results of the analysis of a cantilever pile where the variation of the first
natural frequency w1 is plotted with respect to the normalised support stiffness h and the normalised axial
load P/Pcr, where Pcr is the Euler critical buckling load.

For further details of the instability of piles, readers are referred to Bhattacharya et al. (2008), Bhattacharya
and Adhikari (2008), and Adhikari and Bhattacharya (2009).

9.4.2 Kinematic bending moment in the piles

There will be interaction between the pile motion and the vibrating soil motion. Owing to the discontinuity at
the interface between the soft and stiff layers, a kinematic bending moment will be induced in the pile.
Another aspect of the problem is the inertial bending moment in the pile due to oscillations of the super-
structure. The two types of bending moments are explained in Figure 9.10. The location of the inertial
bending moment is within the top few diameters of the pile. However, the location of the kinematic bending
moment is the interface (i.e. where there is sharp contrast in soil stiffness). If the sharp contrast is in the top
part of the pile, the maximum design bending moment may be the superposition of the two. This shows the
importance of ground investigation to ﬁnd out the layering and stratiﬁcation.

Figure 9.9 Variation of the ﬁrst natural frequency of a cantilever pile with respect to the normalised support
stiffness for different values of the axial load

140

120

100

80
ω1: Hz

0
0 400 600
0.2 0.4 0.6 200
0.8 0
η = kL4/El
P/Pcr

252

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Figure 9.10 (a) Kinematic and (b) inertial bending moments in a pile [Photo courtesy: Prof George Mylonakis]

Soft
layer

Stiff
layer

Seismic
SH waves

(a) (b)

9.5. Criteria for the design of piles in seismic areas

This section of the chapter lists the essential criteria for the design of pile foundations in liqueﬁable areas.
These are based on the work of Bhattacharya and Tokimatsu (2004) and Dash and Bhattacharya (2007). A
safe design procedure should ensure that the piles have enough strength and stiffness to sustain the following.

■ A collapse mechanism should not form in the piles under the combined action of lateral loads
imposed upon by the earthquake in addition to the axial load. At any section of the pile, the bending
moment should not exceed the allowable moment of the pile section. The shear force at any section of
the pile should not exceed the allowable shear capacity.
■ A pile should have sufficient embedment in the non-liquefiable hard layer below the liquefiable layer,
to achieve fixity in order to carry moments induced by the lateral loads. If proper fixity is not
achieved, the piled structure may slide due to the kinematic loads (lateral spreading of the soil).
Typical calculations to find the point of fixity are provided in Appendix A. However, as a rule of
thumb, the fixity is typically 3–6 times the diameter of the pile in the non-liquefiable hard layer.
Figure 9.11 shows a simplified model to consider fixity effects.
■ A pile should have sufficient capacity to carry the axial load acting on it during full liquefaction
without becoming laterally unstable (i.e. buckling instability). It has to sustain the axial load and
vibrate back and forth (i.e. it must be in stable equilibrium when the surrounding soil has almost zero
stiffness owing to liquefaction). Lateral loading due to ground movement, inertia or out-of-straightness
will increase lateral deflections, which in turn can cause plastic hinges to form, reducing the buckling
load, and promoting more rapid collapse. These lateral load effects are, however, secondary to the
basic requirements that piles in liquefiable soils must be checked against Euler’s buckling criterion.
This implies that there is a requirement of a minimum diameter of pile dependent upon the likely
liquefiable depth.

253

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 9.11 Fixity model

Before the earthquake After liquefaction

due to the earthquake

Increasing
1 L′s1 depth fixity
2 Point of fixity L′s2
(non-liquefied soil)
3
L′s1 < L′s2 Point of fixity
(layer 2 fully liquefied)

Soil layers:
1) Non-liquefiable soil
2) Liquefiable soil
3) Non-liquefiable soil (dense sand)

■ The pile should have sufficient capacity to carry the additional dynamic loads along with the static
forces without exceeding the yield. The piles should also be considered as a part of the structure while
carrying out the natural frequency estimation, especially in the liquefied condition. A simplified
method to find the frequency is provided in Chapter 5 (Figure 5.8) and more details are provided in
Figure 9.12(b) and Chapter 10 (Section 10.6).
■ The settlement in the foundation due to the loss of soil support should be within the acceptable limit.
Furthermore, the settlement should not induce end-bearing failure in the pile.

The criteria for design of piles in non-liqueﬁable layered soils is different. There will be interaction between
the dynamics of the structure and the oscillations of the soil layers due to the upward propagating shear
waves. The soil layers will have their own natural frequency depending on the layering of the soil, and the
structure together with the ground will have a coupled soil–structure interaction frequency. The design checks
and considerations are highlighted below.

■ The kinematic bending moment will be a maximum at the soil interface having sharp contrast, as shown
in Figures 9.4 and 9.10(a). The kinematic bending moment depends on the ratio of the shear wave
velocities of the soil layer (VS1 and VS2) at the interface, the pile–soil stiffness ratio in the layers, the
amplitude and the frequency content of the input motion, the location of the soil interface (shallow or
deep interface) and the boundary condition of the pile head.
■ In layered soils (see Figure 9.4), designers need to estimate the location and the magnitude of the
inertial and kinematic bending moments of the piles for various scenarios. For example, in the case
of an LNG tank structure, these include tank empty, tank full and tank partially full. For the tank
example in Figure 9.4, while the kinematic bending moment in the pile at the deep soil layer

254

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

interfaces (Figure 9.4, left diagram) will be almost unaffected by the various tank conditions (full,
partially full or empty), the inertial moments will change considerably with a variable base shear.
■ In a layered soil, if the interface is at a shallow depth (within the active length of the pile), the
position of the maximum inertial moment may be in closer proximity with the maximum kinematic
bending moment, demanding high bending stiffness of the pile.
■ The magnitude of the inertial bending moment depends on the soil–structure interaction period of the
structure–foundation–soil system and the input motion. The location of the maximum inertial bending
is normally a few diameters below the pile head (i.e. within the active length region).
■ For a short, stiff pile, the kinematic bending moment dominates the pile bending design.
■ For ﬂexible piles, a large bending moment may be experienced by the pile at the pile head if the
frequency of excitation is near the effective natural frequency of vibration. Also, if the input motion
excites the higher modes of soil layer vibration, a large kinematic bending moment may also occur at
the interface.

9.6. Design checks and calculations required

Engineers must carry out calculations to satisfy themselves that the failure mechanisms will not develop.
The following checks need to be carried out.

■ Shear failure. Shear failure of a pile may occur due to the lateral loads such as inertial or kinematic
loads or a combination of the above. Figure 9.12 shows the shear failure mechanism due to inertial
load. This is particularly damaging to hollow circular concrete piles (non-ductile) with a low shear
capacity.
■ Bending failure. Bending failure of piles may occur because of lateral loads either due to inertial
or lateral spreading loads, or a combination of the two. The combination depends on the type of
earthquake motion, the time of onset of liquefaction and the regaining of the strength of the soil after
liquefaction. A method to calculate bending moments due to lateral spreading loads is provided in the
Japanese code of practice for bridges (JRA, 2002), and is discussed later in this chapter.
■ Buckling instability. Buckling failure in slender piles may occur due to the effect of an axial load
acting on the pile and the loss of surrounding confining pressure provided by the soil owing to
liquefaction. This theory was proposed by Bhattacharya (2003), and later verified by many researchers
such as Knappet and Madabhushi (2010) and Shanker et al. (2007). Lateral loading due to slope
movement, inertia or out-of-line straightness in the pile will increase lateral deflections, which in turn
can increase the likelihood of instability failure even at lower axial loads. This may cause plastic
hinges in the piles, leading to collapse of the structure. Euler’s method may be used to calculate the
bucking load, and a safety factor of at least three is necessary.
■ Dynamic failure. All of the above failures (shear, bending and buckling) may occur due to static
loads. During earthquakes, dynamic soil–pile interaction becomes much more complicated and has a
significant effect on the pile response. The dynamic properties of the soil and structure change, which
may lead to amplification of the structural response, eventually leading to failure of the structure (see
Figure 9.12). Figure 9.12(b) explains the change in the characteristics of a structure due to a shift in
the time period. This amplification of response is very pronounced for bridge-type structures where a
large deflection of the deck is expected that can often lead to midspan collapse. Readers are referred
to Chapter 10 for the dynamic failure of bridges.
■ Settlement check. As soil liquefies, shaft capacity will reduce drastically and the settlement of the
foundation needs to be checked.

255

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 9.12 (a) Failure mechanisms. (b) Change in dynamic characteristics of a structure due to liquefaction

CM CM CM CM

Typical building with Shear failure Bending failure Buckling Dynamic amplification
pile foundation mechanism mechanism mechanism mechanism

(a)

T2
Spectral acceleration

Before liquefaction

After liquefaction

KRot.
KRot. KLat.
KLat. T1 T2 Period

(b)

256

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

9.6.1 Limit state design and the combination of mechanisms

The above ﬁve forms of failure could be described as the ‘limit state of collapse’. Each of these failure
mechanisms can individually cause complete collapse of a foundation. However, a real failure is perhaps a
non-linear combination of these mechanisms. The various mechanism and their interactions are shown in
Figure 9.13. The interaction of bending and buckling and dynamics has been studied by Dash et al. (2010) and
Bhattacharya et al. (2014) using the well-documented failure of Showa Bridge in Japan.

9.6.1.1 Case study: bending and settlement interaction

Figure 9.14 shows the collapse mechanism of the Kandla Port and Customs Office Tower building in India,
based on the work of Dash et al. (2009). This section describes the fundamental mechanisms that may cause
yielding/failure of a pile. The failure of the Kandla Port tower building shown in Figure 9.14(a) is used as an
example to discuss the various mechanisms. The building is located in laterally spreading ground in the city of
Kachchh, on the coast of the Arabian Sea. The lateral dimension of the building is 13 m, and it is 22 m high
and supported on 18 m piles. Figure 9.14(b) is a schematic of the failure of the Kandla Port tower building
shown in Figure 9.14(a). The founding level of the pile was in liquefiable zone. During the 2001 Bhuj
earthquake, the soil beneath the soft clay liquefied and could not support the vertical load of the building. This
is a case of interaction between the bending and settlement mechanisms.

Figure 9.13 (a) Individual mechanisms of pile failure. (b) Possible interactions between the failure mechanisms

Bending Settlement

Shear Static
instability
Dynamic
instability

(a)
ity
bil
ility

sta
tab
nt

c in
me

ins
g

mi
din
ar

ttle

tic

na
She

Ben

Sta
Se

Shear

Bending
Settlement
Static instability
Dynamic instability

(b)

257

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 9.14 Interaction of two mechanisms

(a) (b)
0.3 m 10 m

2m
Before earthquake
After earthquake

20 m
0.3 m
Boundary wall

Arabian Sea
10 m
soft clay
18 m

12 m
sandy soil
0.45 m

Brown clay

9.6.1.2 Case study: interaction of bending, buckling and dynamics – examples of good
performance (Saraighat Bridge) and poor performance (Showa Bridge)
For this type of analysis, an important design calculation is the change in the natural frequency of vibration of
a pile-supported structure during the process of liquefaction. The frequency of a pile-supported structure will
change with the stiffness degradation of the soil surrounding the pile. Usually, the time period of vibration of a
pile-supported structure is estimated based on formulas that are dependent on the dimensions of the struc-
tures. In some cases, engineers often model the structure in software and obtain the eigenvalues. This time
period depends on the dimensions of the superstructure without any consideration of the foundation.
However, during and after liquefaction, as the pile loses its lateral confinement, it becomes an integral part of
the superstructure. Figure 9.15 shows the Showa Bridge collapse). Figure 9.16 presents the configuration of
the foundation together with the time period of the bridge as soil liquefies progressively. Analysis carried out
by Bhattacharya et al. (2014) showed that the time period of the bridge elongated from about 2 s at no
liquefaction to about 6 s at full liquefaction. Analysis carried out by Halder et al. (2008) indicated that the
time taken to reach maximum liquefaction was about 10 s.

It is of interest to note the difference of behaviour between a bridge supported on a pile foundation and a
caisson (well) foundation as soil progressively liqueﬁes. Figure 9.17 shows a caisson supported bridge that
experienced many earthquakes and performed well. Further details can be found in Dammalla et al. (2017).
Figure 9.18 compares the change in the time period and predicted spectral displacement (Sd) for the Showa
and Saraighat Bridges. Higher spectral displacement may lead to unseating of the decks. The marked dif-
ference in the predicted behaviour can be attributed to the differential increase in the time period. For the same
depth of liquefaction of 10 m, the increase in time period for the Showa Bridge is 300%, but the corre-
sponding increase for the Saraighat Bridge is about 18%. It is important to note that the Showa Bridge
collapsed due to fall of the decks; this is an example of dynamic failure.

258

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Figure 9.15 The pile-supported Showa Bridge that collapsed during the 1964 Niigata earthquake (Japan)

Figure 9.16 Period estimation for Showa Bridge: (a) foundation conﬁguration for period estimation; (b) variation of
the period with the liqueﬁed soil layer thickness

(a) (b)

6.5
Dead weight,
M = 74 t Deck level
6.0

5.5
Air 6m
5.0
Period: s

4.5
Water 3m
4.0

3.5
Liquefied
soil layer Variable 3.0
thickness
2.5

Depth of fixity = 4D, 2.0

where D is the pile diameter 0 2 4 6 8 10
Liquefied soil layer thickness: m
Non-liquefied
stiff soil

259

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 9.17 Saraighat Bridge (India) with caisson foundations

Figure 9.18 Change in the time period and spectral displacements for the Showa and Saraighat Bridges

150

125 ☆
Change in time period: %

☆
100 ☆
ge

Saraighat
rid

☆ ☆ Showa
db

75
rte

☆
po

☆
up

50
e-s

☆
Pil

25 ☆ bridge
supp orted
☆ Ca isson-

0 5 10 15 20 25
(a) Depth of liquefaction: m

200
Saraighat
☆ Showa
160
Change in Sd: %

ge
id

120
br
d
te
or
pp

80
su
-
le
Pi

40
ridge
pported b
Caisson-su
0
0 3 6 9 12 15
Depth of liquefaction: m
(b)

260

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

9.7. Design codes of practice

It is of interest to compare the state-of-the-art understanding with the latest codes of practice for pile design.
Most codes, such as the US FEMA P-750 (Building Seismic Safety Council, 2009), the European Eurocode 8
(CEN, 2004), the ‘Japanese Speciﬁcations for highway bridges’ (JRA, 2002) and the Indian IS-1893 (BIS,
2000), focus on the bending mechanism of the pile. Essentially, the codes recommend that piles should be
treated as beams.

9.7.1 Japanese Road Association (JRA) code on lateral spreading loads

The ‘Japanese Specifications for highway bridges’ (JRA, 2002) suggests designing piles against bending
failure by considering both the inertia and the kinematic force due to the lateral spreading. The action of the
liquefied soil can be considered as a distribution equal to 30% of the total burden pressure. Figure 9.19 shows
that, if there is also a non-liquefiable crust overlaying the liquefiable layer, its action can be accounted for by a
passive earth pressure distribution.

9.7.2 Eurocode 8 on kinematic bending moments in piles

Eurocode 8 Part 5 (CEN, 2004) suggests designing piles considering the bending due to inertial and kinematic
loads arising from the deformation of the surrounding soil, recommending that piles should be designed to
remain elastic. However, the interfaces between layers having different properties should have the capacity
to develop a plastic hinge. For these calculations the lateral resistance offered by the soil layers susceptible to
liquefaction should be ignored. Eurocode 8 Part 5 also states

Figure 9.19 Kinematic load on a pile due to lateral spreading together with idealisation for the seismic design of
bridge foundations suggested by the JRA (2002)

A Non-liquefiable crust
B Liquefiable layer
C Non-liquefiable layer

Passive earth
A pressure

30% of the
overburden pressure

261

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Potential plastic hinging shall be assumed for:

■ a region of 2d from the pile cap

■ a region of ±2d from any interface between two layers with markedly different shear stiffness (ratio
of shear moduli >6)
where d denotes the pile diameter. Such region shall be ductile, using proper conﬁning reinforcements.

Eurocode 8 prescribes that kinematic effects should be taken into account when all of the following conditions
simultaneously exist

■ the seismicity of an area is moderate to high and is characterised by a normalised peak ground
acceleration agS > 0.1g, where ag is the design ground acceleration on type A subsoil (essentially rock)
and S is the soil factor – see Section 5.4.1. For details of ground type A see Appendix A (Table A.8)
■ the subsoil type is D or worse, characterised by sharply different shear moduli between consecutive layers
■ the importance of the superstructure is class III or IV.

See Section A.10 of Appendix A for details of ground types.

The above factors have been echoed in the Italian building code OPCM 3274 (Presidenza del Consiglio dei
Ministri, 2003). This effectively considers the kinematic bending moment in piles as arising from the soil
deformations from the passage of seismic waves that impose curvatures and thereby strains on the piles.

9.7.3 Indian Standard (IS 1893) on the design of piled foundation

The only available guideline for pile foundations in liqueﬁable soils, given in IS 1893 (BIS, 2000), states that
‘The piles should be designed for lateral loads neglecting lateral resistance of soil layers liable to liquefy.’
Hence, in this analysis, the piles need to be considered to have no lateral resistance in the zone of liquefaction.
As discussed earlier, the axial instability of a pile may be enhanced due to the effect of lateral disturbing forces
(i.e. force due to lateral spreading if the ground is on a slope).

9.7.4 Indian Road Congress code (IRC:78) for the design of bridges/ﬂyovers
The IRC:78 code (IRC, 2014) is extensively used to design bridges and ﬂyovers in India. For pile design in
seismic areas, IRC:78 prescribes the following

■ Clause 709.4.2 states that piles may be designed taking into consideration all the load effects, and
their structural capacity examined as a column. The self-load of a pile or the lateral load due to
earthquake, water current force and so on on the portion of free pile up to the scour level and up to
the potential liquefaction level, if applicable, should be accounted for.
■ Clause 709.1.7 stipulates that the minimum diameter of piles should be as listed in Table 9.1.
■ Clause 705.4.1 states that, in soil, the minimum depth of foundations below the point of fixity should
be the minimum length required to develop full fixity as calculated by any rational formula.
■ Clause 709.1.4 requires that permanent steel liner should be provided at least up to maximum scour
level for piles in streams, rivers and creeks. In the case of marine clay or soft soil or soil having
aggressive material, a permanent steel liner of sufficient strength should be used for the full depth
of such strata. The minimum thickness of the liner should be 5 mm.
■ The reinforcements in a pile should be provided for the full length of pile, as per the design
requirements. However, the minimum area of the longitudinal reinforcement should be 0.4% of the
area of the cross-section in all concrete piles. Lateral reinforcement should be provided in the form of
links or spirals with 8 mm-diameter steel as a minimum and a spacing of not less than 150 mm. The
cover to the main reinforcements should not be less than 75 mm.

262

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Table 9.1 Pile speciﬁcation according to IRC:78

Pile type Land bridges: m River bridges: m

Driven cast-in-situ piles 0.5 1.2

Precast piles 0.35 1.0

Bored piles 1.0 1.2

IRC (2014)

9.8. Modelling of the problem

This section provides details of modelling the problem for different types of loading as well as calculation
methods against different type of mechanisms.

9.8.1 Bending due to inertial and kinematic loading

In the bending mechanism, piles are considered as beams where the lateral loads on the piles are due to inertial
effects from the superstructure and kinematic effects due to ground movement. Following Tokimatsu et al.
(1998), pile–soil interaction can be described considering three different phases as illustrated in Figure 9.20

■ Phase 1: prior to the development of pore water pressure, the inertial force from the superstructure
may dominate.
■ Phase 2: kinematic forces from liqueﬁed soil start acting with increasing pore pressure.
■ Phase 3: towards the end of shaking, kinematic forces dominate and have a signiﬁcant effect on the
pile performance, particularly when permanent displacement occurs in laterally spreading soil.

Figure 9.20 Different phases of pile loading based on the bending mechanism of pile failure

Pinertia force Pinertia force

Ground Ground
displacement displacement

B L L

C C C

Phase 1 Phase 2 Phase 3

During shaking before During shaking after Lateral ground movement
liquefaction liquefaction after earthquake

B Liquefiable layer
C Non-liquefiable ground
L Liquefied layer

263

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Based on large-scale shaking table tests in earthquake defence facilities, Tokimatsu et al. (2005) and Tokimatsu
and Suzuki (2005) suggested a methodology to combine inertial and kinematic loads for piles subjected to
seismic loads as follows.

When the natural period of the structure Tb is larger than the natural period of ground Tg, then
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Mmax = Mi 2 + Mk 2 (9:6)

but if Tb is less than Tg, then

Mmax = Mi + Mk (9:7)
where Mmax, Mi and Mk are the maximum bending moment, the bending moment due to the inertia of the
superstructure and the bending moment induced by the kinematic interplay (i.e. lateral spreading),
respectively.

9.8.2 Effect of the axial load on the bending response of the pile
Most often piles are subjected to both axial and lateral loads. During liquefaction, the buckling load or Euler
critical load Pcr of the pile decreases. Considering the vertical load of the pile P, and if Pcr approaches close to
P, the structure can become unstable. The inclusion of the axial load in bending analysis can easily be
implemented in a numerical model, where the lateral and axial loads are considered together. The numerical
model for the lateral loading of the soil on a pile is generally carried out in two different ways, using

■ the force-based method, where lateral loading due to the inertial or lateral spreading load can be
applied, or
■ the displacement-based method, where the predicted displacement of the ground may be applied.

Figure 9.21 shows these two methods of representing a field condition. In the force-based method, the force
from liquefied soil flow is represented by a limiting pressure Fmax on the pile, as shown in Figure 9.21(b).
Figure 9.21(c) illustrates displacement-based method, where the lateral spreading is modelled by applying a
displacement at the free ends of the soil springs. Typical results of the pile response from these two methods
with an axial load are described below.

Figures 9.22 shows the effect of the axial load on the bending response of a pile foundation, in terms of the
normalised pile displacement y/D (where y is the lateral displacement evaluated considering the effect of the
axial load and D is the pile diameter) against either Dsoil/D or F/Fmax (where Dsoil is the ground displacement
and F is the axial load), depending on the method of analysis (i.e. displacement or force based, respectively).
Figure 9.22 clearly shows that when the axial load is higher, the pile head deﬂection y becomes larger. For
a particular axial load (p), there is a sudden increase in the pile head deﬂection, which is the instability
phenomenon.

During liquefaction, as the soil loses its stiffness, the elastic buckling load Pcr also reduces. Assuming a
constant static axial load P on the pile, the ratio P/Pcr will therefore increase. With increasing P/Pcr, the pile
head deﬂection and the bending moment in the pile will also increase. Theoretically, when this ratio is close to
1 (i.e. P is close to the buckling load), the bending moment ampliﬁcation factor becomes very high, which

264

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Figure 9.21 Pile–soil interaction in liqueﬁable soil: (a) typical ﬁeld condition; (b) force-based model;
(c) displacement-based model

Axial load P Axial load P

Pile
Limiting
∆ soil Ground
pressure from
deformation
spreading soil
∆ soil
Liquefied soil

Fmax
Non-liquefied
soil

(a) (b) (c)

Figure 9.22 Pile head deﬂection response due to lateral and axial loads

Pcr
P=
n2
Normalised pile displacement, D
y

Pcr
P=
n1

P=0
(no axial load)

Increasing axial load

n1 > n2

Δsoil/D (displacement-based model) or F/Fmax (forced-based model)

265

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 9.23 Pile bending moment response due to lateral and axial loads

1
Normalised bending moment, Mp /M

Pcr
P=
n2 Pcr
P=
n1

P = 0 (no axial load)

Increasing in axial load

n1 > n2

Δsoil/D (displacement-based model) or F/Fmax (forced-based model)

allows the bending moment in the pile to reach its plastic moment capacity Mp at a lower value of lateral load.
The sudden rise in the pile head deflection demonstrates the failure of the pile when the bending moment
reaches Mp and the pile continues to deflect without any additional loading. A similar behaviour can also be
observed when axially loaded piles in level liquefied ground are subjected to inertial loading. This is
explained in Figure 9.23.

9.8.3 Numerical model of the bending–buckling interaction

In practice, piles are most often modelled as beams on a non-linear Winkler foundation (also known as the p–y
spring approach), where the soil is idealised as p–y springs. In the Winkler method of analysis of piles, the
pile–soil interactions are represented by a set of non-linear soil springs: p–y springs (commonly called p–y
curves, these incorporate the lateral pile–soil interaction), t–z springs (these model the shaft resistance, i.e. the
pile–soil friction) and q–z springs (these model the end-bearing interaction), further details can be found in
Section 5.8. Figure 9.24 shows a simple model of a pile that can be analysed using any standard structural
software and can incorporate advanced features such as P-delta effects and non-linearity in the material of the
pile. For any load or displacement applied to the pile either at the pile head (which represents the inertial load
from the superstructure) or along the pile, the required analysis outputs are pile deﬂection, rotation, bending
moment, shear and soil reaction. However, undoubtedly the critical inputs for a realistic analysis are the
springs that represent the interactions. Some important points regarding the shape of p–y curves are

■ A p–y curve looks self-similar to the stress–strain curve of the soil. Essentially, a stress–strain curve is
scaled to obtain a p–y curve. Readers are referred to Bouzid et al. (2013), Lombardi et al. (2017) and
Dash et al. (2010, 2017) for further details on scaling.
■ For fully liqueﬁed soil, the shape is upward convex, as shown in Figure 9.24. Methods to construct
p–y curves for liqueﬁed soil are given in Chapter 5 (Section 5.8.2).

266

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Figure 9.24 Method of pile analysis using the Winkler approach

Axial load

Load, p

Lateral load
Liquefied soil
Lateral soil
t–z springs

Liquefiable p–y
layer Bottom of
Displacement, y liquefiable
t–z layer

p–y

Load, p
t–z Non-liquefied soil
Non-liquefiable
p–y
layer

q–z

Displacement, y

9.8.3.1 Aside: p–y curves

In practice, p–y curves are obtained from codes of practice (e.g. see American Petroleum Institute (API),
2000), and the input required is the stress–strain of the soil. The process of transforming stress–strain to p–y is
semi-empirical. Figure 9.25(a) shows the typical stress–strain of sand and a typical API-type p–y curve for
sand. In contrast, Figure 9.25(b) shows the stress–strain of a typical clay along with typical API p–y curves for
clay. An interesting feature can be observed: the shapes of the p–y curves for sand and clay are similar to the
plots of their stress–strain behaviour. Figure 9.26 shows shapes of p–y curves for partially liquefied soils and
fully liquefied soils. For partially liquefied soil, a reduction factor (often termed the p-multiplier) is applied to
a standard p–y curve for the non-liquefied condition to obtain the p–y curve for the liquefied soil condition.

9.8.3.2 Simplified model to analyse the transition from stage II to stage III
The dynamics of the problem can be considered in a simplified way, as described in Figure 9.5(b). A pile can
be modelled as a free-standing column fixed at some depth below the ground surface – often referred to as the
depth of fixity (Df). Figure 9.5(b) shows a schematic of the simplified pseudo-static analysis, in which the
pile-supported structure (in stage II loading, as shown in Figure 9.5(a)), is modelled as a single-degree-of-
freedom system, and the seismic action is represented by an equivalent shear force Feq that is proportional to

267

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 9.25 (a) Typical stress–strain curve for sand (Wichtman, 2005) and the API p–y curve. (b) Typical stress–strain
curve for clay (Chai et al., 2007) and the API p–y curve

70
60
q = S1–S3: kPa σ3: kPa API p–y curve
50 200
175

p: kN/m
40 150
30 125
100
20 75
50
10
0
0 5 10 15 20
Axial strain, ε1: m y: m
(a)

1500
Deviatoric stress, q: kPa

510 kPa
382 kPa API p–y curve
p'0 = 127 kPa
1000

p'0: kPa p: kN/m

500 127
382
510
0 Simulated
0 5 10 15 20 y: m
Shear strain, εs: %
(b)

Figure 9.26 (a) p–y curve for saturated sandy soil during the process of liquefaction. (b) Simplified p–y curve for
liquefied soil. ru = degree of liquefaction (0 is non-liquefied and 1 is full liquefaction)

Non-liquefied soil ru = 0
ru = 0.25
ru = 0.5
ru = 0.75 pu

Liquefied soil ru = 1 Liquefied soil ru = 1

Load, P
Load, P

K K
K1

Transition from solid

to fluid state

Displacement, y y1 yu Displacement, y

(a) (b)

268

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

the total mass of the superstructure and spectral acceleration. The equivalent force is applied to the centre of
mass of the superstructure, and hence generates an overturning moment M at the point of ﬁxity, which is
given by

M = Feq (Df + hi ) (9:8)

where hi is the height of application of the inertia force measured from the foundation level.

For equilibrium, the overturning moment must be counterbalanced by two axial loads, which are indicated in
Figure 9.5(b) by Adyn, in which the subscript ‘dyn’ denotes ‘dynamic’, highlighting the fact that Adyn is a
result of the overturning moment rather than the weight of the superstructure.

With the onset of liquefaction and the subsequent reduction in the stiffness of the soil layer, the depth of ﬁxity
can increase signiﬁcantly. This primarily depends on the depth of liquefaction, and is shown schematically in
Figure 9.5(b), denoted by Df_liq. Clearly, the time period of the building/structure will increase, as does the
damping of the structure.

This simpliﬁed depth-of-ﬁxity approach incorporates two important features related to the reduction in soil
stiffness caused by liquefaction

■ Reduction in the inertial force and, consequently, the change in the maximum bending moment, due
to lengthening of the fundamental period of the models.
■ Lowering of the location of the maximum bending moment (i.e. the location of the maximum bending
moment moves to deeper locations, as does the point of ﬁxity). The transient bending moment is also
to be considered, and is explained below through an example.

Example. A typical ﬁve-storey building (16 m high × 9 m wide) is supported on piles that pass through 8 m
of liqueﬁable soil.

Usually, the time period of vibration of a pile-supported structure is estimated based on formulas that are
derived from internationally calibrated data (Chopra and Goel, 2000) such as provided in Eurocode 8 (CEN,
2004) and IS 1893 (BIS, 2002). This time period depends on the dimension of the superstructure without any
consideration of the foundation. However, during and after liquefaction, as the pile loses its lateral con-
finement, it becomes an integral part of the superstructure (stage III). Figure 9.27 shows the typical variation
of the structural configuration of the pile-supported structure during transition of the soil from being a
‘solid-like’ to a ‘liquid-like’ medium. The fundamental time period of the building is now estimated to the
first order for two cases: before and after full liquefaction.

Before liquefaction. The fundamental time period of the building as per clause 7.6.2 of IS 1893-1 is

h 16
T1 = 0:09 pffiffiffi = 0:09 pffiffiffi = 0:48 s (9:9)
d 9

It should be noted that this time period is necessary to estimate the base shear of the building using the
response spectra method. Other formulas are also available as discussed in Chapter 5 (Section 5.2).

269

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 9.27 Structural conﬁguration (a) before liquefaction and (b) after liquefaction. It is assumed in the example
that 8 m of soil liqueﬁed

Stiffness (EI) distribution along the height

Depth of fixity
Depth of fixity

After liquefaction. However, after liquefaction, the piles act like slender columns and will form an additional
storey where the piles act as the bottom-most storey. The fundamental time period of the structure will alter
substantially, and in most cases will increase. Figure 9.27(b) shows an analytical model. If it is assumed that
8 m of soil liqueﬁed and that the piles are ﬁxed a further 1.5 m below, the approximate period of the building
can go up to 2–5 s, depending on the stiffness of the piles.

An increase in the fundamental time period of the structure will increase its flexibility, and it may experience
more lateral deformation. Also, designers must ensure that the frequency of the structure at full liquefaction
should not come close to the driving frequency of the earthquake, to avoid resonance effects. The structure in
fully liquefied soil behaves like an inverted pendulum/open ground-storey structure with piles resembling
long ground columns, which is not considered an ideal design for seismic vibration. This situation is similar to
the soft-ground storey phenomena. Due to loss of the lateral soil stiffness of the piles in liquefiable soil, the
stiffness ratio between the superstructure and the stiffness of the pile group becomes larger. This stiffness
change along with the long unsupported length of piles may induce a large lateral displacement at the pier cap
level. In this soft-storey-like situation, as per IS 1893, the piles should be designed for 2.5 times the shear
force and the bending moment calculated under seismic loads. Due to liquefaction, the soil stiffness reduces
drastically, and the depth of fixity of the piles to the ground increases (see Figure 9.27(b)). The increase in the
unsupported length of the piles increases the moment in the piles. However, as the liquefied soil behaves like
a thick fluid (solid suspension), it also acts as a damper.

The lengthening of the time period of the structure (in this case from 0.48 s to say 3 s) depends on many
factors, including EI (the bending stiffness), the length of the piles, the depth of liquefaction, and the stiffness
of the soil beneath the liquefied layer. As the structure transits from 0.48 s (stage II in Figure 9.5(a)) to 3 s
(stage III in Figure 9.5(a)), the bending moment profile in the piles will constantly change, and this transition
time may vary. In most cases, depending on the input motion and the soil profile, this time will range between
8 and 20 s. Ground response analysis (see Chapter 4) can provide us with this information.

270

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

In some cases, during the transition, the frequency of the structure may become tuned to the frequency of the
earthquake, causing ampliﬁcation of the bending moments along the piles. However, due to enhanced
damping of the structure owing to liquefaction, the ampliﬁcation of the responses will be limited. While the
depth of liquefaction can be obtained from empirical methods (see Chapter 7), the time to reach full lique-
faction is very important for predicting transient bending moments in the piles and ground response analysis
is necessary.

9.8.3.3 Euler critical load and the P-delta effect

The elastic critical load Pcr of a pile can be calculated using Equation 9.10 based on the concept of the
effective length Leff of the pile in the liquefiable zone, and is a function of the boundary condition of the pile at
the top and bottom of the liquefiable zone and the depth of the liquefiable zone

π2 EI
Pcr = (9:10)
L2eff

where Leff = KHC, and HC is the depth from which the pile cannot receive any support – which in a simple case
is the summation of the depth of soil liquefaction and the depth to the point of ﬁxity. This is very similar to the
long column approach in structural design. The effective lengths of piles can be found from Table 9.2,
together with values of K.

Table 9.2 Values of Leff

Boundary condition of the pile at the top and bottom Effective K Example
of the liqueﬁed layer length
Top Bottom

Fixed Fixed (sufﬁcient embedment Leff = 0.5Hc 0.5 Pile group with raked piles
at the dense layer)

Free to translate but Pinned (insufﬁcient Leff = 2Hc 2 See Figure 9.29
restrained against embedment at the dense
rotation – sway frame layer)

Free to translate but Fixed (sufﬁcient embedment Leff = Hc 1 Most cases fall in this category:
restrained against at the dense layer) see, for example, Figure 9.5(b)
rotation – sway frame and Figure 9.28

Fixed in direction but Fixed (sufﬁcient embedment Leff = 0.7Hc 0.7 Pile group with raked piles;
free to rotate at the dense layer) improper pile–pile cap connection

Fixed in direction but Pinned (less embedment at Leff = Hc 1 Pile groups with raked piles;
free to rotate the dense layer) improper pile–pile cap connection

Free, i.e. unrestrained Fixed (sufﬁcient embedment Leff = 2Hc 2 Piles in a row such as the Showa
against rotation and at the dense layer) Bridge piles, see Figure 10.19
displacement

271

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

9.8.3.4 Example 1: estimation of the buckling load for a river bridge pile
Estimate the critical load (buckling load) of a river bridge pile if the soil liqueﬁes to 20 m below the ground
level. The pile is steel tubular section, with a 600 mm outside diameter, a 16 mm wall thickness and a Young’s
modulus E of 210 GPa. The pile is in a pile group and rigidly attached, and the top of the pile can translate
laterally but cannot rotate. The boundary conditions are shown in Figure 9.28.

Assuming the depth of ﬁxity is at three pile diameters below the bottom of the liqueﬁable layer, then

unsupported length L = 20 + ð0:6 3Þ = 21:8 m (9:11)

π2 EI
Pcr = as Leff = L (9:12)
21:82
π
I= (6004 − 5684 ) mm4 = 1:25 109 mm4 (9:13)
64

EI = 210 GPa 1:25 109 mm4

= 210 103 N/mm2 1:25 109 mm4
(9:14)
= 262:5 1012 N mm2
= 262:5 103 kN m2

Figure 9.28 Schematic for Example 1

Note: no rotation here

Liquefiable layer

20 m

10 m

272

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

π2 262:5 103 kN/m2

Pcr = = 5451 kN (9:15)
21:82 m2

Buckling is very sensitive to imperfections, and a minimum factor of safety of 3 is recommended. The safe
load during earthquake and liquefaction as described is therefore 5451/3 kN = 1817 kN.

9.8.3.5 Example 2: computation of the buckling load of a pile for a building

A building is supported by 350 mm concrete piles that are 11 m long. Each pile is a part of a large pile cap, and
is attached with a moment-resisting connection. The top boundary condition of each pile is therefore free to
translate but cannot rotate. The corresponding bottom boundary condition is pinned (see the deformed shape
in Figure 9.29). The soil liqueﬁed to 10 m.

The buckling load is Pcr = π2EI/Leff2 and, in this example, Leff = 2L and Pcr = π2EI/4L2, with E = 25 GPa (about
1/10 that of steel). So

π
I= 0:354 m4 (9:16)
64

L = ð10 m + fixity depthÞ ≈ 11 m (9:17)

Figure 9.29 Schematic for Example 2

273

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Note: the ﬁxity depth is about three times the diameter.

π2 25 103 π
Pcr = 0:354 = 0:37 MN = 370 kN (9:18)
4 112 64

The safe load during earthquake and liquefaction to a depth of 10 m is therefore 370/3 kN = 123 kN.

9.8.4 A practical method of carrying out design

Figure 9.30 schematically shows a procedure for carrying out calculations in the following steps.

(a) Based on seismic hazard studies, obtain the design earthquake magnitude Mw, the peak ground
acceleration and the nature of the earthquake.
(b) Use a simpliﬁed method such as that of Eurocode 8 or Idriss and Boulanger (2012) (discussed in
Chapter 7) to obtain the depth of liquefaction. Note that these methods cannot ascertain the degree
of liquefaction.

Figure 9.30 Deconvolution, site response analysis and soil–structure interaction

acceleration,
Spectral

S2: g

Spectrally matched time history

Time period: s

Liquefied layer A
Deconvolution

Non-liquefied layer B

Bedrock

Input motion

274

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

(c) Based on spectral matching or another accepted method, obtain the design bedrock motion.
Carry out a ground response analysis to obtain the site effects (see Chapter 4). Plot the time
history of the excess pore water pressure at various depths. Figure 9.31 shows the time history
of the excess pore water pressure of 20 m-deep ground for a real earthquake. Based on the pore
water pressure ratio, appropriate p–y springs can be assigned to the pile (i.e. fully or partially
liqueﬁed).

Figure 9.31 Time history and excess pore water pressure ratio for the 1976 Friuli earthquake (Italy) (EPWPR, excess
pore water pressure ratio)

1
Excess pore water pressure ratio at 5 m
EPWPR

0.5

0
(e)

1
Excess pore water pressure ratio at 10 m
EPWPR

0.5

0
(d)

Excess pore water pressure ratio at 15 m

EPWPR

0.5

0
(c)

1
Excess pore water pressure ratio at 20 m
EPWPR

0.5

0
(b)

0.5
Friuli earthquake, 1976
Acceleration: g

0.25
0

–0.25
–0.5
0 2 4 6 8 10 12 14 16 18 20
(a) Time

275

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

9.8.5 Kinematic bending moment in a pile

Kinematic pile–soil interaction has been studied by many researchers, and Table 9.3 summarises the main
features of the different methods. These methods can be classified into three types: (a) simplified methods,
(b) the beam on a Winkler foundation approach, and (c) finite-element and boundary-element methods.

One of the main requirement of the designer is to ﬁnd the kinematic moment at the interface. The problem is
explained in Figure 9.32, and two simpliﬁed approaches are presented here.

9.8.5.1 Dobry and O’Rourke (1983) approach

Based on homogeneous, isotropic, linear elastic soil and uniform static stress ﬁeld assumptions, the moment
at the interface is given by

M ≅ 1:86(Ep Ip )3/ 4 G11/ 4 g1 F (9:19)

where EpIp is the bending rigidity of the piles; G1 is the shear modulus of the top layer; g1 is the peak uniform
soil strain in layer 1, which may be obtained from any simpliﬁed approach; and F is a dimensionless factor

Table 9.3 Kinematic bending moment analysis

Method and example researchers Main analytical features

Simpliﬁed methods

Margason and Holloway (1977), National ■ Pile follows the free-ﬁeld motion
Institute of Building Sciences (2012) ■ Kinematic bending moments derived from the peak
curvature of the pile
■ Soil–pile interaction neglected
■ Pile–soil relative stiffness neglected
■ Pile slenderness neglected
■ Radiation damping neglected

Beam on a Winkler foundation approach

Kavvadas and Gazettas (1993), Matlock and Foo ■ May include linear, equivalent linear or non-linear
(1979), Mylonakis et al. (1997), Novak (1979), analyses
Penzien (1970), Tahghighi and Kanagai (2007) ■ Soil–pile kinematic interaction catered for
■ Several models account for soil non-linearity and
stiffness degradation

Finite element and boundary element

methods

Chau et al. (2008), Kaynia and Mahzooni (1996), Based on continuum mechanics models
Kuhlemeyer (1979), Krishnan et al. (1983), ■ Pile and surrounding soil simulated by axisymmetric
Padron et al. (2008), Wu and Finn (1997), Zhang elements and energy-transmitting boundaries
■ Soil–pile kinematic interaction catered for
et al. (2000)
■ Soil non-linearity and stiffness degradation catered for

276

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Figure 9.32 Kinematic bending moment at the interface with sharp contrast [courtesy: RELUIS project]

(a) Pile end conditions free/fix

(b) Superstructure present yes/no Free-head pile
Pile
Kinematic pile
v1, r1, b1, n1 Soil bending moment

G1 h1 Layer
Ep, lp, m
l interface

Interface
v2, r2, b2, n2
G2 h2
d Seismic excitation Seismic SH
undirectional wave excitation

Bedrock

depending on the stiffness contrast at the interface (G2/G1, where G2 is the shear modulus of the bottom layer).
F is a function of c, where c is given by

(1 − c−4 )(1 + c3 )
F= (9:20)
(1 + c)(c−1 + 1 + c + c2 )
where

1 / 4
G2
c= (9:21)
G1

9.8.5.2 Mylonakis (2001) strain-transmissibility approach

An improvement on the Dobry and O’Rourke (1983) approach was developed by Mylonakis (2001), who
proposed

Ep Ip eP
M=2 g F (9:22)
d g1 1

where ep is the pile peak bending strain and g1 denotes the soil shear strain at the layer interface. The ratio
of these parameters ep/g1 represents a kind of ‘strain transmissibility’ function and can be obtained from
Figure 9.33 depending on h1/d (the system geometry, i.e. the depth of the interface), G1/G2 (the stiffness
contrast of the layers) and Ep/E1 (the pile–soil stiffness ratio). The effect of frequency is taken into account
through F, which can be obtained from Figure 9.33 and using

ep /g1
F= (9:23)
(ep /g1 )w=0

277

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 9.33 Design chart for strain transmissibility

0.50
Design chart

h1 γ1
Ep /E1 = 100
0.40 εp G1, E1
G2 20 ∞
h1/d = 7 10
Strain transmissibility, εp/γ1

> La
Ep 300
0.30
d

1000
0.20

Possible 10 000
∞
0.10 lower limit

0.05

0.0
1 10 100 1000

Layer stiffness contrast, G2/G1

For the range of frequencies of interest, F is generally less than 1.25.

Figure 9.34 illustrates the calculation procedure of strain transmissibility method.

9.8.5.3 Aside: observed performance of piled-supported structures during past

earthquakes
Following the 2011 Tō hoku earthquake (Japan), a survey was carried out in a modern (2005 built) 15-storey
condominium in Higashi-suna in the Sunamachi area on the banks of the Naka River. The condominium has
236 apartments and runs parallel to the Tokyo Metropolitan Area highway. There was no damage to the
superstructure: Figure 9.35 shows the building after the earthquake and a plan of the structure. The building is
L shaped but there is no structural connection between the two sections.

The building is supported on concrete piles that are 1–2 m diameter and 53–54 m long. The diameter of the
piles at the base ranges from 1.4 to 3.5 m, for enhanced end-bearing. The load-carrying capacity of the pile
ranges between 2356 and 15009 kN. Each column was supported on one of these large-diameter piles. The
superstructure is a steel frame encased within reinforced concrete, and the interior is light wood partition
walls. The long side of the building runs along the Naka River. Lateral spreading was not observed along the
river bank due to good performance of the river bank protection work.

278

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Figure 9.34 Calculation methodology of strain-transmissibility method

2.00
h1/d = 20

1.75

1.50

20
1.25
10
Φ = (εp/γ1)/(εp/γ1)ω = 0

G2/G1 = 5

1.00

2.00

h1/d = 10

1.75

1.50

Ep/E1 = 100
1000
10 000
1.25

1.00
0.00 0.05 0.10 0.15

Dimensionless frequency; a0 = ω d/Vs1

Table 9.4 lists the diameter of the piles of ten pile-supported structures in liquefiable soils and their observed
performance following a strong earthquake. The case records were taken from the performance of structures
following the 1964 Niigata earthquake (Japan), the 1995 Kobe earthquake (Japan) and the 2001 Bhuj
earthquake (India). Note that the 1960s design concept of a few small-diameter piles (typically 0.3–0.6 m)
were gradually replaced by a few large-diameter piles in the 1990s. Poor performance in this context signifies
formation of plastic hinges in the pile and cracking. Further details of these case studies and failure modes can
be found in Dash et al. (2008, 2010), Bhattacharya and Madabhushi (2008) and Bhattacharya et al. (2004,
2005). In most of these examples, it has been estimated that the soil liquefied up to 7–10 m in depth. It is clear
from Table 9.4 that small-diameter piles performed poorly.

279

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

̄ hoku
Figure 9.35 Pile-supported condominium in the Sunamachi area: (a) photograph taken after the 2011 To
earthquake; (b) plan

(a)

(b)

280

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Table 9.4 Case histories of pile foundation performance during past earthquakes

Case history Pile diameter: m Performance

NFCH pile-supported building, 1964 Niigata earthquake 0.35 (hollow RC) Poor
(Hamada, 1992a)

Pile-supported Yachiyo Bridge, 1964 Niigata earthquake 0.3 (RC) Poor

(Hamada, 1992a)

Pile-supported NHK building, 1964 Niigata earthquake 0.35 (RCC) Poor

(Hamada, 1992a)

Pile-supported Showa Bridge, 1964 Niigata earthquake 0.6 (steel) Poor

(Bhattacharya et al., 2005)

Pile-supported LPG tank 101, 1995 Kobe earthquake 1.1 (RC) Good
(Ishihara, 1997)

Pile-supported LPG tank 106 and 107, 1995 Kobe earthquake 0.3 (RC hollow) Poor
(Ishihara, 1997)

Pile-supported 14 storey building, 1995 Kobe earthquake 2.5 (RC) Good

(Tokimatsu et al., 1996)

Pile-supported Kandla Port Tower, 2001 Bhuj earthquake 0.4 (RC) Poor
(Dash et al., 2008)

Pile-supported old jetty in Kandla port, 2001 Bhuj earthquake 0.5 (RC) Poor
(Bhattacharya and Madabhushi, 2008)

Pile-supported new jetty in Kandla port, 2001 Bhuj earthquake 1.0 (RC-ﬁlled steel tube) Good

RC, reinforced concrete

Bhattacharya et al. (2004) carried out a comprehensive study of 15 case histories of piled foundation
performance based on buckling analysis parameters. Essentially, the study showed that a pile is laterally
unsupported in the liquefiable zone due to the removal of lateral confinement by the soil, owing to liquefaction.
The slenderness ratio Leff/rmin of the pile in the region that could become unsupported is used to classify pile
foundation performance. Leff is the Euler effective length of an equivalent pin-ended strut and rmin is the
minimum radius of gyration of the pile section. The study showed that a line representing a slenderness ratio of
50 could distinguish between poor and good pile performance. This line is of some significance in structural
engineering, as it is often used to distinguish between ‘long’ and ‘short’ columns. Columns having slenderness
ratios below 50 are expected to fail by plastic squashing, whereas those above 50 are expected to fail by
buckling, both modes being modified by induced bending moments. This slenderness ratio of 50 signifies a
length-to-diameter of about 12 for reinforced concrete columns. Furthermore, the study carried out by Dash
et al. (2010) showed that large-diameter piles were safe against the bending, bucking and bending–buckling
interaction types of failure. The good performance of the building described in Figure 9.35 further reinforces
the notion that a few large-diameter piles are better that many small-diameter piles.

281

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

9.9. Raked or battered piles

While vertical piles are common, there are cases where raked (inclined) piles have been used to support
structures. Typically, raked piles are used to support marine structures (port buildings, wharfs, offshore
platforms, etc.) or bridge foundations where large lateral loads are expected. While broad aspects of the
various failure mechanisms of vertical pile foundations are understood, there is little consensus on the
dominant mechanism of failure. On the other hand, there is confusion regarding the use of raked or inclined
piles in seismic areas, mainly due to their mixed performance during past earthquakes. Many codes of
practice prohibit the use of raked piles in seismic areas: see commentary on the codes by Giannakou et al.
(2007), Gerolymos et al. (2008), Escofﬁer et al. (2008) and AFPS (1990).

This section therefore has two aims

■ to collate and critically analyse the performance of raked piles during past earthquakes
■ to provide a method to obtain lateral stiffness.

9.9.1 Case studies: raked piles

Studies of damage to structures supported on raked piles during past earthquakes have provided observations:
some foundations performed superbly while others collapsed. For example, raked piles remained undamaged
in Landing Bridge in the aftermath of the 1987 Edgecumbe earthquake (New Zealand) and also in the port
structures in Kandla Portfollowing the 2001 Bhuj earthquake (India), but severe damage to raked piles was
observed in Rio Banano Bridge during the 1991 Costa Rica earthquake and in a wharf structure in the Port of
Oakland during the 1989 Loma Prieta earthquake (USA). Table 9.5 lists some of the case studies reporting on
the performance of raked piles.

One of the reasons for using raked piles is their enhanced lateral stiffness. This section aims to provide a
comparative study of vertical and lateral stiffness in the liquefied condition. As a result, soil stiffness is
ignored. Figure 9.36 shows two pile foundations, one having a group of four vertical piles and the other a
group of four raked piles. In both the cases, the length of the pile is L, with a bending rigidity of EI. The raked
piles are defined by an angle of batter (f), which is the angle that the pile forms with the vertical. It is assumed
that the pile is fixed at some level below the liquefiable zone. This point of fixity depends on the relative
pile–soil stiffness. Typical values of fixity lie between 5 and 10 times the diameter of the pile. In this section,
a simple analytical expression is used to find the lateral stiffness of piled foundations for these two cases.

9.9.1.1 Lateral stiffness of a group of vertical piles

The boundary condition is assumed as free to translate at the pile head and free to rotate. The lateral stiffness
of a group of four piles is given by

12EI
K =4 (9:24)
L3

where EI is the bending rigidity of the pile and L is the length of the pile. This expression can be found in most
standard textbooks.

282

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Table 9.5 Case studies of raked pile performance

Case study Details Performance

Performance of Landing Bridge, 1987 406 mm2 prestressed concrete piles at 1:6 rake Good
Edgecumbe earthquake (Berrill et al., (angle of batter 9.5°). The piles passed through 4 m
2001) of liqueﬁed soil. Both piers and abutments were
supported by raked piles.

Port of Oakland, 1989 Loma Prieta Raked piles used in conjunction with vertical piles. The Poor
earthquake (Seed et al., 1991) piles passed through bay mud and were founded on
dense sand. The 7th Street terminal used 406.4 mm2
prestressed concrete piles. At Piers 27 and 29, the pile
section was 508 mm2 prestressed concrete.

Rio-Banano Bridge, 1991 Costa Rica Abutments were supported on two rows of 360 mm2 Poor
earthquake (EERI, 1991) driven precast piles. The front row had a 1:5 rake
(angle of batter 11.3°), and the abutment rotated.

Rio-Bananito Bridge, 1991 Costa Rica Piles installed at rakes of 1:5 and 1:10. Poor
earthquake (EERI, 1991)

Inclined pile- supported quay wall, 1995 Inclined piles supported Maya wharf. Good
Kobe earthquake (Kastranta et al., 1998)

Berths in Kandla Port, 2001 Bhuj Hollow reinforced concrete piles of 0.51 m diameter Good
earthquake and an average length of 17.7 m.

Figure 9.36 (a) Vertical and (b) raked piles

Bending rigidity El, length L
Bending rigidity El, length L

Plane of facity in
non-liquefiable hard layer

(a) (b)

283

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

9.9.1.2 Lateral stiffness of the group of raked piles

The lateral stiffness of a raked pile is given by

EA 12EI 12EI AL2
Kraked = 4 cos2 q + 3 sin2 q =4 cos2 q + sin2 q (9:25)
L L L3 12I

where EA is the axial stiffness of the pile, A is the cross-sectional area of the pile and q is the angle made by the
pile with the horizontal (seen Figure 9.36).

The second moment of area I is given by

2
I = Armin (9:26)

where rmin is the minimum radius of gyration.

Combining Equations 9.25 and 9.26 we get

12EI AL2 12EI l2
Kraked =4 3 2
cos2 q + sin2 q =4 3 − 1 cos q + 1
2
(9:27)
L 12Armin L 12

where l = L/rmin and is the slenderness ratio of the pile in the likely unsupported zone.

Comparing Equations 9.24 and 9.27, we deﬁne an improvement factor Zi as

Kraked l2
Zi = = − 1 cos2 q + 1 (9:28)
K 12

This equation suggests that raked piles are always laterally stiffer that vertical piles. It can be recast using the
angle of batter f as

Kraked l2
Zi = = − 1 sin2 f + 1 (9:29)
K 12

Equation 9.29 is plotted in Figure 9.37. The factor Zi represents the ratio of the stiffness of a raked pile with
respect to a vertical pile, and could be termed the ‘improvement factor’. Zi is plotted in Figure 9.37 against the
angle of batter for various slenderness ratios l.

284

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Figure 9.37 Normalised stiffness ratio of a raked pile plotted against the angle of batter

100
λ = 40
90
λ = 60
80 λ = 80
λ = 100
70
Improvement factor, Zi

1
0 2 4 6 8 10 12 14 16 18 20
Angle of batter, φ: °

9.10. Example: pile-supported LNG tanks

LNG is natural gas (primarily methane) in its liquid form, and is one of the cleanest forms of hydrocarbon
fuel. It has been reported that the world is discovering roughly four times as much gas as it uses every year.
However, the main gas consumption areas, for example the USA, Japan, UK, India, China and Korea, are
long distances from the location of reserves (e.g. Russia). As a result, storage and transmission is an engi-
neering challenge. LNG storage tanks are typically 50–100 m in diameter and about 20–40 m in height,
providing a storage capacity of between 100 000 and 200 000 m3. When required, LNG is revaporised and
piped into the gas distribution network. Therefore, LNG storage tanks represent critical infrastructure
facilities since they are required to provide a steady and safe energy supply even after a natural disaster, such
as an earthquake. An LNG tank is a containment-type structure, similar to a nuclear reactor, and comprises
(a) an inner stainless steel (nickel steel) tank that contains the LNG at about –160°C and (b) an outer tank of
approximately 1 m thick insulation that surrounds the inner tank to safeguard against any disaster in case of an
emergency.

285

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Seismic hazards that can cause partial to complete failure of an LNG tank foundation include

■ displacement of the foundation due to fault rupture, slope stability failure or lateral spreading effects
■ loss of vertical load bearing capacity in the case of a raft foundation or end-bearing piles in liqueﬁable
soils
■ excessive settlement due to liquefaction of the subsoil
■ structural failure of the piles due to pile instability, shear or bending.

Modern seismic analyses are based on the probabilistic response spectrum rather than deterministic peak
ground acceleration. Large-scale LNG storage tanks have a fundamental natural period of about 2–3 Hz,
depending on their shape and geometry, and are more or less within the range of maximum excitation of
typical severe earthquakes. Increasingly, LNG storage tanks are constructed in highly seismic areas and
where the supporting soil is loose to medium-dense sands. In such areas, special remedial measures such as
ground improvement need to be taken to counter soil liquefaction. Often, piles are used to support these
structures. A typical design of an LNG tank will include the following steps

(a) Site-specific seismic hazard studies. The aim of these studies is to arrive at a design ground motion
in an accurate and defensible manner, and to verify whether or not the site is not susceptible to
ground failure, liquefaction or fault rupture hazards. This task includes the identification and
characterisation of all seismic sources (location, size, type and style of movement, maximum
magnitude, seismogenic width, seismic slip rate, etc.), development of a seismotectonic model of
the region, performance of probabilistic seismic hazard analyses (PSHAs) with full Monte Carlo
simulation for the inclusion of uncertainly in the input parameters, deaggregation of the results, and
identification of the magnitude and site-to-source distance of the controlling earthquake scenario,
selection of recorded time histories based on the results of the deaggregation of PSHAs, and spectral
matching in the time domain for the development of site-specific design ground acceleration time
histories at the surface of weathered rock (outcrop motion). The site-specific seismic studies need
to be carried out for the site for two types of motion: operation basis earthquake (OBE) and safe
shutdown earthquake (SSE).
(b) Site response analyses. This is the second step, and is needed to study changes in the ground motion
characteristics (intensity and frequency content) as seismic waves propagate from the underlying
weathered rock to the ground surface. If the soil liquefaction is a potential problem and whether or
not it can be mitigated. However, if the soil liquefaction potential is not mitigated, a non-linear
effective stress site response analysis needs to be carried out. This is mainly because liquefied soils
act as a natural base isolator, preventing strong ground motion propagating through the soil column.
This would result in a major reduction in the inertial load (base shear) and the resulting lateral load
demand on the piles. In summary, two types of analyses may performed: equivalent linear analysis
with full p–y soil springs (no soil liquefaction) and non-linear analysis effective stress analysis with
reduced p–y springs (soil liquefaction).
(c) Structural design of piles. The bending moments and shear forces in the pile are obtained using a
full dynamic model.
(d) Pile drivability studies. These studies are undertaken to explore the necessity of ground
improvement.

286

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

9.10.1 Pile-supported LNG tank in layered soils (non-liqueﬁable)

In the ﬁrst example, an LNG tank is considered that has an outside diameter of 92 m and an inside diameter of
88 m. The capacity of the tank is 180 000 m3, which is equivalent to about 84 600 t of LNG. The time period
of vibration in the impulsive mode time period is about 0.39 s. The total mass of the tank and LNG – referred
to as the ‘full’ condition – is about 1500 MN. The soil at the site is layered, with 9.5 m of soft soil and a shear
wave velocity of 80 m/s (bulk density of 15 kN/m3); below this is stiff soil with a shear wave velocity of
330 m/s (bulk density of 20 kN/m3). The input acceleration used in this study at the engineering base layer
(bedrock or outcrop motion) is shown in Figure 9.38. This is a synthetically generated motion rich in
frequency, and the fast Fourier transform of the motion is shown in Figure 9.39. The maximum input
acceleration is 1.25 m/s. Figure 9.40 compares the response spectrum of the input motion with Eurocode 8,
and shows a very good match with the bedrock spectra.

As a second example, the LNG tank described above is considered to be supported on 300 concrete piles that
are 1.3 m in diameter and 15.5 m long. The soil proﬁle suggests that due to the stiffness contrast of the
supporting soil at the interface (ratio of the shear modulus is about 23) there will be a signiﬁcant kinematic
bending moment induced in the piles. This is because of the curvature imposed by the surrounding soil due to
the passage of seismic waves, and is termed the ‘kinematic bending moment’, which is different from the
‘inertial bending moment’ arising from the superstructure. The kinematic and inertial moment in the piles is
estimated using a research program (SPIAB 3.0) (Mylonakis, 2001; Mylonakis et al., 1997). The envelope of
the maximum kinematic and the inertial bending moments in the pile is shown in Figure 9.41. The following
points can be observed

■ A substantial kinematic bending moment developed at the interface and is independent of the tank
condition (full or empty). A simple back-of-the-envelope check based on Dobry and O’Rourke (1983)
and Mylonakis (2001) is given below.
■ As expected, the inertial bending moment in the top part of the pile changes with the condition of the
tank (full or empty).

Figure 9.38 Input motion used in the study

1.2
1.0
Response acceleration: m/s2

0.8
0.6
0.4
0.2
0
–0.2
–0.4
–0.6
–0.8
–1.0
–1.2
–1.4
0 1 2 3 4 5 6 7 8 9 10 11 12
Time: s

287

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Figure 9.39 Fast Fourier transform of the input motion

0.8

0.7

0.6

0.5
Fourier amplitude

0.4

0.3

0.2

0.1

0
5 10 15 20
Frequency: Hz

Figure 9.40 Response spectrum of the input motion and its comparison with Eurocode 8

4.5
4.0 Response spectrum of input motion

Eurocode 8: type 1, ground type D

Spectral acceleration: m/s2

3.5
Eurocode 8: type 1, ground type A (bedrock)
3.0
2.5
2.0

1.5
1.0

0.5
0.0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
Period: s

288

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Figure 9.41 Inertial and kinematic bending moments in the pile for tank full and tank empty

Bending moment: kN m
0 3000 6000 9000 12 000
0

6
Depth of the pile: m

8 Inertial moment (full tank)

Inertial moment (empty tank)

10 Kinematic moment (empty tank)

Kinematic moment (full tank)

9.10.1.1 Veriﬁcation based on Dobry and O’Rourke (1983)

The moment at the interface is given by

M ≅ 1:86(Ep Ip )3 / 4 G11 / 4 g1 F (9:30)

EpIp is the bending rigidity of the piles (3500 MN m2); the second moment of area is 0.14 m4 and Young’s
modulus for concrete is taken as 25 GPa. G1 is the shear modulus of the top layer (9.6 MPa in this case). F is a
dimensionless factor depending on the stiffness contrast at the interface G2/G1, where G2 is the shear modulus
of the bottom layer. For the present problem, the ratio of (G2/G1) is 23, and this factor is about 0.4. g1 is the
peak uniform soil strain in layer 1, which may be obtained from any simpliﬁed approach – estimated to be
0.0051.

289

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Based on the above data, the peak moment at the interface has been estimated to be 3200 kN m, which is of
similar magnitude of the moment obtained using SPIAB. This apparent difference is understandable if the
inherent assumptions in the formulation are considered. The most notable limitations are the absence of the
dynamic nature of the excitation and the weak assumption of constant shear strain within each layer.

9.11. Example: Kandla Port tower building on piles

The building considered here is the Port and Customs Tower situated in the Kandla Port area, Gujarat, India
(see Figure 9.1(a)). This building tilted during the 2001 Bhuj Earthquake, and liquefaction was observed
around the building. Figure 9.42 shows a modelling technique allowing investigation of the dynamic
behaviour of the soil–pile–structure interaction during liquefaction. The superstructure is modelled as an
equivalent single-degree-of-freedom system, which gives the same fundamental period of the actual building.
The pile–soil interaction can be modelled using appropriate p–y springs. Therefore, the ﬁrst step is to carry out
a ground response analysis whereby the behaviour of the ground can be ascertained.

Figure 9.42 Soil–pile–structure interaction model used in the SAP 2000 software

Foundation mat

Three-dimensional model Section of front-row piles

of the foundation in SAP 2000

Lateral soil spring Axial soil spring Bearing soil springs

End-bearing soil spring
(p–y spring) (t–z spring) under the foundation mat

290

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

9.12. Summary
Piles are columns embedded deep into the ground, and are one of the most common forms of foundations in
seismic areas with weak soils. Pile-supported structures founded on liqueﬁable soils continue to collapse
during earthquakes despite being designed with the required factors of safety against bending due to lateral
loads and axial capacity (shaft resistance and end bearing). Conventional pile design requires two types of
calculations

■ For the bearing capacity (axial and lateral capacity); that is, how much vertical load can be safely
applied, which depends on the pile geometry, pile material and the soil layering. This is often termed
the geotechnical bearing capacity. Essentially, it is the summation of the shaft resistance (friction
between the soil and the pile) and the end bearing, applying a factor of safety of 2.5–3. The lateral
capacity also needs to be calculated, and this is often guided by the pile type and its bending moment
capacity. The lateral load capacity of short piles is dictated by the soil failure next to the pile. In
contrast, long slender piles fail by formation of a plastic hinge.
■ For the bending moment along the length of the pile, it is assumed that piles are beams supported on
non-linear Winkler springs. For a given lateral load acting on the pile head, the bending moment can
be calculated and checked with the capacity.

Pile design in seismic areas requires additional considerations

■ When soil liqueﬁes, it loses much of its stiffness and strength, so the piles now act as long slender
columns, and can simply buckle (buckling instability) under the combined action of the axial load and
inevitable imperfections (out-of-line straightness, lateral perturbation loads due to inertia and/or soil
ﬂow, etc.). The conventional design approach thus recommends that piles are designed as laterally
loaded beams.
■ The natural frequency of pile-supported structures may decrease considerably owing to the loss of
lateral support offered by the soil to the pile, and the damping ratio of structure may increase to values
in excess of 20%. These changes in dynamic properties can have important design consequences.

REFERENCES
Adhikari S and Bhattacharya S (2012) Dynamic analysis of wind turbine towers on flexible foundations. Shock and
Vibration 19(1): 37–56.
API (American Petroleum Institute) (2007) Recommended Practice for Planning, Designing and Constructing
Fixed Offshore Platforms – Working Stress Design. API, Washington, DC, USA.
Berrill JB, Christensen SA, Keenan RP, Okada W and Pettinga JR (2001) Case studies of lateral spreading forces
on a piled foundation. Géotechnique 51(6): 501–517.
Bhattacharya S (2003) Pile Instability During Earthquake Liquefaction. PhD thesis, University of Cambridge,
Cambridge, UK.
Bhattacharya S (2006) Safety assessment of existing piled foundations in liquefiable soils against buckling
instability. ISET Journal of Earthquake Technology 43: 133–147.
Bhattacharya S and Adhikari S (2011) Experimental validation of soil–structure interaction of offshore wind
turbines. Soil Dynamics and Earthquake Engineering 31(5–6): 805–816.
Bhattacharya S and Madabhushi SPG (2008) A critical review of methods for pile design in seismically liquefiable
soils. Bulletin of Earthquake Engineering 6(3): 407–446.
Bhattacharya S, Madabhushi SPG and Bolton MD (2004) An alternative mechanism of pile failure in liquefiable
deposits during earthquakes. Géotechnique 54(3): 203–213.
Bhattacharya S, Bolton MD and Madabhushi SPG (2005) A reconsideration of the safety of the piled bridge
foundations in liquefiable soils. Soils and Foundations 45(4): 13–26.

291

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Bhattacharya S, Blakeborough A and Dash SR (2008) Learning from collapse of piles in liquefiable soils.
Civil Engineering 161(2): 54–60 (special issue).
Bhattacharya S, Hyodo M, Goda K, Tazoh T and Taylor CA (2011) Liquefaction of soil in the Tokyo Bay area from
the 2011 Tohoku (Japan) earthquake. Soil Dynamics and Earthquake Engineering 31: 1618–1628.
Bhattacharya S, Tokimatsu K, Goda K, Sarkar R, Shadlou M and Rouholamin M (2014) Collapse of Showa Bridge
during 1964 Niigata earthquake: a quantitative reappraisal on the failure mechanisms. Soil Dynamics and
Earthquake Engineering 65: 55–71.
BIS (Bureau of Indian Standards) (2000) IS-1893. Criteria for earthquake resistant design of structures. Part 1:
General provisions and buildings. BIS, Delhi, India.
Bouzid DJ, Bhattacharya S and Dash SR (2013) Winkler Springs (p–y curves) for pile design from stress–strain of
soils: FE assessment of scaling coefficients using the mobilized strength design concept. Geomechanical
Engineering 5(5): 379–399.
BTL Committee (2000) Study on Liquefaction and Lateral Spreading in the 1995 Hyogoken-Nambu Earthquake.
Building Research Institute, Ministry of Construction, Tokyo, Japan. (In Japanese.)
Building Seismic Safety Council (2009) FEMA P-750. Recommended seismic provisions for new buildings and
other structures. Federal Emergency Management Agency of the US Department of Homeland Security,
Washington, DC, USA.
Cairo R and Dente G (2007) Kinematic interaction analysis of piles in layered soils. ISSMGE-ERTC 12 Workshop
‘Geotechnical Aspects of EC8’, Madrid, Spain. Pàtron Editore, Bologna, paper 13.
CEN (Comité Européen de Normalisation) (2004) EN 1998-5. Design of structures for earthquake resistance.
Part 5: Foundations, retaining structures and geotechnical aspects CEN, Brussels, Belgium.
Chai JC, Carter JP and Hayashi S (2007) Modelling strain-softening behaviour of clayey soils.
Lowland Technology International 9(2): 29–37.
Chau KT, Shen CY and Guo X (2008) Non-linear seismic soil-pile-structure interaction: shaking table tests and
FEM analyses. Soil Dynamics and Earthquake Engineering 29: 300–310.
Chopra AK and Goel RK (2000) Building period formulas for estimating seismic displacements. Earthquake
Spectra 16(2): 533–536.
Dammala, PK, Bhattacharya S, Krishna, A, Kumar SS and Dasgupta K (2017) Scenario based seismic re-qualification
of caisson supported major bridges – a case study of Saraighat Bridge. Soil Dynamics and Earthquake
Engineering 100: 270–275.
Dash SR and Bhattacharya S (2007) Criteria for design of piled foundations in seismically liquefiable deposits.
4th International Conference on Earthquake Geotechnical Engineering, Thessaloniki, Greece, paper 1724.
Dash SR, Bhattacharya S, Blakeborough A and Hyodo M (2008) p–y curve to model lateral response of
pile foundations in liquefiable soils. 14th World Conference on Earthquake Engineering (14WCEE), Beijing,
China.
Dash SR, Govindaraju L and Bhattacharya S (2009) A case study of damages of the Kandla Port and Customs
Office Tower supported on a mat–pile foundation in liquefied soils under the 2001 Bhuj earthquake.
Soil Dynamics and Earthquake Engineering 29(2): 333–346.
Dash SR, Bhattacharya S and Blakeborough A (2010) Bending–buckling interaction as a failure mechanism of
piles in liquefiable soils. Journal of Soil Dynamics and Earthquake Engineering 30(1–2): 32–39.
Dash S, Rouholamin M, Lombardi D and Bhattacharya S (2017) A practical method for construction of p–y curves
for liquefiable soils. Soil Dynamics and Earthquake Engineering 97: 478–481.
Dobry R and O’Rourke MJ (1983) Discussion on ‘Seismic response of end-bearing piles’ by Flores-Berrones R and
Whitman RV.’ Journal of Geotechnical and Geoenvironmental Engineering, ASCE 109: 778–781.
Editorial Committee for the Report on the Hanshin-Awaji Earthquake Disaster (1998) Report on the Hanshin-
Awaji Earthquake Disaster. Maruzen, Tokyo, Japan. (In Japanese.)
EERI (Earthquake Engineering Research Institute) (1991) Earthquake Spectra ‘Costa Rica earthquake
Reconnaisance Report’. Supplement B to Volume 7. EERI, Oakland, CA, USA.,

292

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Fujii S, Isemoto N, Satou Y and Kaneko O (1998) Investigation and analysis of a pile foundation damaged by
liquefaction during the 1995 Hyogoken-Nambu earthquake. Soils and Foundations, pp. 179–192 (special
issue).
Gazetas G, Fan K, Tazoh T, Shimizu K, Kavvadas M and Makris N (1992) Seismic pile-group–structure
interaction. In Piles under Dynamic Loads (Prakash S (ed.)), ASCE, Reston, VA, USA, pp 56–93.
Gerolymos N, Giannakou A, Anastasopoulos I and Gazetas G (2008) Evidence of beneficial role of inclined piles:
observations and summary of numerical analyses. Bulletin of Earthquake Engineering 6(4): 705–722.
Giannakou A, Gerolymos N and Gazetas G (2007) Kinematic response of groups with inclined piles. Proceedings
of the 4th International Conference on Earthquake and Geotechnical Engineering, Thessaloniki, Greece.
Haldar S, Sivakumar Babu SL and Bhattacharya S (2008) Buckling and bending response of slender piles in
liquefiable soils during earthquakes. Geomechanics and Geoengineering 3(2): 129–143.
Hamada M (1992a) Large ground deformations and their effects on lifelines: 1964 Niigata earthquake. In Case
Studies of Liquefaction and Lifelines Performance During Past Earthquakes (O’Rourke TD and Hamada M
(eds)). National Center for Earthquake Engineering Research, Buffalo, NY, USA, vol. 1, Technical Report
NCEER-92-0001.
Hamada M (1992b) Large ground deformations and their effects on lifelines: 1983 Nihonkai-Chubu earthquake. In
Case Studies of Liquefaction and Lifelines Performance During Past Earthquakes (O’Rourke TD and Hamada
M (eds)). National Center for Earthquake Engineering Research, Buffalo, NY, USA, vol. 1, Technical Report
NCEER-92-0001.
Idriss IM and Boulanger RW (2008) Soil Liquefaction During Earthquakes, 2nd edn. Earthquake Engineering
Research Institute. Earthquake Engineering Research, Oakland, CA, USA.
IRC (Indian Road Congress) (2014) IRC:78. Standard Specifications and Code of Practice for Road Bridges,
Section VII- Foundations and Substructures. IRC, Delhi, India.
Ishihara K (1997) Terzaghi oration: geotechnical aspects of the 1995 Kobe earthquake. Proceedings of 14th
International Conference on Soil Mechanics and Foundation Engineering, Hamburg, Germany, vol. 4,
pp. 2047–2073.
JRA (Japan Road Association) (2002) Specifications for highway bridges. Part V: Seismic design. JRA, Tokyo,
Japan.
Kastranta G, Gazetas G and Tazoh T (1998) Performance of three quay walls in Maya Wharf: Kobe 1995.
Proceedings of the European Conference on Earthquake Engineering, Paris, France, vol. 2, p. 65.
Kavvadas M and Gazetas G (1993) Kinematic seismic response and bending of free-head piles in layeres soil.
Géotechnique 43(2): 207–222.
Kawamura S, Nishizawa T and Wada H (1985) Damage to piles due to liquefaction found by excavation twenty
years after earthquake. Nikkei Architecture, May, pp. 130–134.
Kaynia AM and Mahzooni S (1996) Forces in pile foundations under seismic loading. Journal of Engineering
Mechanics, ASCE 122: 46–53.
Knappett JA and Madabhushi SPG (2009) Influence of axial load on lateral pile response in liquefiable soil. Part I:
physical modelling. Géotechnique 59(7): 571–581.
Krishnan R, Gazetas G and Velez A (1983) Static and dynamic lateral deflection of piles in non-homogeneous soil
stratum. Géotechnique 33(3): 307–326.
Kuhlemeyer RL (1979) Static and dynamic laterally loaded floating piles. Journal of the Geotechnical Engineering
Division, ASCE 105(GT2): 289–304.
Lombardi D (2013) Dynamics of Pile-supported Structures in Seismically Liquefiable Soils. PhD thesis, University
of Bristol, Bristol, UK.
Lombardi D, Dash SR., Bhattacharya S, Ibraim E, Wood DM and Taylor CA (2017) Construction of simplified
design p–y curves for liquefied soils. Géotechnique 67(3): 216–227.

293

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Margason E and Holloway DM (1977) Pile bending during earthquakes. Proceedings of the 6th World Conference
on Earthquake Engineering, Sarita Prakashan, Meerut, India, vol. II, pp. 1690–1696.
Matlock H and Foo SHC (1979) Axial analysis of piles using a hysteretic and degrading soil model. Proceedings of
a Conference on Numerical Methods in Offshore Piling, London, UK, pp. 127–134.
Mylonakis G (2001) Simplified model for seismic pile bending at soil layer interfaces. Soils and Foundations
41: 47–58.
Mylonakis G, Nikolaou A and Gazetas G (1997) Soil–pile–bridge seismic interaction: kinematic and inertial
effects. Part I: soft soil. Earthquake Engineering and Structural Dynamics 26: 337–359.
National Institute of Building Sciences (2012) 2009 NEHRP Recommended Seismic Provisions: Design Examples.
Federal Emergency Management Agency, Washington, DC, USA.
Nikolaou S, Mylonakis G, Gazetas G and Tazoh T (2001) Kinematic pile bending during earthquakes: analysis and
field measurements. Géotechnique 51(5): 425–440.
Novak M (1979) Soil-pile interaction under seismic loads. Proceedings of the International Symposium on
Numerical Methods in Offshore Piling, London, UK, pp. 41–50.
Padron LA, Aznarez JJ and Maero O (2008) Dynamic analysis of piled foundations in stratified soils by a BEM-
FEM model. Soil Dynamics and Earthquake Engineering 5: 333–346.
Penzien J (1970) Soil–pile foundation interaction. In Earthquake Engineering (Wiegel RL (eds)). Prentice-Hall,
Englewood Cliffs, NJ, USA.
Presidenza del Consiglio dei Ministri (2003) OPCM 3274. Primi elementi in materia di criteri generali per la
classificazione sismica del territorio nazionale e di normative tecniche per le costruzioni in zona sismica.
Presidenza del Consiglio dei Ministri, Rome, Italy. (In Italian).
Rouholamin M, Bhattacarya S and Orense RP (2017) Effect of initial relative density on the post-liquefaction
behaviour of sand. Soil Dynamics and Earthquake Engineering 97: 25–36.
Saitoh M (2005) Fixed-head pile bending by kinematic interaction and criteria for its minimization at optimal pile
radius. Journal of Geotechnical Engineering, ASCE 131(10): 1243–1251.
Seed RB, Dickenson SE and Mok CM (1991) Seismic response analysis of soft and deep cohesive sites: a brief
summary of recent findings. Proceedings of the CALTRANS First Annual Seismic Response Workshop,
Sacramento, California.
Shanker K, Basudhar PK and Patra NR (2007) Buckling of piles under liquefied soil conditions. Geotechnical and
Geological Engineering 25(3): 303–313.
Sica S, Mylonakis G and Simonelli AL (2007) Kinematic bending of piles: analysis vs. code provisions.
Proceedings of the 4th International Conference on Earthquake Geotechnical Engineering, Thessaloniki,
Greece.
Tahghighi H and Kanagai K (2007) Numerical analysis of non-linear soil-pile group interaction under lateral loads.
Soil Dynamics and Earthquake Engineering 5: 463–474.
Tokimatsu K and Suzuki H (2005) Effect of inertial and kinematic interactions on seismic behaviour of pile
foundations based on large shaking table tests. Proceedings of the 2nd CUEE Conference on Urban Earthquake
Engineering, Tokyo, Japan.
Tokimatsu K, Suzuki H and Sato M (2005) Effects of dynamic soil–pile structure interaction on pile stresses.
Journal of Structural and Construction Engineering, Architectural Institute of Japan 587: 125–132.
Tokimatsu K, Oh-oka H, Satake K, Shamoto Y and Asaka Y (1998) Effects of Lateral ground movements on
failure patterns of piles in the 1995 Hyogoken-Nambu earthquake. Proceedings of a Speciality Conference:
Geotechnical Earthquake Engineering and Soil Dynamics III, Seattle, WA, USA, pp. 1175–1186.
Wu G and Finn W (1997) Dynamic elastic analysis of pile foundations using finite element method in the time
domain. Canadian Geotechnical Journal 34: 44–52.

294

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Pile foundations

Yoshida N and Hamada M (1991) Damage to foundation piles and deformation pattern of ground due to
liquefaction-induced permanent ground deformations. Proceedings of the 3rd Japan–US Workshop on
Earthquake Resistant Design of Lifeline Facilities and Countermeasures for Soil Liquefaction, San Francisco,
CA, USA, pp. 141–161.
Zhang F, Kimura M, Nakai T and Hoshikawa T (2000) Mechanical behaviour of pile foundations subjected to
cyclic lateral loading up to the ultimate state. Soils and Foundations 40: 1–17.

295

Downloaded by [ Auckland University of Technology] on [27/01/19]. Copyright © ICE Publishing, all rights reserved.
Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

ISBN 978-0-7277-6166-8
https://doi.org/10.1680/sdof.61668.297
ICE Publishing: All rights reserved

Chapter 10
Analysis of foundations for major bridges
10.1. Introduction
This chapter

■ reviews the collapse of bridge foundations during earthquakes

■ reviews foundations for bridges together with closed-form solutions for the stiffness of deep
foundations for
– ﬂexible/rigid pile foundations
– circular and double-D caisson foundations
■ provides a methodology for the natural frequency estimation of bridge structures
■ gives an example calculation to predict the natural frequency of a caisson-supported bridge
■ compares the performance of bridges supported on caisson and pile foundations.

10.2. Survey of the performance of bridge foundations in seismic areas

From the point of view of foundations, bridge failures can be classiﬁed into the following types

■ Midspan collapse, where only a few spans in the middle of the bridge fail while the remaining part of
the bridge performs satisfactorily. The midspan fails due to unseating of the decks, possibly due to a
large displacement at the central pier heads.
■ Collapse or failure of abutments of the bridge due to lateral spreading. Large kinematic loads are
applied at the foundations.
■ Structural failure of the foundation, which may be common for slender piles in liqueﬁable zones.
In these cases, plastic hinges are formed in the pile.

Table 10.1 lists some of the failed major bridges in liqueﬁable soils during seismic events. Figures 10.1–10.10
presents the aerial/schematic views of the failed bridges. It may be noted from Table 10.1 and Figures (10.1–10.9)
that the central part (midspan) of the bridges collapsed.

10.2.1 Performance of the Rokko Bridge during the 2011 To ̄ hoku earthquake
(Japan)
The 2011 Great East Japan Earthquake occurred at 2:46 pm on 11 March 2011 with a moment magnitude
(Mw) of 9.0. Many highway bridges were damaged due to both the large ground motion and tsunami effects.
Very similar damage to that to the Showa Bridge in the 1964 Niigata earthquake (Japan) occurred to the
Rokko Bridge in Ibaraki Prefecture (Chen and Duan, 2003). The soil around the bridge liqueﬁed, and the two
central piers collapsed, resulting in the loss of the three central spans of the bridge (Figure 10.1).

297

Seismic Design of Foundations

Table 10.1 Abridged list of Bridge failures due to seismic liquefaction

Earthquake Bridge Remarks

2011 To ̄ hoku earthquake Rokko Bridge This steel girder bridge, which was supported by steel pile–bent
(Japan) columns located in Ibaraki prefecture, collapsed from the effects
of strong ground motion (see Figure 10.1)

2008 Sichuan/Wenchuan Miaoziping Bridge One of the ﬁve approaching spans of Miaoziping Bridge
earthquake (China) collapsed due to the earthquake (see Figure 10.2)

2008 Sichuan/Wenchuan Gaoyuan Bridge The middle span of the bridge fell off the piers,
earthquake and evidence of liquefaction was present (see Figure 10.3)

1995 Kobe earthquake Nishinomiya Liquefaction was prevalent at the bridge site, and the simply
(Japan) Bridge supported span next to the arched span fell off the piers
(see Figure 10.4)

1991 Costa Rica Rio Viscaya Bridge One internal supporting pier was missing and was assumed to
earthquake have settled due to liquefaction

1990 Luzon earthquake Magsaysay Bridge Piers settled and/or tipped over. Two few bridge decks dropped
(Philippines) into the water (see Figure 10.5)

1976 Tangshan Zhuacun Bridge The girders of the middle spans (Nos. 10 and 11) dropped off
earthquake (China) the piers (see Figure 10.6)

1976 Tangshan Shahe Bridge Piers inclined, and among them one crashed down, and the
earthquake girders supported by that pier fell. The maximum relative dis-
placement at the level of the top of the pier between the girders
and piers was 1.05 m in the direction of the bridge axis

1976 Moro Gulf Quirino Bridge The midspan collapsed in this truss bridge (see Figure 10.7)
earthquake (Philippines)

1975 Haicheng Panshan Highway Evidence suggests that one pier (No. 7) out of 14 piers sank
earthquake (China) Bridge almost 15 cm. Other piers inclined, and cracks were induced in
them (see Figure 10.8)

1964 Niigata earthquake Showa Bridge The deck near the middle piers, and those adjacent to it fell into
(Japan) the river (Figure 10.9)

298

Seismic Design of Foundations

Analysis of foundations for major bridges

Figure 10.1 Failure of the Rokko Bridge in Ibaraki Perfecture

10.2.2 Performance of the Miaoziping Bridge, 2008 Wenchuan earthquake

(China)
The Wenchuan earthquake occurred in Sichuan Province, China, at 2:28 pm (local time) on 12 May 2008. It
was estimated to have an Mw of 8.0 by California Earthquake Authority and 7.9 by the US Geological Survey.
The main structure of the Miaoziping Bridge is a 450 m-long three-span continuous prestressed concrete
frame bridge with a main span of 220 m. The approach spans are of prestressed concrete T girders that are
50 m long and simply supported with concrete slabs. All the spans were supported by nearly 100 m-tall
reinforced concrete piers resting on pile foundations in the dam reservoir. One of the two end-spans of a ﬁve-
span continuous T-girder bridge collapsed, as shown in Figure 10.2. It was reported that the distance between
the two adjacent piers that supported the fallen span was extended by about 0.69 m at the top, based on
measurements taken after the earthquake (Kawashima et al., 2009).

299

Seismic Design of Foundations

Figure 10.2 Failure of the Miaoziping Bridge (Kawashima et al., 2009)

10.2.3 Performance of the Gaoyuancun Bridge, 2008 Wenchuan earthquake

(China)
This bridge consisted of four spans, each of 23.45 m. The girder of span 2 fell into the river, as shown in the
Figure 10.3 on the side of Gaoyuan due to liquefaction at the site.

Figure 10.3 Failure of the Gaoyuan Bridge (Han et al., 2009)

300

Seismic Design of Foundations

Analysis of foundations for major bridges

10.2.4 Performance of the Nishinomiya Bridge, 1995 Kobe earthquake (Japan)

This arch bridge linked two reclaimed islands, Nishinomiya and Koshien. The top 6–8 m portion basi-
cally comprised loose reclaimed sand, with an standard penetration test (SPT) N value of 10 or less.
During the 1995 Kobe earthquake the soil liqueﬁed, and the span resting in between pier P99 and P98
fell off (Figure 10.4).

Figure 10.4 Failure of the Nishinomiya Bridge (Alim, 2014) (Ref: Alim, Haﬁzul (2014) ’Reliability Based Seismic
Performance Analysis of Retroﬁtted Concrete Bridge Bent‘, Thesis, Lamar University, Texas, US)

301

Seismic Design of Foundations

Figure 10.5 Magsaysay Bridge (a) before and (b) after the earthquake (Ishihara et al., 1993)

144 m
15 m 23 m 23 m 23 m 15 m 15 m 15 m 15 m

AA AB
P5 P6 P7
P1 P4
P3
P2
(a)

AA AB
P5 P6 P7
P4
P1 P3
P2
(b)

10.2.5 Performance of the Magsaysay Bridge, 1990 Luzon earthquake

(Philippines)
The Luzon earthquake occurred on 16 July 1990 in the middle of Luzon island, with an Mw of 7.8.
The Magsaysay Bridge is basically a eight-span reinforced concrete bridge supported by seven piers and
two abutments. The bridge is 144 m long, and comprised eight simply supported reinforced concrete
girders resting on piers supported by concrete piles about 10 m in length. Piers P1, P2 and P3 (Figure 10.5)
settled and/or tipped over. Few bridge spans dropped into the water. The ﬁrst and second piers (P1 and P2)
moved towards the river channel (P3), and the third pier sank about 2 m due to loss of bearing capacity of the
liqueﬁed riverbed deposit (Ishihara et al., 1993).

10.2.6 Performance of the Zhuacun Bridge, 1976 Tangshan earthquake (China)

The Tangshan earthquake occurred on 28 July 1976 with an Mw of 7.8. The Zhuacun Bridge crossing the
Luanhe River at Zhuacun Village in Qianan County was an important major structure on the Lengkou–
Daqinghekou Highway. The superstructure consisted of 31 spans of 22.20 m- long pre-cast reinforced
concrete T girders, which were simply supported. The total length of the bridge was 695 m, and the deck
width was 7 m with a 0.75 m sidewalk on each side. The girders of spans 10 and 11 from the southern bank
dropped due to seismic shaking (Figure 10.6).

10.2.7 Performance of the Shahe Bridge, 1976 Tangshan earthquake (China)

The Shahe River Bridge consisted of 12 spans of reinforced concrete T girders and seven spans of rein-
forced concrete voided slabs. The width of the deck was 7 m, with a 0.75 m sidewalk on each side. The
main girders of spans 12 and 13, which were at the boundary of these two different types of slabs, dropped
on the ground.

302

Seismic Design of Foundations

Analysis of foundations for major bridges

Figure 10.6 Side view of the failure of the Zhuancun Bridge (Huixian et al., 2002)

10.2.8 Performance of the Quirino Bridge, 1976 Moro Gulf earthquake

(Philippines)
The 1976 Moro Gulf earthquake and tsunami took place on 17 August 1976 at 00:11 local time near the
islands of Mindanao and Sulu in the Philippines. The Mw was calculated to be 8.0. The Quiriono Bridge was
badly damaged, as shown in the Figure 10.7.

303

Seismic Design of Foundations

Figure 10.7 Side view of the failure of the Quirino Bridge

10.2.9 Performance of the Panshan Highway Bridge, 1975 Haicheng earthquake

(China)
The Haicheng earthquake occurred on 4 February, 1975 in the northeastern region of China with an Mw of
7.3. The bridge had reinforced concrete superstructure with 14 spans. Each of two abutments and 13 piers
were supported by four reinforced concrete piles with a diameter of approximately 1 m and a length of
approximately 30 m. Due to the liquefaction caused by the earthquake, pier 7 sunk, and spans G7, G8, G9 and
G12 fell during the major aftershock (Figure 10.8).

Figure 10.8 (a) Panshan Bridge after the earthquake and (b) the soil stratigraphy near pier 4 (P4) (Shengcong and
Tatsuoka, 1984)
15 cm

17 cm 45 cm 32 cm 55 cm 66 cm 2 cm 38 cm 2 cm 2 cm 4 cm 13 cm 10 cm 5 cm

G1 G2 G3 G4 G5 G6 G7 G8 G9 G10 G11 G12 G13 G14

Panshan Yingkou
1.5 m

11.5 m

6m
Clay
7m
1m Fine sand
2m
P0 P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 2m
Silt
5m

Before the earthquake

After the earthquake

304

Seismic Design of Foundations

Analysis of foundations for major bridges

Figure 10.9 Failure of the Showa Bridge (Bhattacharya et al., 2014)

10.2.10 Performance of the Showa Bridge, 1964 Niigata earthquake (Japan)

The Niigata earthquake occurred on the 14 June 1964, and registered an Mw of 7.6. Located some 55 km from
the epicentre, crossing the Shinano River, the Showa Bridge (a simple steel girder bridge with pile foun-
dations) was one of the bridges that collapsed as a result of the earthquake. The total length of the bridge was
about 307 m. The bridge had 12 composite girders, and its breadth was about 24 m. The length of the main
span was about 28 m, and length of the side span was about 15 m (Fukuoka, 1966). A view of the collapsed
Showa Bridge from the southwest side is shown in Figure 10.9. During the earthquake the bridge site was
subjected to extensive liquefaction and lateral spreading. It can be observed that the central span of the bridge
collapsed due to the deck unseating and falling into the river. Several failure mechanisms were identiﬁed as
the cause, such as bending, buckling, the combined action of bending and buckling, and dynamic changes in
the structure, which were all the result of the seismically induced liquefaction (Bhattacharya et al., 2014;
Mohanty et al., 2017).

10.3. Bridge foundations

Bridge foundations include both shallow and deep foundations. Shallow foundations are typical for rock sites.
However, for the case of bridges with large spans and heavy trafﬁc loads or ground with soft soils or rivers
with scour/strong current, deep foundations are mostly preferred. Pile and caisson foundations are traditional
deep foundations used to support major bridges; see Figure 10.10 for a schematic classiﬁcation of deep
foundations based on the slenderness ratio and the depth of embedment in accordance with the work of
Gerolymos and Gazetas (2006). Methods of analysis of deep foundations have been described in detail in the

305

Seismic Design of Foundations

Figure 10.10 Crude classiﬁcation of deep foundations based on the slenderness ratio and the depth of embedment

Pier

L L

D B
Pile Caisson
L/D > 8 Rigid: 0.5 < L/B > 4
Flexible: 3 < L/B > 8

Chapter 9, and the present chapter focuses on the application of the described methodologies to major bridges
supported on deep foundations. One of the important aspect of seismic design is the stiffness of foundations.
Closed-form solutions for the analysis of laterally loaded deep foundations (both piles and caissons) are
introduced in this chapter. The application of stiffness in the calculation of the fundamental frequency of a
bridge together with examples are also provided.

10.4. Caisson foundations

Caissons (often called well foundations) are the preferred foundations for heavy superstructure loads and high
lateral loads. Therefore, they are typical in a long-span bridge such as a suspension or cable-stayed bridge.
These types of foundations are traditionally used for major river bridges with a high discharge rate and also in
seismically active regions (e.g. river bridges in seismic zones of India). Caissons offer many advantages
compared with traditional pile foundations: for example, they can support high unsupported lengths
(potentially the result of liquefaction and scour). The following caisson shapes are used in practice

■ circular
■ twin circular
■ double D
■ double octagonal
■ single and double rectangular

306

Seismic Design of Foundations

Analysis of foundations for major bridges

Of these, circular, twin circular and double-D caissons are the most used types in many countries.
Figure 10.11 shows bridge structures supported on circular, twin circular and double-D caissons, along with
their typical geometry. Circular caissons are simple to construct, easy to sink in and their (theoretical) stiffness
is uniform in all directions. The tilt of circular caissons can be effectively controlled. Practically, their size is
limited to a diameter of about 12 m, because of which they are considered ideal only for single-lane trafﬁc
bridges with a narrow pier width. For multi-lane trafﬁc and wide piers, twin circular and double-D caissons
are traditionally preferred, due to their high seating width.

For twin circular caissons, two widely spaced (<3 m) individual caissons are connected with a well cap at
the top, over which a monolithic pier can be built. They are ideal for shallow water, as the depth of
sinking increases the possibility of tilt. Also, twin circular caissons have a tendency to tilt towards each
other due to the spacing limitation in between. The common well cap should be laid in such a way so as
to account for the possibility of differential settlement during construction or the lifetime of the caissons.
Another drawback to these caissons is the scouring of soil in between the piers due to the formation of
wake vortices (Figure 10.11(b)).

Double-D caissons are probably the ideal type of foundation for wide piers with multi-lane trafﬁc together
with a double-deck structure. Their shape allows easy casting and sinking, and they do not require much
construction effort due to the presence of two dredge holes. They are also relatively effective in controlling the
scour effects because of their intrinsic shape, which does not allow turbulent ﬂow to be continuous along the
width of their cross-section (Figure 10.11(c)).

Figure 10.11 Typical bridges supported on (a) circular, (b) twin circular and (c) double-D caisson foundations

Railway

Pier

Well cap 10–20 m

5 to 12 m
5 –10 m

Wake
vortices
(a) (b) (c)

307

Seismic Design of Foundations

10.5. Stiffness of deep foundations

For performance-based design calculations, deep foundations should principally be designed to remain
elastic, and the stiffness of the foundation is a prerequisite for such calculations. The stiffness is deﬁned as
the force (or moment) to be applied to the foundation top to produce a unit displacement (or rotation)
along the same direction, keeping the displacements and rotations in the other directions zero (CEN,
2004)). A method adopted in Eurocode 8 (CEN, 2004)), called the lumped spring approach, is considered
appropriate to model the behaviour of deep foundations resisting the static and dynamic loading condi-
tions. In simple terms, the foundation and surrounding soil are replaced by a set of elastic springs at the
foundation head level (Figure 10.12(b)), namely a lateral spring KL, a rotational spring KR, a coupling
spring KLR and a vertical spring KV. Usually, foundations are vertically stiff, and the axial stiffness is often
avoided in the analysis (i.e. KV is omitted from dynamic calculations). With the use of elastic stiffness
values, one can reliably estimate the response of the structure by performing an eigenvector modal analysis
in standard ﬁnite-element programs such as SAP 2000.

Based on the elasto-dynamic Eurocode 8 approach (CEN, 2004), any layered soil stratum can be idealised
using the stiffness (E) variation along the depth (z) with any of three proﬁles (Figure 10.13(a))

■ homogeneous – which can be considered appropriate for over-consolidated soils

■ linear inhomogeneity – typical for normally consolidated soils
■ parabolic heterogeneity – ideal for sandy soils.

Figure 10.12 (a) Real ﬁeld conditions. (b) Idealised mathematical model

Z
Superstructure mass
(msuperstructure)
Pier
(E, I, L, D, m)pier
Lpier

Bed level Water (E, I, L, D)foundation

Llp
Length, L

KMM KHH
X
Caisson
KMH

Diameter, D
(a) (b)

308

Seismic Design of Foundations

Analysis of foundations for major bridges

Figure 10.13 (a) Idealisation of soil proﬁles. (b) Variation of the soil stiffness with depth

Young’s modulus of soil, E

Depth, z

Homogeneous Linear Parabolic

(uniform) variation variation
(a)

1 E/Es

Homogeneous
(over consolidated clay)
Stiffness at 1 diameter depth
1
Linear inhomogeneous
(sandy soils)
z/D

Parabolic inhomogeneity
(normally consolidated clays)

(b)

Figure 10.13(b) shows the idealisation of the soil stiffness with depth. The stiffness variation with depth (z)
can be represented by the following formula according to Eurocode 8 (CEN, 2004; Jalbi et al., 2018; Latini
and Zania, 2017; Shadlou and Bhattacharya, 2016)
z a
E = Es (10:1)
D

where Es is the stiffness of the soil at one diameter depth D, z is the depth and a is the degree of inhomo-
geneity. Homogeneous soils will have an a value of zero, linear inhomogeneity has a value of 1 while a
varies from 0 to 1 for the parabolic case.

309

Seismic Design of Foundations

10.5.1 Flexible pile foundations

Eurocode 8 (CEN, 2004) suggested the formulas in Table 10.2 to estimate the stiffness terms of the ﬂexible pile
foundations, which are a function of the geometry of the pile (pile diameter D) and the pile–soil stiffness ratio Ep/Es.

10.5.2 Rigid foundations

Shadlou and Bhattacharya (2016) proposed closed-form solutions to estimate the impedance functions of
deep foundations exhibiting rigid behaviour based on advanced numerical analyses. Table 10.3 lists the
impedance functions of rigid deep foundations. Unlike ﬂexible foundations, rigid foundations depend on the
aspect ratio of the pile L/D and Poisson’s ratio of the soil f(vs ) .

Table 10.2 Impedance functions for ﬂexible pile foundation

Ground proﬁle Homogeneous Linear Parabolic

KHH 0:21 0:35 0:28

Ep Ep Ep
DEs 1:08 0:60 0:79
Es Es Es

KMM 0:75 0:80 0:77

Ep Ep Ep
D3 Es 0:16 0:14 0:15
Es Es Es

KMH 0:50 0:60 0:53

Ep Ep Ep
D2 Es − 0:22 − 0:17 − 0:24
Es Es Es

CEN (2004)

Table 10.3 Impedance functions for rigid deep foundations

Ground proﬁle Homogeneous Linear Parabolic

KHH 0:62 1:53 1:07

L L L
DEs f(vs ) 3:2 2:35 2:65
D D D

KMM 2:50 3:45 3:00

L L L
3
D Es f(vs ) 1:65 1:58 1:63
D D D

KMH 1:56 2:50 2:00

L L L
D2 Es f(vs ) − 1:80 − 1:80 − 1:80
D D D

Shadlou and Bhattacharya (2016)

f(vs ) = 1 + 0:6jvs − 0:25j

310

Seismic Design of Foundations

Analysis of foundations for major bridges

10.5.2.1 Example calculation

A river bridge supported on a rigid circular caisson will be used to demonstrate the methodology for
estimating the stiffness: the Tista Bridge in West Bengal (India), which is close to the Himalayan seismic
belt. The bridge is supported on single circular caissons of diameter D = 6.6 m and an embedment depth
L = 26 m below the river bed level. For simpliﬁcation, the stiffness variation is assumed linear with a value of
75 MPa at a depth of one caisson diameter. Poisson’s ratio is taken as 0.30. The methodology for estimating
the stiffness is given below in three steps.

Step 1. Estimation of the lateral stiffness KHH

1:53
KHH 26
= 2:35 = 19:145 (10:2)
DEs f(vs ) 6:6

The effect of Poisson’s ratio can be considered as

f(vs ) = 1 + 0:6j0:30 − 0:25j = 1:03 (10:3)

KHH = 19:145 6:6 75 000 1:03 = 9 761 0782 kN=m

(10:4)
= 9:761GN=m

Step 2. Estimation of the coupled stiffness KMH

2:5
KMH 26
= −1:8 = −55:43 (10:5)
D2 Es f(vs ) 6:6

Step 3. Estimation of the rotational stiffness KMM

3:45
KMM 26
3
= 1:58 = 178:95 (10:6)
D Es f(vs ) 6:6

KMM = 178:952 6:63 75 000 1:03 = 3 974 376 584 kNm=rad

(10:7)
= 3974 GNm=rad

10.5.3 Rigid double-D caisson foundations

Double-D caissons can be considered to be a combination of a rigid rectangular section and a solid circular
section of diameter D (see Figure 10.14 for details). The geometry may have a signiﬁcant effect on the
stiffness of deep rigid foundations, and has been investigated by Dammala et al. (2018). Considering the
shape of double-D caissons, Dammala et al. (2019) proposed closed-form solutions for the stiffness esti-
mation of double-D caissons embedded in three different soil proﬁles. The formulas are based on advanced

311

Seismic Design of Foundations

Figure 10.14 Idealised shape of a double-D caisson foundation

B
D/2 B/2

D
Y–Y
MY

X–X

three-dimensional ﬁnite-element analyses, and provide solutions based on the geometry of the double-D
caisson (the width of the rectangular portion B and the diameter of the circular section D). The double-D
stiffness values can be obtained by multiplying the circular rigid caisson stiffness by the multiplication
constants (a, b, y) for lateral, rocking and coupled stiffness, respectively, as follows

KHH-caisson = a ðKHH Þ (10:8)

KMM-caisson = b ðKMM Þ (10:9)

KMH-caisson = y ðKMH Þ (10:10)

The stiffness constants are provided below for three different soil proﬁles.

For homogeneous soils

D B
a = 0:15 0:17 + 0:84 +1 (10:11)
2:5 D

D B
b = 0:22 0:22 + 0:83 +1 (10:12)
2:5 D

D B
y = 0:17 0:17 + 0:85 +1 (10:13)
2:5 D

312

Seismic Design of Foundations

Analysis of foundations for major bridges

For the linear variation of stiffness

D B
a = 0:17 0:86 ln +1 +1 (10:14)
2:5 D

D B
b = 0:20 0:98 ln +1 +1 (10:15)
2:5 D

D B
y = 0:18 0:85 ln +1 +1 (10:16)
2:5 D

For the parabolic variation of stiffness

D B
a = 0:19 0:80 ln +1 +1 (10:17)
2:5 D

D B
b = 0:22 0:85 ln +1 +1 (10:18)
2:5 D

D B
y = 0:18 0:88 ln +1 +1 (10:19)
2:5 D

10.5.3.1 Example calculation

The Saraighat Bridge (Figure 10.15) is located in Assam, a northeastern state of India and a highly active
seismic region. It is a typical example of a bridge supported on double-D caissons. The depth of embedment
of the caissons is 41 m, the width is 9.2 m and the diameter is 9.6 m (see Figure 10.15). The soil at the centre
of the bridge consists mostly of loose to dense silty sand, with the density increasing with depth, which can be
approximated as a linearly varying stiffness proﬁle, using the formulas below.

The soil stiffness at one diameter depth Es is approximately 130 MPa. The stiffness terms of a double-D
caisson (KHH, KMM and KMH) can be estimated based on Equations 10.8–10.10, in a simple three-step
procedure.

Step 1. The ﬁrst step involves the determination of the stiffness for a circular shaft (B = 0) of given diameter.

The effect of Poisson’s ratio can be considered as

f(vs ) = 1 + 0:6j0:30 − 0:25j = 1:03 (10:20)

313

Seismic Design of Foundations

Figure 10.15 (a) Saraighat Bridge; (b) central portion of the bridge; (c) cross-section through the foundation (Z–Z);
(d) cross section through the pier (X–X); (e) typical soil proﬁle at the centre of the bridge (Dammala et al., 2017a)

1298
33.22 10 × 123.07 m C/C 33.22

A1 A2

P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11

(a)
Roadway
Railway

14.0 11 Loose to medium-dense

x x fine sand
SPT Navg = 8

River bed level

z z
Medium to dense
14 silty sand mixed
41 56
with gravels
SPT Navg = 26

P6
Very dense to
17
dense grey, silty sand
(b) SPT Navg = 36

Very hard and

4.20 4.5 6 stiff clayey silt
Pier
SPT Navg = 47
16.40 8.7

Deep dense sand

2.70 4.5 14 mixed with gravels
Pier
SPT Navg = >100

9.60 (e)
9.60
(c) (d)
All dimensions are in metres

314

Seismic Design of Foundations

Analysis of foundations for major bridges

(a) The lateral stiffness of a rigid circular shaft KHH with D = 9.6 m, Es = 130 MPa and length L = 40 m is

1:53
KHH 40
= 2:35 = 20:861 (10:21)
DEs f(vs ) 9:6

KHH = 20:861 9:6 130 000 1:03 = 26:81 106 kN/m

(10:22)
= 26:81 GN/m

(b) The rocking stiffness of a circular shaft KMM according to Table 10.3 is
3:45
KMM 40
3
= 1:58 = 217:294 (10:23)
D Es f(vs ) 9:6

KMM = 217:294 9:63 130 000 1:03 = 25742 106 kN m/rad

(10:24)
= 25742 GN m/rad

(c) Similarly, the coupled stiffness of a circular shaft KMH is

2:5
KMH 40
= −1:8 = −63:80 (10:25)
D2 Eso f(vs ) 9:6

KMH = −63:80 9:62 130 000 1:03 = −787:325 106 kN

(10:26)
= −787:325 GN

Step 2. This step involves the determination of multiplication constants (a, b and y) to account for the width
(B = 9.2 m) of the double-D caissons using the proposed formulas.

The value of a according to Equation 10.14 is

9:6 9:2
a = 0:170 0:86 ln +1 + 1 = 1:351 (10:27)
2:5 9:6

The value of b according to Equation 10.15 is given as

9:6 9:2
b = 0:20 0:98 ln +1 + 1 = 1:444 (10:28)
2:5 9:6

315

Seismic Design of Foundations

The coupled stiffness constant y according to Equation 10.16 is

9:6 9:2
y = 0:18 0:85 ln +1 + 1 = 1:369 (10:29)
2:5 9:6

Step 3. The ﬁnal step involves multiplication of the constants by the circular stiffness values in order to obtain
the geometry-based stiffness of double-D caissons.

The lateral stiffness of a double-D caisson KHH-caisson according to Equation 10.8 is

KHH-caisson = 1:351 26:81 = 36:22 GN/m (10:30)

The rocking stiffness of a double-D caisson KMM-caisson according to Equation 10.9 is

KMM-caisson = 1:444 25 742 = 37171 GN m/rad (10:31)

The coupled stiffness of a double-D caisson KMH-caisson according to Equation 10.10 is

KMH-caisson = 1:369 ( − 787:325) = −1077:848 GN m/rad (10:32)

Once the stiffness of the foundation is established, the results can be used for the simpliﬁed dynamic
analysis of the structures supported on such foundations. Solutions are provided by Arany et al. (2016)
and Shadlou and Bhattacharya (2016) to identify the fundamental frequency of offshore wind turbines
using the impedance functions, which are based on beam theory and single-degree-of-freedom systems
with lumped masses. This essentially means that these solutions can also be used for other types of stiff
vibrating masses such as a concrete deck supported on bridge piers, ﬁnally resting on deep foundations as
shown in Figure 10.12.

10.6. Methodology for fundamental frequency estimations

Arany et al. (2016) provided closed-form solutions to estimate the fundamental frequency of offshore wind
turbines supported on deep foundations such as piles and caissons. The same methodology can also be
applied to bridge systems, and, hence, the methodology in obtaining the fundamental frequency of the bridge
pier supported on deep foundation will be described brieﬂy here.

It is first required to calculate the fixed-base natural frequency ffb, assuming the fixity of the foundation at
the bed level (ground level). Then, empirical foundation flexibility coefficients (CHH, CMM) accounting for
the soil foundation interaction are evaluated based on the values of KHH, KMM and KMH. The product of the
flexibility coefficients with the fixed-base natural frequency yields the soil–foundation–superstructure system
natural frequency fnz as

fnz = CHH CMM ffb (10:33)

316

Seismic Design of Foundations

Analysis of foundations for major bridges

ffb accounts for the non-uniform geometry of the superstructure; that is, enhanced stiffness at the bottom of the
superstructure through the factor CMP (Arany et al., 2016)
rffiffiffiffi
1 k
ffb = = CMP ffb:pier (10:34)
2π m

where ffb.pier is the ﬁxed-base natural frequency of the bridge pier

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 3Epier Ipier
ffb:pier = 3
(10:35)
2π L (msuperstructure + 0:227mpier )

where Epier and Ipier are the elastic modulus and the moment of inertia of the bridge pier, respectively, and
msuperstructure and mpier are the mass of the superstructure (the deck and the loads acting) and the pier,
respectively.

CMP, the ﬂexibility coefﬁcient of the foundation, can be estimated as

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
CMP = (10:36)
1 + ð1 + y Þ3 c − c

where c and y are the bending stiffness ratio and the foundation–pier length ratio, respectively, and

Epier Ipier
c= (10:37)
Efoundation Ifoundation

Llp
y = (10:38)
Lpier

where Lpier is the height of the pier; Efoundation and Ifoundation are the modulus of elasticity of the foundation
material and the moment of inertia of the pier, respectively; and Llp is the length of the foundation from the
mudline to the bottom of the pier.

The non-dimensional foundation stiffness parameters (hHH, hMM and hMH) need to be determined in order to
find the flexibility coefficients (CHH and CMM)

KHH L3pier
hHH = (10:39)
Epier Ipier

KMM Lpier
hMM = (10:40)
Epier Ipier

KMH L2pier
hMH = (10:41)
Epier Ipier

317

Seismic Design of Foundations

Once the values of hHH, hMM and hMH are known, the ﬂexibility coefﬁcients can be calculated using the
relationships

1
CMM = 1 − (10:42)
1 + aðhMM − h2MH / hHH Þ

1
CHH = 1 − (10:43)
1 + bðhHH − h2MH / hMM Þ

where a and b are the constants, estimated to be 0.6 and 0.5, respectively (Arany et al., 2016).

10.6.1 Example calculation of fnz for a double-D caisson-supported bridge

The Saraighat Bridge (see Figure 10.15(a)) in Assam, India, a region of high seismicity, has been chosen
to demonstrate the applicability of calculated stiffness values in evaluating the fundamental frequency of a
bridge structure. The bridge, with a span of 1298 m, is a double Warren steel truss type with the lower
deck accommodating a single-lane railway track while the upper deck houses a double-lane roadway (see
Figure 10.15). The soils in this region are highly prone to liquefaction (Dammala et al., 2017b) and, hence,
researchers have initiated requaliﬁcation strategies by monitoring the response of the structure in terms of
fundamental frequencies using high-quality accelerometers. The fundamental frequency thus estimated has
a range between 0.90 and 0.92 Hz (Debnath et al., 2015).

The fundamental frequency fnz of the soil–caisson–pier system is estimated using the evaluated stiffness
functions. The lateral, rocking and coupled stiffnesses of the Saraighat Bridge foundation were evaluated in
Section 10.5.3.1, and are listed in Table 10.4 along with the input parameters. Once the stiffness functions are
evaluated, fnz of the soil–caisson–pier system can be approximated using simple cantilever beam theory
(Arany et al., 2016). The procedure of obtaining fnz is described stepwise below.

Step 1: calculate the fixed-base natural frequency. It is first necessary to calculate the fixed-base natural
frequency ffb, assuming the fixity of the foundation at the bed level (ground level). Both the bending and
foundation–pier length ratios need to be evaluated

20 20:14
c= = 0:0216 (10:44)
20 929:3

15
y = = 0:83 (10:45)
18

Calculate the substructure flexibility coefficient, to account for the enhanced stiffness
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
CMP = = 0:95 (10:46)
1 + ð1 + 0:83Þ 0:0216 − 0:0216
3

318

Seismic Design of Foundations

Analysis of foundations for major bridges

Table 10.4 Input parameters for the frequency estimation

Parameter Value

Lateral stiffness, KHH-caisson: GN/m 36.22

Rocking stiffness, KMM-caisson: GN.m/rad 37171

Coupled stiffness, KMH-caisson: GN –1073.85

Mass of the deck live and dead load, msuperstructure: t 3500

Mass of the pier, mpier: t 750

Length of the pier, Lpier: m 18

Length of foundation above the mudline, Llp: m 15

Diameter of the pier, Dpier: m 4.5

4
Second moment of the area of the pier, Ifoundation: m 20.14

Second moment of the area of the well, Ifoundation: m4 929.3

Modulus of elasticity of concrete, Epier or Efoundation: GPa 20

Obtain the ﬁxed-base natural frequency of the caisson-supported pier

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 3 20 109 20:14
ffb,pier = = 1:19 Hz (10:47)
2π ½3 500 000 + (33 750 000) / 140 183

ffb = 1:19 0:95 = 1:1305 Hz (10:48)

Step 2: calculate the non-dimensional foundation stiffness parameters. The non-dimensional foundation
stiffness parameters (hHH, hMM and hMH) are required in order to find the flexibility coefficients (CHH and
CMM)

36:22 183
hHH = = 524:41 (10:49)
20 20:14

−1074 182
hMH = = −863:90 (10:50)
20 20:14

37171 18
hMM = = 1661 (10:51)
20 20:14

319

Seismic Design of Foundations

Step 3: calculate the foundation flexibility factors. Once the values of hHH, hMM and hMH are known, the
flexibility coefficients can be calculated using the relationships below proposed by Arany et al. (2016)

1
CMM ðhHH , hMH , hMM Þ = 1 − = 0:993 (10:52)
1 + 0:6 ½1661 − (− 863:92/524:4)

1
CHH ðhHH , hMH , hMMR Þ = 1 − = 0:974 (10:53)
1 + 0:5½524:41 − (− 863:92/1661)

Step 4: calculate the ﬂexible natural frequency of the system. The natural frequency based on Equation
10.33 is therefore given by

fnz = 0:993 0:974 1:1305 = 1:0934 Hz (10:54)

The obtained natural frequency of the bridge pier system (1.093 Hz) is very close to the measured frequency
of the bridge (0.9–0.92 Hz).

10.6.2 Liquefaction effects on caisson-supported bridges

The Saraighat Bridge was designed and constructed in the late 1950s, before our present understanding of
seismic liquefaction (1964). A performance assessment study was carried out to determine the maximum
possible seismic event, through a ground response analysis. The surficial loose sandy stratum (to 11 m)
was found to be liquefiable according to the semi-analytical formulas of Idriss and Boulanger (2006). The
length of the embedment was decreased accordingly, to account for the increased unsupported length in
the stiffness formulations, and the resulting fundamental frequencies at different depths of liquefaction
estimated. The time period (the inverse of the fundamental frequency) variation with increasing depth of
liquefaction is shown in Figure 10.16(a). It can be seen that the change in the time period is less than
20% even for the full depth of liquefaction (top 11 m). The resulting spectral displacements are estimated
and shown in Figure 10.16(b). The net increase in the spectral displacements reaches 130 mm, which is
much lower than the provided seating length of the deck on the pier. Hence, it can be concluded that the
caisson foundation provides sufficient support even in the case of significant liquefaction of the sur-
rounding soil.

10.7. Dynamic behaviour of pile-supported bridges

10.7.1 Collapse of the Showa Bridge in liqueﬁable soil during the 1964 Niigata
earthquake (Japan)
The case study of the collapse of the Showa Bridge is very well documented in Bhattacharya et al. (2014).
The Niigata earthquake occurred on 14 June 1964, and registered an Mw of 7.6. The Showa Bridge,
spanning the Shinano River, was located some 55 km from the epicentre, and collapsed as a result of the
earthquake. The total length of the bridge was about 307 m. The bridge had 12 composite girders with a
main span length of 28 m and a breadth of about 24 m (Fukuoka, 1966). The collapsed view from the
southwest of the Showa Bridge is shown in Figure 10.9.

320

Seismic Design of Foundations

Analysis of foundations for major bridges

Figure. 10.16 (a) Variation in the natural period and the percentage increase in the natural period; (b) Sd and the
percentage change in Sd with Dliq (Dammala et al., 2017a)

1.5 50

1.4 40
Time period

Change in the time period: %

Change in the time period

1.3 30
Time period: s

1.2 20

1.1 10

1.0 0
0 3 6 9 12
Depth of liquefaction: mm
(a)
800 50
Spectral displacement
Change in spectral displacement
Cut off for unseating = 750 mm
750

Change in the spectral displacement: %

40
Spectral displacement: mm

700
Margin for unseating failure for
the maximum possible event 30

650
Dliq for Mw 8.1_0.186g
Dliq for Mw 7.0_0.186g
20
600 Dliq for Mw 6.0_0.186g

10
550

500 0
0 3 6 9 12
Depth of liquefaction: mm
(b)

321

Seismic Design of Foundations

10.7.1.1 Liquefaction proﬁle

The Showa Bridge was situated in the coastal alluvial plain of the city of Niigata, which consists mainly of
marine sediments. The sand was uniformly graded medium sand, and its 60% diameter (D60) was about
0.3 mm (Fukuoka, 1966). From published SPT data (Iwasaki, 1986), it is clear that the upper strata was
liquefiable, as the average, pre-earthquake N value was found to be about 9.3 for the top 15 m strata of the
Showa Bridge site. Hamada and O’Rourke (1992) estimated the ground liquefaction profile, and this is shown
in Figure 10.17. The soil at the site liquefied to a maximum depth of about 10 m below the river bed, and this
depth decreased towards the abutments.

10.7.1.2 Foundation and structural details of the Showa Bridge

The foundation of each supporting pier was a single row of nine tubular steel piles connected laterally by a
pile cap. Each pile was 25 m long with an outer diameter D of 0.609 m. The wall thickness of the upper 12 m
of the pile was 16 mm, and the thickness for the bottom 13 m was 9 mm. The material of the piles, in
accordance with Japanese standards (JSA, 2014), was assumed to be SKK490 grade steel pipe with a yield
strength of 315 MPa and an ultimate strength of 490 MPa.

A dead load of 6662 kN from the girder and the slab was transferred to the pier, which in turn was shared by the nine
piles (Bhattacharya et al., 2014). The axial load carried by each pile was estimated as 740 kN. The design live loads
are ignored, as there was no signiﬁcant trafﬁc on the bridge during its failure. The bridge collapsed just 1 month after
its construction (Figure 10.18), ruling out the possibility of bridge failure due to material degradation.

Figure 10.17 Soil liquefaction proﬁle (shaded area)

North Showa Bridge South

Hakusan Park B′
10
FL FL FL FL
N 0 10 20
0 50 100m 0 10 20
N N 0 10 20
N 0 10 20
0 50 0 50 0 50 0 50
Ts FL FL
Elevotion: m

0 N 0 10 20 N 0 10 20
As-1 0 50 0 50

–10
As-2

–20

Figure 10.18 Collapse of the Showa Bridge along with deﬂections of the pile caps (Iwasaki, 1986)

Left bank P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11

Right bank

Abutment Abutment

322

Seismic Design of Foundations

Analysis of foundations for major bridges

The procedure introduced in Chapter 9 to determine the elongated natural period of piers as result of liq-
uefaction is revisited here in the context of the Showa Bridge. The relevant dynamic design parameters for the
Showa Bridge required for evaluation of the natural period and the peak lateral displacement in the case of
liqueﬁed soil are discussed below.

(a) Unsupported length of the pile. The unsupported lengths of the pile in the pre- and post-liquefaction
stages (L0-pre and L0-post, respectively) are determined with the appropriate depth of ﬁxity, following
Bhattacharya and Goda (2013) and as explained in Chapter 9. Considerations for the unsupported lengths for
both stages for pier P6 are illustrated in Figure 10.19. The unsupported length for the piles is estimated using
the data from the available literature (Bhattacharya et al., 2014) and is presented in Table 10.5. The analysis
also assumes that the pile is stable under vertical settlement.

(b) Natural period. The natural period of a pier can be estimated following the simple formulas given in
Lombardi and Bhattacharya (2013) and from the theory explained in Chapter 9 (see Figure 9.12). For the pre-
liquefaction stage, the natural period Tpre can be estimated using
sffiffiffiffiffiffiffiffiffiffiffi
Me
Tpre = 2π (10:55)
Ke-pre

3EI
Ke-pre = (10:56)
L30-pre

Figure 10.19 Model of Showa Bridge pile P6: (a) pre-liquefaction stage; (b) post-liquefaction stage

Sdpre Sdpost

W W W W
Feq-pre Feq-post

6m Air
6m Air

3m Water 3m Water
5D

Liquefiable Liquefiable
10 m soil 10 m
soil

4D
Non-
6m liquefiable Non-liquefiable
soil soil

(a) (b)

323

Seismic Design of Foundations

For the post-liquefaction stage, the natural period Tpost can be estimated by
sffiffiffiffiffiffiffiffiffiffiffiffiffi
Me
Tpost = 2π (10:57)
Ke-post

3EI
Ke-post = (10:58)
L30-post

Here, Me is the equivalent mass, Ke-pre and Ke-pre are the pre- and post-liquefaction stiffnesses, respectively,
and EI is the ﬂexural stiffness of the pier.

The geometry of each pile foundation, the length of each pile above the ground surface and the depth of the
liqueﬁed soil have been taken from the available literature for estimation of the natural period before and after
liquefaction at the Showa Bridge, and are shown in Figure 10.19 for pier P6. The natural periods of the
different piers were calculated for both extreme conditions, and are presented in Table 10.5. The published
work of Lombardi and Bhattacharya (2013) can be referred to for further details.

Table 10.5 Analysis for the Showa Bridge (1964 Niigata earthquake)

Pier Hair: m Hwater: m Hliq: m L0-pre: m Tpre: s Sd-pre: m L0-post: m Tpost: s Sd-post: m Increase Remarka
in Sd: %

P1 6 0 5 9 1.60 0.95 13.4 2.91 1.17 23.15 NC

P2 6 2.5 5 11.5 2.31 1.2 15.9 3.77 1.76 46.66 NC

P3 6 3 6.5 12 2.47 1.3 17.9 4.50 2.8 115.3 C

P4 6 3 8 12 2.47 1.3 19.4 5.08 3.3 153.8 C

P5 6 3 9 12 2.47 1.3 20.4 5.48 3.78 190.7 C

P6 6 3 10 12 2.47 1.3 21.4 5.88 4.0 207.6 C

P7 6 4 4.5 13 2.78 1.6 16.9 4.13 2.5 56.25 NC

P8 6 4.5 1 13.5 2.95 1.6 13.9 3.08 2.2 37.5 NC

P9 6 5 1 14 3.11 1.79 14.4 3.25 2.4 34.07 NC

P10 6 2 0.5 11 2.17 1.09 11 2.14 1.5 37.61 NC

P11 6 0 0.5 9 1.60 0.63 9.4 1.57 0.675 7.142 NC

Hair, mean height of each pier in air; Hwater, mean height of water column at each pile; Hliq, mean depth of liquefaction
a
C, collapse – the piers collapsed after full liquefaction; NC, no collapse – the piers did not collapse after full liquefaction

324

Seismic Design of Foundations

Analysis of foundations for major bridges

(c) Estimation of the peak acceleration and displacement. The response spectrum of a type I (level 2)
earthquake as prescribed by the Japanese code of practice ‘Design speciﬁcations for highway bridges’ (JRA,
2002) is used for the analysis. The peak acceleration for different piers are estimated from the expressions
proposed in Table 6.4.1 of the Japanese code (JRA, 2002). This ground motion corresponds to an inter-plate
earthquake with less probability of occurrence during the bridge service life but strong enough to cause
critical damage. Tpre and Tpost (see Table 10.5) are determined for different piers, and are used to estimate the
corresponding peak accelerations Apre and Apost. These acceleration values are used to ﬁnd the peak dis-
placements Dpre and Dpost at the head of each of pier using
2
Tpre
Dpre = Apre (10:59)
2π

2
Tpost
Dpost = Apost (10:60)
2π

Both equations can be combined to establish the relationship between the peak displacement of the pier top
before liquefaction and that of at full liquefaction (Figure 10.20).

Figure 10.20 Variation of the peak displacement with increasing natural period

30
Sustained end pier (P1) – the Showa Bridge
Failed central pier (P5) – the Showa Bridge
25 Failed central pier (P6) – the Showa Bridge
Failed central pier – the Rio Viscaya Bridge Sa-post/Sa-pre = 1.0
Saraighat Bridge 0.90

20 0.80

0.70
Sd-post/Sd-pre

15 0.60

0.50

10 0.40
0.35
0.30
0.25
5 0.20
0.15
0.10

0
1 2 3 4 5
Tpost/Tpre

325

Seismic Design of Foundations

The following can be noted from Table 10.5

■ First, it can be seen that the natural period of different piers increases due to the liquefaction. For soil
with a greater depth of liquefiable soil, the margin of increase is even higher. For instance, the natural
period of pier P6 becomes almost 5.8 s at full liquefaction, whereas before the liquefaction it was only
2.47 s. This resulting increased natural period of the piers falls in the displacement sensitive zone of
the response spectra. So, the lateral displacement of the pile top may increase significantly.
■ Second, the peak displacement prior to liquefaction Sd-pre for the different piers ranges from 0.9
to 1.7 m. This value increases as liquefaction sets in. For example, the peak displacement at full
liquefaction Sd-post for pier P6 increases up to 4 m. It can also be seen from Figure 10.20 that as the
natural period of pier P6 at full liquefaction increases to almost 2.3 times its pre-liquefaction value,
its spectral displacement increases by almost 200%. A similar relationship for piers P5 and P1 is also
illustrated in Figure 10.20. As piers have been simplified to be single-degree-of-freedom systems and
the liquefied soil still offers some resistance, the actual displacement can be less than this. Noted that
as the natural period of the pile increases to a higher value because of liquefaction, the equivalent
static force acting at the pile top reduces due to a reduction in the value of the peak acceleration.
Nonetheless, the lateral displacement due to this effect becomes so high that it may have unseated the
deck, to cause the failure.

10.7.2 Collapse of the Panshan Bridge in the 1975 Haicheng earthquake (China)
The Haicheng earthquake occurred on 4 February 1975 in the northeastern region of China with a surface-
wave magnitude (Ms) of 7.3. The hypocentre depth was about 12 km. Many bridges, buildings, embankments
and other structures were seriously damaged due to soil liquefaction – including the Panshan Bridge. The
overall length of this bridge was 315.64 m, and the length of each span was about 22.2 m. The width of the
bridge was about 7 m. The bridge was supported on piles 0.9 m in diameter and 30 m in length, with a pile cap
thickness of 1.7 m. The piers had a diameter of 1 m and a length of 7 m. The literature is silent about the
strength of the soil around the piers as well as the number of piles beneath each support. There is also no
information on SPT N values for the surrounding soil. Hence, it is difﬁcult to predict the superstructure load
acting on the piles for this bridge. However, looking at the bridge superstructure and substructure and noting
its similarity to the Hangu Bridge, which was damaged during the 1976 Tangshan earthquake, the super-
structure load acting from each of the span was assumed to be 170 t, with three columns and piles beneath
each support. Figure 10.8 is a schematic of the undamaged and damaged Panshan Bridge.

The soil near the Panshan Bridge is highly stratified, with clay, fine sand and silty soil. There is a shallow
layer of silty soil of around 1.5 m near the ground surface, which liquefied. However, for pier P7, this silty
layer was absent, as the water depth at this location was greater than for other piers. The soil layer
arrangement with respect to depth from the ground surface around all the piers was assumed to be as
described by Shengcong and Tatsuoka (1984), and used to determine the soil stratigraphy near pier P7.

10.7.2.1 Liquefaction proﬁle

The depth of liquefaction near the bridge was about 3 m (Liu et al., 1991). So it is assumed that the top silty
soil would have liqueﬁed along with the top portion of the clay, which would have suffered from cyclic
mobility.

326

Seismic Design of Foundations

Analysis of foundations for major bridges

However, for piers P6, P7, P8 and P9, the water depth was greater. For pier P7, the water depth was about 8 m,
while for other piers it was around 1 m. Hence, the depth of liquefaction can be taken to be about 11 m (up to
the interface of the silt and ﬁne sand).

10.7.2.2 Estimation of the dynamic design parameters

The unsupported length of the piles in the pre- and post-liquefaction stages can be determined with the
appropriate depth of ﬁxity, following Bhattacharya and Goda (2013). The unsupported length for the piles
was estimated using the data from the available literature (Shengcong and Tatsuoka, 1984), and presented in
Table 10.6. The geometry of each pile foundation, the length of each pile above the ground surface and the
depth of the liqueﬁed soil were taken from the available literature (Shengcong and Tatsuoka 1984) for
estimation of the natural period of the Panshan Bridge piers before and after liquefaction. Using Equations
10.25 and 10.27, the natural period of the different piers has been calculated for both extreme conditions, and
are given in Table 10.6. The response spectrum as prescribed in the GB 500110 standard (Chinese Gov-
ernment, 2010) was used for the analysis. The peak acceleration for different piers was estimated from the

Table 10.6 Analysis for the Panshan Bridge (1975 Haicheng earthquake)

Pier Hair: Hwater: Hliq: L0-pre: L0-post: Tpre: Tpost: Tpost/ Dpre: Dposte: Increase Remarka
m m m m m s s Tpre m m in D: %

P1 7 1 2 12.5 13.6 2.2 2.5 1.134 0.127 0.146 15 NC

P2 6 2 2 12.5 13.6 2.2 2.5 1.134 0.127 0.146 15 NC

P3 6 2 3 12.5 14.6 2.2 2.8 1.262 0.127 0.164 29 NC

P4 5 2 3 11.5 13.6 1.9 2.5 1.286 0.111 0.146 32 NC

P5 5 2 3 11.5 13.6 1.9 2.5 1.286 0.111 0.146 32 NC

P6 4.11 2.89 3 11.5 13.6 1.9 2.5 1.286 0.111 0.122 10 NC

P7 7 8 11 19.5 29.6 4.3 8 1.9 0.265 0.709 167 C

P8 4.5 3.31 3 12.3 14.4 2.1 2.7 1.266 0.124 0.143 15 NC

P9 4.73 2.27 3 11.5 13.6 1.9 2.5 1.286 0.111 0.122 10 NC

P10 5.25 1.75 3 11.5 13.6 1.9 2.5 1.286 0.111 0.146 32 NC

P11 5.35 1.65 2 11.5 12.6 1.9 2.2 1.146 0.111 0.129 16 C

P12 5.87 1.13 2 11.5 12.6 1.9 2.2 1.146 0.111 0.129 16 NC

P13 6 0.31 1 10.8 10.9 1.8 1.8 1.013 0.100 0.101 1 NC

a
C, collapse – the piers collapsed after full liquefaction; NC, no collapse – the piers did not collapse after full liquefaction

327

Seismic Design of Foundations

expressions proposed in Figure 5.1.5 of GB 50011. Tpre and Tpost were determined for different piers, and
used to estimate the peak accelerations. These acceleration values were used to obtain the peak displacements
at the head of each of pier, and are listed in Table 10.6. The analysis was carried out taking into account the
support system of the bridge, the earthquake conditions and the depth of liquefaction.

The following can be noted from Table 10.6

■ First, it can be seen that the natural period of different piers increases due to liquefaction: for example,
the natural period of pier P7 becomes almost 8 s at full liquefaction, whereas before liquefaction it
was only 4 s. This resulting increased natural period of the piers falls in the displacement sensitive
zone of the response spectra. So, the lateral displacement of the pile top may increase signiﬁcantly.
■ Second, the peak displacement prior to liquefaction Dpre for the different piers ranges from 0.1 to
0.2 m. It can be seen that this displacement increases for every pier after liquefaction. However,
the increase in the peak displacement for pier P7 is much greater than for the other piers.

10.8. Performance comparison of bridges supported on pile and caisson

foundations in liquefiable soils
It is of interest to note the difference in behaviour between a bridge supported on a pile foundation and a
caisson (well) foundation as soil progressively liquefies. In this context, the well-documented pile-supported
Showa Bridge and the caisson-supported Saraighat Bridge are considered. It has been reported by
Bhattacharya et al. (2014a) that the period of the Showa Bridge increased from 2 to 6 s as the soil liquefied to
a depth of about 10 m, resulting in increased spectral displacement. Figure 10.21 compares the change in the
predicted time periods of both bridges. A steep increase in the time period for the Showa Bridge can be
observed. Figure 10.22 presents a comparison of the spectral displacement for both bridges, and the marked
difference in the predicted behaviours can be attributed to the differential increase in the time period (see
Figure 10.21). For the same depth of liquefaction of 10 m, the increase in the time period is 300% for the
Showa Bridge; however, the corresponding increase for the Saraighat Bridge is about 18%. It is important to
note that the Showa Bridge collapsed due to the decks falling. Based on the study, it can be inferred that the
predicted good performance of the Saraighat Bridge can be attributed to the high rigidity of the caisson

Figure 10.21 (a) Change in time period and (b) percentage change in time period for Saraighat and Showa Bridges

150
6 Saraighat
Change in time period: %

125 Showa
5
ge

100
Time period: s

rid
db

4
75
rte
po
up

3 50
e-s
Pil

2 ge
25 ed brid
-s upport
Caisson
1
0 5 10 15 0 5 10 15
Depth of liquid: m Depth of liquid: m
(a) (b)

328

Seismic Design of Foundations

Analysis of foundations for major bridges

Figure 10.22 (a) Change in Sd and (b) percentage change in Sd for both the bridges

5000 200
Saraighat
Spectral displacement: mm

Showa
4000 160

Change in Sd: %

eg
id
3000 120

br
d
rte
p po
2000 80

su-
le
Pi
1000 40
ge
ported brid
Caisson-sup
0 0
0 3 6 9 12 15 0 3 6 9 12 15
Depth of liquid: m Depth of liquid: m
(a) (b)

foundations, which need a relatively low stiffness contribution from the surrounding soil. Many major bridges
in high-seismicity areas of the world are supported on caisson foundations. These result show that the change
in the natural period due to liquefaction is insigniﬁcant for caisson-supported bridges, and will result in a very
low spectral displacement of the pier head. Though they consume enormous amounts of concrete, caisson
foundations may be the preferable foundation type for major bridges in liqueﬁable soils.

10.9. Summary
Bridges are backbone of an economy and they need to be working even after an earthquake. This chapter
provides examples of collapses and failures of bridges in the past earthquakes where ground was one of the
main contributor. Mechanisms of failures are discussed and calculation procedure are shown. One of the
important mechanism is the elongation of bridges period due to subsurface liquefaction. The relative per-
formance of pile supported bridge and caisson supported bridge are also discussed by taking an example.

REFERENCES
Arany L, Bhattacharya S, Macdonald JHG and Hogan SJ (2016) Closed form solution of Eigen frequency of
monopile supported offshore wind turbines in deeper waters incorporating stiffness of substructure and SSI.
Soil Dynamics and Earthquake Engineering 83: 18–32.
Bhattacharya S and Goda K (2013) Probabilistic buckling analysis of axially loaded piles in liqueﬁable soils.
Soil Dynamics and Earthquake Engineering 45: 13–24.
Bhattacharya S, Tokimatsu K, Goda K, Sarkar R, Shadlou M and Rouholamin M (2014) Collapse of Showa Bridge
during 1964 Niigata earthquake: a quantitative reappraisal on the failure mechanisms. Soil Dynamics and
Earthquake Engineering 65: 55–71.
CEN (Comité Européen de Normalisation) (2004) EN 1998-5. Design of structures for earthquake resistance.
Part 5: Foundations, retaining structures and geotechnical aspects. CEN, Brussels, Belgium.
Chen WF and Duan L (2003) Bridge Engineering: Seismic Design. CRC Press, Boca. Raton, FL, USA.
Chinese Government (2010) GB 50011. Code for seismic design of buildings. China Architecture and Building
Press, Beijing, China.
Dammala PK, Bhattacharya S, Krishna AM, Kumar SS and Dasgupta K (2017a) Scenario based seismic re-
qualiﬁcation of caisson supported major bridges – a case study of Saraighat Bridge. Soil Dynamics and
Earthquake Engineering 100: 270–275.

329

Seismic Design of Foundations

Dammala PK, Krishna AM, Bhattacharya S, Nikitas G and Rouholamin M (2017b) Dynamic soil properties
for seismic ground response studies in Northeastern India. Soil Dynamics and Earthquake Engineering
100: 357–370.
Dammala PK, Jalbi S, Bhattacharya S and Adapa MK (2018) Simpliﬁed methodology for stiffness estimation of
double D shaped caisson foundations. In New Solutions for Challenges in Applications of New Materials and
Geotechnical Issues. GeoChina 2018. Sustainable Civil Infrastructure (Wang S, Xinbao Y and Tefe M (eds)).
Springer, Berlin, Germany.
Dammala PK, Jalbi S, Bhattacharya S, Adapa MK and Bouzid DA (2019) Impedance functions for double-D
shaped caisson foundations. Journal of Testing and Evaluation, doi: 10.1520/JTE20180075.
Debnath N, Deb SK and Dutta A (2015) Multi-modal vibration control of truss bridges with tuned mass dampers
under general loading. Journal of Vibration and Control 22(20): 4120–4140.
Fukuoka M (1966) Damage to civil engineering structures. Soils and Foundations 6(2): 45–52.
Gerolymos N and Gazetas G (2006) Winkler model for lateral response of rigid caisson foundations in linear soil.
Soil Dynamics and Earthquake Engineering 26(5): 347–361.
Hamada M and O’Rourke TD (1992) Case Studies of Liquefaction and Lifeline Performance During Past
Earthquakes, vol. 1. Japanese Case Studies. National Center for Earthquake Engineering Research, Buffalo,
NY, USA, Technical Report NCEER-92-0001.
Han Q, Du X, Liu J et al. (2009) Seismic damage of highway bridges during the 2008 Wenchuan earthquake.
Earthquake Engineering and Engineering Vibration 8(2): 263–273.
Huixian L, Housner GW, Lili X and Duxin H (2002) The Great Tangshan Earthquake of 1976. California Institute
of Technology, Pasadena, CA, USA, EERL.2002.001.
Idriss IM and Boulanger RW (2006) Semi-empirical procedures for evaluating liquefaction potential during
earthquakes. Soil Dynamics and Earthquake Engineering 26(2–4): 115–130 (special issue).
Ishihara K, Alex AA and Towhata I (1993) Liquefaction induced ground damage in Dagupan in the July 16, 1990
Luzon Earthquake. Soils and Foundations 33(1): 133–154.
Iwasaki T (1986) Soil liquefaction studies in Japan: state of the art. Soil Dynamics and Earthquake Engineering
5(1): 2–68.
JRA (Japan Road Association) (2002) Speciﬁcations for highway bridges. Part V: seismic design. JRA, Tokyo,
Japan. (In Japanese.)
JSA (Japanese Standards Association) (2004) JIS A 5525. Steel pipe piles. JSA, Tokyo, Japan.
Kawashima K, Takahashi Y, Ge H, Wu Z and Zhang J (2009) Reconnaissance report on damage of bridges in 2008
Wenchuan China earthquake. Journal of Earthquake Engineering 13(7): 965–996.
Latini C and Zania V (2017) Dynamic lateral response of suction caissons. Soil Dynamics and Earthquake
Engineering 100: 59–71.
Liu H, Wang C and Wong L (1991) Slides of river banks concerning lateral spread of liquefaction. Proccedings of
the 2nd International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil
Dynamics, Saint Louis, MO, USA, pp. 507–514.
Lombardi D and Bhattacharya S (2013) Modal analysis of pile-supported structures during seismic liquefaction.
Earthquake Engineering and Structural Dynamics 43(1): 119–138.
Mohanty P, Dutta SC and Bhattacharya S (2017) Proposed mechanism for mid-span failure of pile supported river
bridges during seismic liquefaction. Soil Dynamics and Earthquake Engineering 102: 41–45.
Shadlou M and Bhattacharya S (2016) Dynamic stiffness of monopiles supporting offshore wind turbine
generators. Soil Dynamics and Earthquake Engineering 88: 15–32.
Shengcong F and Tatsuoka F (1984) Soil liquefaction during Haicheng and Tangshan earthquake in China:
a review. Soils and Foundations 24(4): 11–29.

330

Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

Chapter 11
Foundations in slopes and for retaining
walls
11.1. Introduction
The natural topography of many seismically active regions, such as Japan or New Zealand, can be described
as either hilly or rolling terrain. Therefore, often in engineering projects, such as those involving bridge
abutments, building foundations and road pavements, foundations requiring to be constructed near sloping
ground are encountered. Moreover, earth-retaining structures, such as retaining walls, bridge abutments, quay
walls, anchored bulkheads, braced excavations and mechanically stabilised walls, are used throughout such
seismically active areas.

When designing retaining structures and foundations in slopes to incorporate the effects of seismic loading,
settlement/deformation criteria typically govern over bearing capacity criteria. While ﬁnite-element methods
may be the preferred approach in analysing important structures, limit equilibrium methods have been
commonly used due to their simplicity; however, they are limited by the need to assume a failure mechanism.

In this chapter the seismic performance of foundations in slopes and of retaining structures is discussed in
terms of the observed response during previous earthquakes, and the procedures involved in analysing their
seismic behaviour through a pseudostatic method are introduced.

11.2. Concept of seismic coefﬁcients for slopes and retaining walls

11.2.1 Simpliﬁed pseudo-static approach
When analysing the seismic stability of slopes and retaining structures, pseudostatic methods have been the
conventional approach. In this method, the stability of a potentially sliding mass is determined for static
loading conditions, and the effects of earthquake forces are accounted for by including equivalent horizontal
forces acting on the mass. The horizontal force is expressed as the product of the weight of the sliding mass
and a seismic coefﬁcient that is expressed as a fraction of the acceleration due to gravity

S = kW

where S is the pseudostatic earthquake force, k is the seismic coefﬁcient and W is the weight of the
potentially unstable mass. Using this approach, the slope stability can be calculated by the usual limit
equilibrium static methods; it is merely necessary to introduce the equivalent static force S to the other forces
in play (Figure 11.1).

331

Seismic Design of Foundations

Figure 11.1 Slice method in pseudostatic analysis

H
S
W

τm

Si Ei
Wi h
Ei+1

Xi+1

τm

Pseudostatic methods are popular because

■ they do not involve signiﬁcant computing costs
■ much has been written about them, and reﬁnements of varying sophistication have been suggested,
so there are many possibilities for comparison
■ for soil characterisation, they use parameters readily obtained from static laboratory tests that are
widely known and easy to perform.

On the other hand, they have the following shortcomings

■ they cannot take into account the complex nature of seismic phenomena that involve cyclic actions
alternating in time and space, changing in intensity and sign from moment to moment and from point to point
■ they do not adequately characterise the behaviour of soils, ascribing static resistance parameters to
these, while, in fact, the geotechnical characteristics of a site subjected to alternating seismic stresses
vary in time (depending on the degree of strain) and may be reduced drastically
■ all limit equilibrium methods provide only one statement of the degree of over-all stability along a
given slip surface, while offering no indication of the stress and strain states at various points within
the medium concerned.

332

Seismic Design of Foundations

Foundations in slopes and for retaining walls

Nevertheless, in spite of these shortcomings, regulatory codes and published design guides for the seismic
design of slopes and retaining walls are typically based on the simple limit equilibrium design philosophy
where the seismic load is typically represented as an additional pseudostatic load.

11.2.2 Seismic coefﬁcient

It has been conventionally assumed, albeit wrongly, that the seismic coefﬁcient k and the magnitude of
acceleration amax during earthquakes are equivalent to each other. However, a surface acceleration with the
maximum value of amax = 0.7g is not as destructive as taking k = 0.7. This is because the duration of
acceleration is limited and may not give structures time to move signiﬁcantly, while the static seismic inertia
force with k = 0.7 is hypothesised to last forever.

A compilation of recommended values for k is summarised in Table 11.1. Note from the table that the
recommended maximum value of k is about 0.33–0.50 of the peak ground acceleration (PGA), with 0.5 as the
maximum (for catastrophic earthquakes). In Japan, Noda et al. (1975) analysed past earthquake damage to
quay walls at harbours, and they estimated the k values needed to induce failure using limit equilibrium. They
also estimated the maximum horizontal acceleration amax by using acceleration records obtained at nearby
sites or by mathematical modelling. The results of their studies are shown in Figure 11.2.

Currently, there are no specific rules for the selection of an appropriate seismic coefficient for design. It is
recommended that the seismic coefficient should be based on the anticipated level of acceleration within the
failure mass and should also depend on the importance of the structure and an acceptable level of seismic
performance (i.e. amount of allowable seismic displacement).

When the horizontal seismic coefﬁcient kh is set equal to the PGA, the seismic active loadings can become
very large, resulting in the design of the retaining structure becoming increasingly large and uneconomical.

Table 11.1 Recommended values for the horizontal seismic coefﬁcient k from various sources

k Description Reference

0.1 ‘Severe’ earthquakes (Rossi-Forel IX) Terzaghi (1950)

0.2 ‘Violent, destructive’ earthquakes (X)

0.5 ‘Catastrophic’ earthquakes

0.1–0.2 FS > 1.15 Seed (1979)

0.10 ‘Major’ earthquake, FS > 1.0 US Army Corps of Engineers (1982)

0.15 ‘Great’ earthquake, FS > 1.0

⅓–½ PHA FS >1.0 Marcuson and Franklin (1983)

½ PHA FS >1.0 Hynes-Grifﬁn and Franklin (1984)

FS, factor of safety; PHA, peak horizontal acceleration

333

Seismic Design of Foundations

Figure 11.2 Relation between the seismic coefﬁcient k and the maximum ground acceleration amax (Noda et al.,
1975). Note that 1 gal = 1 cm/s2

1/3
(
0.3
a
k = max
g
k=
3 (
1 amax
g

*
(1935)
Seismic coefficient, k

(1952) (1952)
0.2 (1952)

(1964) (1968)

(1973)

0.1 * (1930)

* Measured acceleration

0 100 200 300 400 500

Maximum ground acceleration, amax: gal

Designing for kh = PGA will limit the deformations to zero; however, if deformations can be tolerated within
the performance limits of the structure (e.g. within the allowable seismic displacement), then kh can be
reduced to some fraction of the anticipated peak acceleration.

Bray and Travasarou (2009) proposed a rational basis for selecting the seismic coefficient based on a project-
specific allowable level of seismic displacement. The seismic response characteristics of the slope are rep-
resented by the fundamental period of the potential sliding mass, and the site-dependent seismic demand is
characterised by the 5% damped elastic design spectral acceleration at the degraded period of the potential
sliding mass. Bray et al. (2010) extended the concept to the seismic design of retaining walls, where the
seismic coefficient is selected based on proper consideration of the seismic hazard at the site and the amount
of seismic displacement that defines the threshold between satisfactory and unsatisfactory seismic perfor-
mance of the retaining structure.

11.3. Foundation considerations for buildings on slopes

Due to the lack of suitable ﬂat land, as well as other factors related to scenery and other aesthetic issues, many
structures such as residential houses are now being constructed on hillsides and sloping ground. Constructing
these buildings on sloped sites in seismically active areas is more challenging, because it not only requires
additional site preparation and foundation work but also consideration of the stability of the ground sup-
porting the foundation.

334

Seismic Design of Foundations

Foundations in slopes and for retaining walls

When building on sloped ground, any of the following options are typically adopted

■ Cut and fill. In this process, a level base is carved out on the sloping site so that the structure can be
designed for use on a level site. Any spoil that is removed from the slope is used to fill the lower edge
of the slope (Figure 11.3(a)).
■ Stilt support. Building residential houses and light structures on stilts is one way of addressing
steeply sloped sites (see Figure 11.3(b)). This avoids the need for expensive foundations and also
leaves the ground untouched.
■ Basements and retaining walls. Building a basement on the lower edge of a slope is an option for
bringing the slope up to the level of the foundation slab (Figure 11.3(c)). The basement walls are
typically subjected to significant earth pressure from the surrounding ground, and are designed as
retaining walls. In some cases, a retaining wall is used to prop up the sloping ground so that the
building can be constructed adjacent to it.

Whatever option is selected, the design of shallow foundations on slopes is governed by the two criteria
mentioned in Chapter 8: bearing capacity criteria and deformation (settlement) criteria. The bearing capacity
on sloping grounds has often been assessed using limit equilibrium methods (e.g. Castelli and Motta, 2010;
Hansen, 1970), upper-bound methods (e.g. Kumar and Rao, 2003) and ﬁnite-element analysis (e.g.
Georgiadis, 2010; Loukidis and Salgado, 2009). Of these approaches, limit equilibrium methods have been
commonly used due to their simplicity; however, they are limited by the need to assume a failure mechanism.
As indicated in Figure 11.4, the proximity of the foundation to the inclined ground surface affects the for-
mation of the failure wedges at ultimate load.

Figure 11.3 Typical options when building on sloping ground: (a) cut and ﬁll; (b) use of stilts; (c) basement and
retaining walls

Cut
Original slope
New level surface Retaining wall

Retaining wall
Fill Stilts

(a) (b)

Basement

(c)

335

Seismic Design of Foundations

Figure 11.4 Bearing capacity of strip footing on a slope

Following many large-scale earthquakes, buildings constructed on or adjacent to slopes have been affected
not because the ultimate bearing capacity of the ground was reached but because of instability of the ground
where footings were built, as well as slope collapse occurring upslope impacting the structure. Although the
main problems with shallow foundation performance during earthquakes have been related to poor soil
conditions, which induce liquefaction, cyclic softening, lateral spreading and other instability, many instances
of sloping sites not susceptible to soil liquefaction failed due to poor site performance, where slope instability
(either upslope or underneath the building foundation) led to foundation failure; some case studies are dis-
cussed in the next section. Thus, when constructing structures on slopes, a full site assessment should be
carried out to decide whether the site will be suitable for shallow foundations, or whether ground
improvement is ﬁrst required, or whether deep foundations would be more appropriate.

11.4. Case studies: seismic performance of buildings on slopes

11.4.1 2011 Tōhoku earthquake (Japan)
One of the devastating impacts of the 2011 Tōhoku earthquake (moment magnitude (Mw) 9.0) off the Pacific
coast was the severe damage to houses, especially those built on hillside embankments and artificial valley
fills on sloping residential land surrounding Sendai City. Reports indicate that as many as 950 slope failures in
nine prefectures occurred in such valley fills, with more than 4000 housing lots involving 7000 houses
severely damaged (Hyodo et al., 2012; Mori et al., 2012). Some of these affected areas have been damaged
previously following the 1978 Miyagi earthquake and then subsequently rebuilt.

Based on observations from previous earthquakes and the 2011 earthquake, Mori et al. (2012) classiﬁed the
types of failures in hillside embankments for residential use as

■ class 1 – landslides in the soft original ground

■ class 2 – landslide-type failure along the boundary between the ﬁll and the original ground
■ class 3 – ground displacements due to shallow slips
■ class 4 – ground displacement without slippage due to seismic motion.

336

Seismic Design of Foundations

Foundations in slopes and for retaining walls

Figure 11.5 Classiﬁcation of the failure type in hillside embankments for residential use (Mori et al., 2012; reprinted
with permission from the author)

Slip surface Slip surface Slip surface

Fill Fill
Original ground Original ground

Class 1: landslide-type failure in the soft original ground Class 2: landslide-type failure along the boundary between the
fill and the original ground

Slip surface

Fill Fill
Original ground Original ground

Class 3: ground displacement due to shallow slip Class 4: ground displacement without slip

These are illustrated in Figure 11.5. Class 1 is a typical slope failure through the soft original ground, and is
characterised by the familiar main scrap at the top of the slope and upheaval at the toe. Class 2 is similar to
class 1, but the failure surface is located along the boundary of the ﬁll and the original ground. Classes 3 and 4
generally involve weak ﬁll materials and a shallow water table, and the apparent modes of ground dis-
placement at the surface include cracks, differential settlements, steps in the ground and slope failures. Houses
and other structures located within the moving mass or along the failure surface are expected to undergo
various types of deformation.

Following the 2011 earthquake, post-disaster investigations of the damage and failure of hillside embank-
ments showed mostly classes 2 and 3 damage: that is, valley-filled embankments, which were created by
filling valleys with soil cut from both sides of the original ground, and the boundary area between cut and fill
were susceptible to severe housing damage due to the failure of embankments.

Hyodo et al. (2012) reported five slope failures that occurred in a residential area on artificial valley fills in
Taiyo New Town, Yamamoto, Miyagi Prefecture. The site was constructed by levelling the hilly area and
using the cut materials as fills for the valleys to provide foundation ground for houses. The fill material was
sandy, derived from the weathering of tuffaceous sandstone that formed the natural ground. Each of the five
slope failures was observed to occur either at the edge of the artificial valley fill or at embankment sections
widened for road construction.

Figure 11.6 shows one of the failed blocks on the western section of the site. Hyodo et al. (2012) reported
that a two-storey house on top of the slope was damaged due to vertical offsets and extensional failure
resulting from slope movement. The house was tilted, exposing its foundation. Fences, railings and other
light-weight structures also collapsed. The road on the crest of the failed slope subsided, and debris
covered the drainage channels. Portions of road pavements as well as damaged sewage facilities were seen
among the flowed debris. The soil debris was observed to be sand and sandy soils, representing the
material of the fill. Figure 11.7 shows the cross-sectional profile of the slope.

337

Seismic Design of Foundations

Figure 11.6 Slope failure at one of the blocks. A house hangs precariously on the edge of the failed slope
(Hyodo et al., 2012; reprinted with permission from the author)

Figure 11.7 Cross-sectional proﬁle of the failed slope (Hyodo et al., 2012; reprinted with permission from the
author)

Fill
A
B5-2

1 1
1
2
3 B5-2
3
7 Failed fill
4
9
5 0
7
6 1
9 0/45
7 2
14 2

Tuffaceous sandstone 8
15
3

4
1/50

(weathered)
9

10
12
Tuffaceous 5
8

8
13
11
15 sandstone 6
8
12 7
15 13
8
13
9
12
0 5 10 15 20 30 m 10
14
11
13
12 14

338

Seismic Design of Foundations

Foundations in slopes and for retaining walls

Several valley-ﬁll-type landslides were also observed in Oritate 5-chome in Sendai City. The housing
complex in the area was developed in the latter half of the 1960s, and the lots were sold in the early 1970s.
Figure 11.8 shows a plan view of Oritate 5-chome residential area, indicating the distribution of ground
cracks and landslides. It was clear that the landslides occurred along the valley that existed prior to the
construction of the ﬁll. Some of the observed damage to houses in the area is illustrated in Figure 11.9.

Figure 11.8 Plan view of the Oritate 5-chome residential area, showing the location of the landslide zone and the
distribution of cracks (Courtesy of Fukken Gijutsu Consultants)

Tension

Cracks Compression

Tension with strike

slip

Landsliding area

Direction of
movement

100 m

Figure 11.9 Observed damage to residential houses at the Oritate 5-chome landslide site (Courtesy of M. Hyodo)

339

Seismic Design of Foundations

From the post-earthquake reconnaissance reports, it was clear that most of the damage to the housing stock
could be attributed to the ground displacement of the soft filled soils used for the hillside embankment. The
fill materials used were generally composed of soils containing a high ratio of fines, which could have resulted
in poor compaction during construction work, the shallow water table, and the degradation of the hillside
embankments with elapsed time.

Thus, in addressing the stability of house foundations on slopes, it is important to consider not only the
ground deformations that will occur in the vicinity of the footing (i.e. when a footing straddles the boundary
of ﬁll and the original ground) but also the overall slope stability of the hillside embankment itself.
Appropriate countermeasures (ground improvement, restraining piles, etc.) should be implemented to min-
imise the expected ground displacements in the slopes during a strong earthquake. House owners should
therefore understand the need to create earthquake-resistant residential land.

11.4.2 2011 Christchurch earthquake (New Zealand)

The 22 February 2011 earthquake (Mw 6.3), with an epicentre located just south of Christchurch, New Zealand,
caused widespread devastation, including building collapse, liquefaction and landslides. Throughout the Port
Hills of Banks Peninsula on the southern fringes of Christchurch, landslide and ground damage caused by
the earthquake included rock fall (both cliff collapse and boulder roll), incipient loess landslides, and
retaining wall and ﬁll failures. Four deaths from rock fall occurred during the main shock, and one during an
aftershock later in the afternoon. Hundreds of houses were damaged by rock falls and landslide-induced
ground cracking (Dellow et al., 2011).

One of the distinct landslide or ground failure types that have caused damage to residential houses built on
slopes was the catastrophic collapse of over-steepened present-day and former sea cliffs, as shown in
Figure 11.10. Houses situated on top of the hills were damaged by tension cracks on the slopes above the
cliff faces, while those located at the toe of the slopes were destroyed by debris inundation. In addition,

Figure 11.10 Cliff collapse between Sumner and Moncks Bay following the 2011 Christchurch earthquake (Dellow
et al., 2011)

340

Seismic Design of Foundations

Foundations in slopes and for retaining walls

Figure 11.11 Tension cracks in a hilly residential area near Moncks Bay, which affected houses and roads

incipient movement of landslides in loess, ranging from a few millimetres up to 0.4 m, occurred at several
locations, as depicted in Figure 11.11. Again, houses and roads were damaged by extension ﬁssuring at the
head of these features and compressional movement at the toe.

11.5. Earth pressure theory

Okabe (1926) and Mononobe and Matsuo (1929) developed the basis of the pseudostatic analysis of seismic
earth pressures on retaining structures that has become popularly known as the Mononobe–Okabe (M-O)
method. The M-O theory of seismic earth pressure is based on modifying the Coulomb theory on limit
equilibrium by applying a pseudostatic acceleration to the Coulomb active (or passive) wedge.

11.5.1 Active earth pressure condition

Figure 11.12 illustrates the situation that served as the basis for the development of the M-O theory. As it is an
extension of the Coulomb theory for the static case, the same set of assumptions apply, that is

■ backﬁll soil is composed of dry, cohesionless material

■ the slip plane passes through the bottom
■ the failure wedge is a rigid body (acceleration is uniform within the wedge)
■ displacement of the wall is large enough to attain the active and passive states of stress in the backﬁll.

In addition to the forces that exist under static conditions, the dry soil wedge at the back of a rigid wall is
subjected to inertial (or pseudostatic) forces in both the horizontal and vertical directions. For active pressure,
the horizontal inertia force is oriented outwards. The rear face of the wall is inclined by b, with a frictional

341

Seismic Design of Foundations

Figure 11.12 Seismic active earth pressure

KvW

KhW
W
H

δ
φ
β F
PAE

αAE

angle d with the soil. The total active thrust can be expressed in a form similar to that developed for static
conditions
1
PAE = K g H 2 (1 − kv ) (11:1)
2 AE

cos2 (f − q − b)
KAE = rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2 (11:2a)
sin(f + d ) sin(f − q − i)
cos q cos2 b cos (d + b + q) 1 +
cos(d + b + q) cos(i − b)

where

−1 kh
f−i≥q q = tan (11:2b)
1 − kv

PAE is the total seismic active thrust, KAE is coefficient of dynamic active pressure, g is the unit weight of the
soil, H is the wall height, f is the angle of internal friction of the soil, d is the angle of interface friction
between the wall and the soil, b is the inclination of the wall, i is the slope of the backfill soil, and kh and kv are
the horizontal and vertical seismic coefficients, respectively.

The M-O total seismic active thrust acts at the elevation of H/3 above the base of the wall. Recent studies,
however, recommend values of 0.45 to 2/3 of the wall height. Seed and Whitman (1970) divided the total
active thrust PAE into the static component PA and the dynamic component DPAE

PAE = PA + DPAE (11:3)

342

Seismic Design of Foundations

Foundations in slopes and for retaining walls

The static component PA, and the earth pressure coefﬁcient KA are given by

1
PA = K g H2 (11:4)
2 A

cos2 (f − b)
KA = rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 (11:5)
sin(f + d ) sin(f − i)
cos b cos(d + b) 1 +
2
cos(d + b) cos(i − b)

PA is known to act at H/3 above the base of the wall. It was suggested that the dynamic component is taken to act
at approximately 0.6H from the base. Thus, the point of application Y measured above the base of the wall is

PA ðH/ 3Þ + DPAE ð0:6H Þ

Y= (11:6)
PAE

M-O analysis shows that kv, when taken as one-third to one-half the value of kh, affects PAE by less than 10%.
Seed and Whitman (1970) concluded that vertical accelerations can be ignored when the M-O method is used
to estimate PAE for typical wall designs.

11.5.2 Passive earth pressure conditions

The total passive thrust on a wall retaining a dry, cohesionless backﬁll (Figure 11.13) is given by

1
PPE = K g H 2 (1 − kv ) (11:7)
2 PE
cos2 (f − q + b)
KPE = rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2 (11:8)
sin(f − d ) sin(f − q + i)
cosq cos2 b cos (d − b + q) 1 −
cos(d − b + q) cos(i − b)

Figure 11.13 Seismic passive earth pressure

KvW
KhW
PPE
W
H δ

F
β φ

343

Seismic Design of Foundations

The total passive thrust can also be divided into static and dynamic components

PPE = PP − DPPE (11:9)

Note that the dynamic component acts in the opposite direction to the static component, thus reducing the
available passive resistance. Recall that

1
PP = K g H2 (11:10)
2 P

cos2 (f + b)
KP = rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2 (11:11)
sin(f − d ) sin(f + i)
cos b cos (d − b) 1 −
2
cos(d − b) cos(i − b)

11.5.3 Advantages and disadvantages

Although conceptually quite simple, M-O analysis provides a useful means of estimating earthquake-induced
loads on retaining walls. A positive horizontal acceleration coefﬁcient causes the total active thrust to exceed
the static active thrust, and the total passive thrust to be less than the static passive thrust. Since the stability of
a particular wall is generally reduced by an increase in the active thrust and/or a decrease in the passive thrust,
the M-O method produces seismic loads that are more critical than the static loads prior to an earthquake.

As a pseudostatic extension of Coulomb analysis, however, M-O analysis is subject to all the limitations of
pseudostatic analyses as well as the limitations of Coulomb theory (the Coulomb failure wedge assumption
and a cohesionless backfill). For high accelerations or for steep back slopes, the M-O theory leads to
excessively high pressures that asymptote to infinity at critical acceleration levels or back slope angles. For
the latter conditions, no real solutions to the equation exist, implying equilibrium is not possible (see the first
part of Equation 11.2). Typically, the use of the M-O theory should be limited to the following criteria,
provided that the limiting condition of Equation 11.2 is met

■ backﬁll slopes, i ≤ 18.4° (3H:1V or ﬂatter)

■ limited to KAE ≤ 0.60
■ free-draining backﬁll materials (cohesionless soils) behind the wall should extend throughout the
seismic active wedge.

In addition, as in the case of pseudostatic slope stability analyses, determination of the appropriate
pseudostatic coefficient is difficult, and analysis is not appropriate for soils that experience significant loss of
strength during earthquakes (e.g. liquefiable soils). M-O analysis will over-predict the actual total passive
thrust when d > f/2. For these reasons, the M-O method should be used and interpreted carefully.

11.5.4 Modiﬁcations to the original M-O theory

Because of the various limitations associated with the M-O theory, various approaches have been formulated
to address them. Saran and Prakash (1968) extended the expression to cohesive backﬁlls (c–f backﬁlls),
taking into account only the horizontal earthquake acceleration (no vertical acceleration). However, the
analytical approach resulted in the pressure due to the weight of the soil wedge and cohesion being maximised
separately, leading to different failure planes – contrary to what is observed in reality. Saran and Gupta (2003)

344

Seismic Design of Foundations

Foundations in slopes and for retaining walls

later extended the analysis for retaining walls with an inclined c–f backfill. Richards and Shi (1994) also
developed an expression for active thrust applied by cohesive backfills by using a generalisation of the
elastoplastic solution for the free field with uniform acceleration for granular materials (Richards et al., 1990),
and considering both horizontal and vertical earthquake coefficients. Shukla et al. (2009) developed an
expression for the total active thrust from c–f backfills on the retaining wall based on the Coulomb sliding
wedge concept, taking into account both horizontal and vertical earthquake accelerations. Some of the more
common alternatives to the M-O theory are described below.

11.5.4.1 Trial wedge method

When the M-O theory is not appropriate for analysing the case on hand, the trial wedge method can be used to
estimate the critical seismic-induced active forces on retaining walls. The method can accommodate various
types of soil behind the wall and relatively complex surface geometries, and is therefore more adaptable.

The method is illustrated in Figure 11.14. First, the backﬁll is divided into wedges by selecting planes through
the heel of the wall. The forces acting on each of these wedges are combined in a force polygon so that the
magnitude of the resultant earth pressure can be obtained. A force polygon is constructed, although the forces
acting on the wedge are in general not in moment equilibrium; this method is therefore an approximation,
with the same assumptions as the equations for Coulomb’s conditions. For the seismic active pressure case,
the maximum value of PAE is required, and the limiting value is obtained by interpolating between the values
for the wedges selected. Details on conducting the trial wedge method of analysis can be found in Ebeling
et al. (2007) and Bowles (1982).

Figure 11.14 Trial wedge method (Anderson et al., 2008; reproduced with permission from the publisher, National
Academies Press, Washington, D.C.)

2 3 4 5
1

khW

W
PAE cℓ
5
ℓ
R
φ
Forces acting on 4
trial wedges for
earthquake loading Maximum PAE
PAE 3
Force polygon for
a typical wedge
2
R W
1
khW Combination
of force polygons
to obtain maximum PAE
cℓ

345

Seismic Design of Foundations

11.5.4.2 Seismic coefﬁcient charts

Anderson et al. (2008) made use of the trial wedge method for a generalised c–f soil failure criterion. The
active earth pressure PAE was calculated to satisfy the condition of moment equilibrium of each of the
combinations of the assumed trial wedge and soil shear strength values over the failure surface. The con-
figurations of the trial wedges were varied until the relative maximum PAE value was obtained for various
horizontal seismic coefficients kh. Figure 11.15 presents active earth pressure coefficient charts for two dif-
ferent soil friction angles with different values of cohesion for horizontal backfill, assuming no tension cracks
and zero wall cohesion. For each chart, earth pressure coefficients are presented as a function of the seismic
coefficient kh and various values of cohesion c. The c value was normalised by the product gH, where g is the
unit weight of the soil and H is the wall height in the presented design charts.

Figure 11.15 Seismic coefﬁcient charts for c–ϕ soils for (a) ϕ = 35° and (b) ϕ = 40° (Anderson et al., 2008;
reproduced with permission from the publisher, National Academies Press, Washington, D.C.)

2.0

1.5
C/γ H = 0
C/γ H = 0.05
C/γ H = 0.1
Kae

1.0 C/γ H = 0.15

C/γ H = 0.20
C/γ H = 0.25
C/γ H = 0.30
0.5

0.0
0 0.2 0.4 0.6 0.8 1
kh

(a)
2.0

1.5
C/γ H = 0
C/γ H = 0.05
C/γ H = 0.1
Kae

1.0 C/γ H = 0.15

C/γ H = 0.20
C/γ H = 0.25
C/γ H = 0.30
0.5

0.0
0 0.2 0.4 0.6 0.8 1

kh
(b)

346

Seismic Design of Foundations

Foundations in slopes and for retaining walls

11.6. Foundations for retaining walls

There are many types of retaining walls. Some of the more common walls (illustrated in Figure 11.16) are
listed below

■ Gravity retaining wall rely on their huge weight to retain the material behind them and achieve
stability against failures. They are generally constructed from concrete, stone or even brick masonry.
■ Cantilever retaining walls consist of a relatively thin stem and a base slab, and are generally
constructed of reinforced concrete. They use much less concrete than monolithic gravity walls, but
require more design and careful construction.
■ Anchored bulkheads, which are extensively used in waterfront areas, are anchored close to the upper
ends of sheet pile walls. The ties or anchored rods reduce the lateral deflection, the bending moment
and the depth of penetration.
■ Tieback walls are brick walls reinforced with wires for stability. The wires, referred to as tiebacks,
have one end secured to the wall while the other end is anchored to a stable structure (e.g. a concrete
deadman or anchored into earth with sufficient resistance) and resist earth pressures that would
otherwise cause the wall to lean.
■ Reinforced soil retaining walls are mechanically stabilised earth walls; that is, they are constructed
using steel or geotextile soil reinforcements placed in layers within a controlled granular fill.

Among these types of walls, gravity and cantilever retaining structures have foundations to transmit the forces
acting on them into the ground; the other types have other mechanisms in place to transmit the forces to the
ground or backﬁll and make them stable. While a cantilever retaining wall has a base slab that acts as its
foundation, gravity retaining walls can have a base with or without a heel; as mentioned above, a key feature
of the gravity wall is that the weight of the wall itself plays a signiﬁcant role in the support of the retained
material. Figure 11.17 shows typical forms of gravity-type and cantilever-type retaining walls.

Figure 11.16 Common types of earth-retaining structures

Gravity wall Cantilever wall Anchored bulkhead

Tieback wall Reinforced soil wall

347

Seismic Design of Foundations

Figure 11.17 (a) Gravity retaining walls; (b) cantilever retaining walls

Keys

Stone masonry Brick Plain concrete

(a)

Stem Stem Stem

Base Base Base

Toe Toe Toe

Heel Heel Heel

(b)

When designing foundations for gravity-type and cantilever type retaining walls, the stability of the wall
should be considered with respect to the following modes of failure

■ Sliding failure. The wall will fail by sliding when the magnitude of the lateral earth pressure (static
and seismic) exceeds the capacity of the footing in resistance. This resistance, as developed solely by
the footing, consists of resistance to shear developed on the contact surface between the footing and
the ground underneath; passive earth pressure acting on the front edge of the footing (and on the front
edge of the shear key, if used); and on the shear key projecting from the footing into the ground
beneath in order to add resistance against sliding.

348

Seismic Design of Foundations

Foundations in slopes and for retaining walls

■ Overturning failure. The wall will fail by overturning when the overturning effect of the seismic
earth pressure exceeds the capacity of the wall in resistance. Generally, the resistance consists of
the following: the weight of the wall; the weight of the ground above the base of the wall and the
weight of any surcharge acting on this portion of the soil, provided this surcharge contributes to the
overturning effect; and the shear resistance of the soil on the vertical plane located at the back edge of
the footing.
■ Bearing failure. The wall will fail by bearing when the magnitude of normal stresses on the
foundation ground induced by both the weight of the wall and surcharge as well as the lateral earth
pressure exceeds the bearing capacity of the footing. A bearing failure of the soil generally occurs at
the toe of the foundation, and leads to forward rotation of the wall.
■ Global instability. This type of failure involves not only the wall but also the surrounding ground,
and occurs along a slip surface (typically cylindrical) passing through the heel of the retaining wall
due to excessive shear stresses along the slip surface within the soil mass. This condition may be
exacerbated by the presence of a soft liqueﬁable layer underneath the retaining wall.

These modes of failure are illustrated in Figure 11.18. A complete design should address each of these modes
of failure where appropriate.

Figure 11.18 Modes of failure of retaining walls: (a) sliding; (b) overturning; (c) bearing capacity; (d) global
instability

Rotation
Movement

(a) (b)

Rotation

Slip surface Slip surface

349

Seismic Design of Foundations

11.7. Worked example

11.7.1 Problem statement
Consider the gravity retaining structure shown in Figure 11.19. For a design horizontal seismic coefﬁcient
kh = 0.15 (with vertical seismic coefﬁcient kv = 0.0), calculate the following

(a) total active thrust on the wall

(b) point of application of the total active thrust
(c) factor of safety against sliding
(d) factor of safety against overturning.

11.7.2 Solution
(a) To estimate the total active thrust on the wall, ﬁrst solve for PAE. Using Equation 11.2

k
q = tan−1 1 −hk = tan−1 (0:15) = 8:53°
v (11:12)
b = 0° i = 0° d = 15° f = 32° f − i > q OK!

Figure 11.19 Gravity retaining wall for the worked example

Concrete
γwall = 25 kN/m3
0.8 m

Sand backfill
γ = 17 kN/m3
φ = 32°
8m δ = 15°

Natural sand
3m
γ = 18 kN/m3
φ = 32°

350

Seismic Design of Foundations

Foundations in slopes and for retaining walls

Solving for the dynamic active earth pressure coefﬁcient (Equation 11.1)

cos2 (f − q − b)
KAE = sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!2
sin(f + d ) sin(f − q − i)
cos q cos2 b cos(d + b + q) 1 +
cos(d + b + q) cos(i − b)

cos2 (32 − 8:53 − 0) (11:13)

= sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!2
sin(32 + 15) sin(32 − 8:53 − 0)
cos(8:53) cos2 (0) cos(15 + 0 + 8:53) 1 +
cos(15 + 0 + 8:53) cos(0 − 0)
= 0:380

The active seismic thrust (per unit m of wall) is

1
PAE = K g H 2 (1 − kv )
2 AE
(11:14)
= 0:5 0:38 17 kN/m3 (10:0)2 (1 − 0)
= 322:75 kN/m

(b) To estimate the point of application, PA is required. The coefﬁcient of static active pressure from the
Coulomb theory is given by Equation 11.5

cos2 (f − b)
KA = sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!2
sin(f + d ) sin(f − i)
cos b cos(d + b) 1 +
2
cos(d + b) cos(i − b)

cos2 (32 − 0) (11:15)

= sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!2
2 sin(32 + 15) sin(32 − 0)
cos (0) cos(15 + 0) 1 +
cos(15 + 0) cos(0 − 0)
= 0:279

The active thrust (per unit metre of wall) is given by

1
PA = K g H2
2 A (11:16)
= 0:5 0:279 17 kN/m (10 m) = 237:15 kN/m
3 2

This acts at a distance H/3 = 10/3 = 3.33 m from the base. Therefore, the dynamic component (per
unit metre of wall) is calculated using Equation 11.3 as

DPAE = PAE − PA = 322:75 − 237:15 = 85:60 kN/m (11:17)

351

Seismic Design of Foundations

This acts at (approximately) a distance of 0.6H = 6.0 m from the base. The point of application of PAE
is then (from Equation 11.6)

ð237:15 3:33Þ + ð85:60 6:0Þ

Y= = 4:04 m (11:18)
322:75
above the base of the wall.

(c) To solve for the factor of safety against sliding, the contribution of the dynamic passive pressure
exerted by the soil in front of the wall may be ignored (conservative approach). If this is not ignored,
the magnitude of PPE is calculated using Equation 11.7

cos2 (f − q + b)
KPE = sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!2
sin(f + d ) sin(f − q + i)
cos q cos b cos(d − b + q) 1 +
2
cos(d − b + q) cos(i − b)

cos2 (32 − 8:53 + 12:4) (11:19)

= sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!2
2 sin(32 + 15) sin(32 − 8:53 − 0)
cos(8:53) cos (12:4) cos(15 + 0 + 8:53) 1 +
cos(15 − 12:4 + 8:53) cos(0 − 12:4)
= 3:524

1
PPE = K g H 2 (1 − kv )
2 PE
(11:20)
= 0:5 3:524 17 kN/m3 (2:0)2 (1 − 0)
= 119:82 kN/m

The point of application of PPE is not really deﬁned; for simplicity, this is considered to be at a
distance of H/3 = 0.67 m from the base. The factor of safety against sliding is

resisting force FR
FSsliding = = (11:21)
driving force FD

FD includes the horizontal component of PAE plus the inertia force on the wall. The weight of the
wall is

W1 = 25 kN/m3 0:80 10 = 200 kN/m (11:22)

W2 = 25 kN/m3 0:5 2:2 10 = 275 kN/m (11:23)

and

FD = kh ðW1 + W2 Þ + PAE cos d

(11:24)
= 0:15 ð200 + 275Þ + 322:75 cos 15° = 385:08 kN/m

352

Seismic Design of Foundations

Foundations in slopes and for retaining walls

The resisting force is calculated a: FR = N tan 32° + horizontal component of PPE

FR = ½ðW1 + W2 Þ + PAE sin d + PPE sinðd + 12:4°Þ tan 32° +PPE cosðd + b Þ
= ½ð200 + 275Þ + 322:75 sin(15° ) + 119:82 sin(15° +12:4° ) tan 32°
(11:25)
+ 119:82 cos(15° +12:4° )
= 438:83 kN/m

Hence
FR 438:83
FSsliding = = = 1:14 (11:26)
FD 385:08

If the effect of PPE is ignored, then

FR 550:34 tan 32°

FSsliding = = = 0:79 (11:27)
FD 385:08

(d) To solve for the factor of safety against overturning

resisting moment MR
FSoverturning = = (11:28)
driving moment MD

Again, the contribution of passive pressure may be ignored (conservative approach). If this is not
ignored, then MD includes the horizontal component of PAE plus the inertia force on the wall

MD = 0:15 (200 5) + 0:15 (275 3:33) + 322:75 cos 15° 4:04

(11:29)
= 1555:25 kN m/m

and MR includes the vertical component of PAE plus the weight of the wall plus the effect of PPE

MR = (200 2:6) + (275 1:47) + 322:75 sin 15° 3:0

+ 119:82½cos 27:4° (0:67) + sin 27:4° (0:16) (11:30)
= 1231:24 kN m/m

MR 1231:24
FSoverturning = = = 0:79 (11:31)
MD 1555:25

If PPE is neglected, then the resisting moment is decreased

MR = (200 2:6) + (275 1:47) + 322:75 sin 15° 3:0

(11:32)
= 1150:28 kN m/m

MR 1150:28
FSoverturning = = = 0:74 (11:33)
MD 1555:25

353

Seismic Design of Foundations

Most design codes set a minimum factor of safety against sliding and overturning to 1.5 if wind/seismic
forces are considered. Considering that both FSsliding and FSoverturning are both <1.5, it can be said that
the wall is unstable for the speciﬁed design seismic coefﬁcient. Moreover, if the effect of the passive
seismic thrust of the soil in front of the wall (PPE) is neglected, the factors of safety are smaller than
those for the case when PPE is taken into account; hence neglecting PPE in the analysis is a conservative
approach.

11.8. Case studies: seismic performance of retaining walls

11.8.1 1995 Kobe earthquake (Japan)
Following the 17 January 1995 Hyogoken Nambu (Kobe) earthquake (Mw 7.2), many conventional
masonry and unreinforced concrete gravity retaining walls were seriously damaged; similarly, many
modern cantilever-type reinforced concrete retaining walls also collapsed. These failures were not
associated with liquefaction of the foundation ground or of the backﬁll soil, but were caused by the high
seismic forces on the wall structure and backﬁll due to the earthquake.

Tatsuoka et al. (1996a, 1996b) examined the seismic performance of an extensive length of retaining
structures for railway embankments. They reported that different types of soil-retaining walls in the severely
shaken areas performed differently: masonry retaining walls, leaning (supported) unreinforced concrete
retaining walls and gravity-type unreinforced concrete retaining walls were the most seriously damaged,
while the damage to cantilever walls was generally less serious.

Figure 11.20 shows the collapse of a leaning-type unreinforced concrete retaining wall, which was con-
structed in the 1930s to support the embankment for the Tōkaidō Main Line railway. Tatsuoka et al. (1996b)
reported that a 600 m-long portion of the retaining wall collapsed, either being broken at the level of the
ground surface and the upper part overturned completely or overturned about the bottom, resulting in the
back-face upside down conﬁguration post-earthquake. They argued that this damage was induced by both a
large horizontal seismic force acting on the retaining wall itself and a large seismic earth pressure imposed on
its back face from the backﬁll.

An example of collapsed gravity-type unreinforced concrete retaining wall is illustrated in Figure 11.21.
Constructed in the 1920s based on the standard design of that time using a horizontal seismic coefﬁcient of
kh = 0.2, a 200 m section of the wall completely overturned. Tatsuoka et al. (1996b) reported that the
collapse can be attributed not only to the large seismic force on the structure and the large seismic earth
pressure but also to the low strength of the wall structure itself and the low bearing capacity of the
supporting subsoil below the toe of the wall.

Figure 11.22 shows the collapse of a cantilever-type reinforced concrete wall that tilted largely in similar
fashion to the adjoining gravity wall shown in Figure 11.16. Constructed in the 1960s, this relatively modern
type of retaining wall mostly tilted outwards, causing a large settlement at the top of the railway embankment
(Tatsuoka et al., 1996b).

Additionally, Tatsuoka et al. (1996a, 1996b) reported the very good performance of a geosynthetic reinforced
soil (GRS) wall at Tanata with a full-height rigid facing. Constructed 3 months before the Kobe earthquake on

354

Seismic Design of Foundations

Foundations in slopes and for retaining walls

Figure 11.20 Damage to a leaning-type reinforced concrete wall supporting an embankment along the Tōkaidō
Main Line between Setsu-Motoyama and Sumiyoshi Station: (a) sketch of the failure (Tatsuoka et al., 1996b;
reprinted with permission from the author); (b) collapsed retaining wall (Courtesy of F. Tatsuoka)

Original position of retaining wall

2.6 m

Moved position
(a)

(b)

355

Seismic Design of Foundations

Figure 11.21 Damage to a gravity-type unreinforced concrete wall supporting an embankment along the Hanshin
Main Line adjacent to Ishiyagawa Station: (a) sketch of the failure (Tatsuoka et al., 1996b; reprinted with permission
from the author); (b) collapsed retaining wall (Courtesy of F. Tatsuoka)

5.0 m Original position

Moved position

(a)

(b)

356

Seismic Design of Foundations

Foundations in slopes and for retaining walls

Figure 11.22 Damage to a cantilever-type reinforced concrete wall supporting an embankment along the Hanshin
Main Line adjacent to Ishiyagawa Station: (a) sketch of the failure (Tatsuoka et al., 1996b; reprinted with permission
from the author); (b) collapsed retaining wall (Courtesy of F. Tatsuoka)

0.75 m
Moved position

Original position

5.0 m
Crack

3.4 m
4.0 m
(a)

(b)

357

Seismic Design of Foundations

the southern slope of an existing embankment for the JR Kobe Line, the wall underwent a residual lateral
displacement of about 26 cm at the top and 10 cm at the ground surface level. Figure 11.23 shows the wall
adjacent to collapsed wooden houses, indicating a shaking intensity estimated to be as high as 0.7g. However,
it was reported that the GRS wall was designed for only kh = 0.2.

Figure 11.23 The GRS retaining wall in Tanata supporting an embankment for the JR Kobe Line: (a) cross-section
(Tatsuoka et al., 1996b; reprinted with permission from the author); (b) condition after the earthquake (Courtesy
of F. Tatsuoka)

Noise barrier 3.5 m

Original slope
3.0 m
0.8 m 2.0 m

Cap 2.5 m
concrete

H–shaped pile
4.5 m
0.5 m

Geogrid
(a)

(b)

358

Seismic Design of Foundations

Foundations in slopes and for retaining walls

Based on field investigations and back analyses on the seismic performance of damaged retaining walls,
Koseki et al. (1996, 1999) showed that there is a large difference between the seismic coefficients (kh)design
used in the current design for retaining walls and the ratios of the highest peak horizontal ground accelerations
to the gravitational acceleration (amax/g) estimated at the damaged retaining walls. Tatsuoka et al. (1998)
explained that such a difference may have been affected by various factors, including the underestimation of
the backfill soil shear strength and the ductility of the retaining wall, which is not considered in pseudostatic
approaches. Moreover, the M-O theory is based on assumptions that are different from the actual phenom-
enon, including the fact that the actual direction of seismic force is reversed alternately (unlike the one-
direction assumption in the M-O theory).

11.8.2 1999 Chi-Chi earthquake (Taiwan)

An earthquake of Mw 7.6 occurred in Taiwan on 21 September 1999 – the largest to hit the country in recent
history. The Chi-Chi (Ji-Ji) earthquake recorded horizontal PGAs as high as 0.79g, and caused signiﬁcant
damage to a number of soil structures in the central part of the country due to landslides, soil liquefaction,
foundation failures and ground movements. Huang (2000) and Fang et al. (2001, 2003) discussed the failure
of many retaining structures induced by the earthquake, including masonry walls, gravity walls, modular-
block retaining walls and reinforced earth-retaining structures.

In areas where mountain roads were constructed on steep slopes by cut-and-fill, many gravity-type walls
supporting the embankment as well as leaning-type concrete walls collapsed following the earthquake,
typically by tilting and sliding. Figure 11.24 illustrates an example of a collapse of a gravity-type retaining
wall along Mountain Route 49 between Tsaolin and Chushan, as well as the estimated cross-section. The road
is cut-and-fill constructed on a steep slope with an inclination of about 50°. The retaining wall had a height of
about 7 m and was standing almost vertically, and may thus be categorised as a gravity-type retaining wall;
however, it is rather slender, and the foundation width was small for a gravity retaining wall (Uchimura et al.,
2000). The wall base was insufficiently embedded into the foundation ground and no special treatment was
made to strengthen the bearing capacity. Moreover, it appeared that the 50 cm-thick asphalt pavement was
overlaid in several layers, indicating that instability had occurred at this location previously, possibly due to
heavy rainfalls. Therefore, it can be considered that the cause of the collapse may be due to the low bearing
capacity of the foundation ground and the low safety margin of the wall against external loadings.

Another gravity-type wall that overturned is shown in Figure 11.25. The wall, which was located along
Provincial Road 21 near Tou-Sheh, was 2.5 m high, 0.5 m wide at the top and 1.5 m wide at the base; the
collapse portion involved a section 40 m long. Weep holes, 80 mm in diameter with a spacing of 2.5 m, had
been constructed through the wall. The site was quite close to the epicentre of the Chi-Chi earthquake. The
horizontal PGA recorded at a nearby strong motion station was 0.45g (Fang et al., 2001). The unexpectedly
strong ground shaking was probably the main cause of the overturning; it was also possible that insufﬁcient
bearing capacity of the foundation ground below the toe was another contributing factor to the collapse of the
wall.

From the results of field investigations reported by Koseki and Hayano (2000) and Fang et al. (2001, 2003), it
can be concluded that the main factors that caused damage to conventional retaining walls were the following:
(a) permanent ground displacements associated with fault movements; (b) large-scale movement of the
backfill slope; (c) loss of bearing capacity of the underlying soils; (d) excessive seismic force on the wall and
backfill; and (e) insufficient compaction of the backfill soil.

359

Seismic Design of Foundations

Figure 11.24 Collapse of a gravity-type retaining wall on a hillside slope in Taiwan: (a) estimated cross-section;
(b) condition after the earthquake (Uchimura et al., 2000)

5.0 m
Roadway
0.6 m 3.6 m Asphalt
0.7 m 70°
Asphalt

~7.0 m
Estimated slip surface

Sandy gravel

(a)

(b)

360

Seismic Design of Foundations

Foundations in slopes and for retaining walls

Figure 11.25 Overturning failure of a gravity-type retaining wall in Taiwan: (a) estimated cross-section; (b) condition
after the earthquake (Fang et al., 2003; reproduced with permission from the publisher, Canadian Science Publishing)

0.5 m 18°
Overturning

.3
Backfill

1:0
Filter
material

2.50 m
PVC
weep hole 0.3 m
Concrete

1.25 m
(a)

(b)

361

Seismic Design of Foundations

11.9. Summary
Past severe earthquakes have shown that building foundations on slopes and retaining structures are vul-
nerable to failure. Due to seismically induced forces, structures on slopes may undergo either bearing capacity
failure or general failure due to the instability of the foundation ground and adjacent portions within the slope.
Similarly, retaining structures are subjected to higher lateral earth pressure resulting from seismic shaking.

In simplified pseudostatic analysis of slopes and retaining structures, the effects of earthquake force is
accounted for as an equivalent horizontal force acting on the sliding mass, expressed as the product of the
weight of the mass and a seismic coefficient. While there are no current specific rules for the selection of an
appropriate seismic coefficient for design, it is usual practice to base it on the anticipated level of acceleration
within the failure mass, taking into account the importance of the structure and an acceptable level of seismic
performance (i.e. the amount of allowable seismic displacement).

Recent large-scale earthquakes have shown that most of the damage to houses on slopes is due to the ground
displacement of the soft ﬁlled soils used for the hillside embankment. Thus, when analysing the stability of
house foundations on slopes, both the stability of the foundation ground (bearing capacity problem) as well as
the overall slope stability of the slope itself should be taken into account. Similarly, the seismic design of
retaining structures should address the various modes of failure, including sliding, overturning and bearing
failures as well as overall instability. To minimise the expected displacements of foundations in slopes as well
as of retaining walls during a strong earthquake, appropriate countermeasures (ground improvement,
restraining piles, etc.) should be implemented.

REFERENCES
Anderson DG, Martin GR, Lam I and Wang JN (2008) Seismic Analysis and Design of Retaining Walls, Buried
Structures, Slopes and Embankments. National Cooperative Highway Research Program, Transportation
Research Board, Washington DC.
Bowles JE (1982) Foundation Analysis and Design, 3rd edn. McGraw-Hill, New York, NY, USA.
Bray JD and Travasarou T (2009) Pseudostatic coefﬁcient for use in simpliﬁed seismic slope stability evaluation.
Journal of Geotechnical and Geoenvironmental Engineering, ASCE 135(9): 1336–1340.
Bray JD, Travasarou T and Zupan J (2010) Seismic displacement design of earth retaining structures. 2010 Earth
Retention Conference (ER2010), Bellevue, Washington, DC, USA, pp. 639–655.
Castelli F and Motta E (2010) Bearing capacity of strip footings near slopes. Geotechnical and Geological
Engineering 28: 187–198.
Dellow G, Yetton M and Massey C et al. (2011) Landslides caused by the 22 February 2011 Christchurch
earthquake and management of landslide risk in the immediate aftermath. Bulletin of the New Zealand Society
for Earthquake Engineering 44(4): 227–238.
Ebeling RM and Morrison EE (1992) The Seismic Design of Waterfront Retaining Structures. Corps of Engineers
Waterways Experiment Station, Vicksburg, MS, USA, Technical Report ITL-92-11.
Fang YS, Chen TJ, Yang YC and Tang CC (2001) The behavior of retaining walls under 1999 Chi-Chi earthquake.
Proceedings of the 4th International Conference on Recent Advances in Geotechnical Earthquake Engineering
and Soil Dynamics, San Diego, CA, USA, paper 10.06.
Fang YS, Yang YC and Chen TJ (2003) Retaining wall damage in the Chi-Chi earthquake. Canadian Geotechnical
Journal 40: 1142–1153.
Georgiadis K (2010) Undrained bearing capacity of strip footings on slopes. Journal of Geotechnical and
Geoenvironmental Engineering, ASCE 136(5): 677–686.
Hansen J (1970) A Revised and Extended Formula for Bearing Capacity. Bulletin No. 28. Danish Geotechnical
Institute, Copenhagen, Denmark.

362

Seismic Design of Foundations

Foundations in slopes and for retaining walls

Huang CC (2000) Investigation of the damaged soil retaining structures during the Chi-Chi earthquake. Journal of
the Chinese Institute of Engineers 23(4): 417–428.
Hynes-Griffin ME and Franklin AG (1984) Rationalizing the Seismic Coefficient Method. US Army Corps of
Engineers Waterways Experiment Station, Vicksburg, MS, USA.
Hyodo M, Orense R, Furukawa M, Noda S and Furui K (2012) Slope failures in residential land on valley fill in
Yamamoto town. Soils and Foundations 52(5): 975–986 (special issue).
Koseki J and Hayano K (2000) Preliminary report on damage to retaining walls caused by the 1999 Chi-Chi
earthquake. Bulletin of Earthquake Resistant Structures Research Center 33: 23–34.
Koseki J, Tateyama M, Tatsuoka F and Horii K (1996) Back analyses of soil retaining wall for railway
embankments damaged by the 1995 Hyogoken Nanbu earthquake. In The 1995 Hyogoken-nanbu Earth-
quake—Investigation into Damage to Civil Engineering Structures (Committee of Earthquake Engineering
(ed.)). Japan Society of Civil Engineers, Tokyo, Japan, pp. 101–114.
Koseki J, Munaf Y, Tateyama M, Kojima K and Horii K (1999) Back analyses of case histories and model tests on
seismic stability of retaining walls. Proceedings of the 11th Asian Regional Conference on Soil Mechanics and
Geotechnical Engineering. Balkema, Rotterdam, the Netherlands, vol. 1, pp. 399–402.
Kumar J and Rao V (2003) Seismic bearing capacity of foundations on slopes. Géotechnique 53: 347–360.
Loukidis D and Salgado R (2009) Bearing capacity of strip and circular footings in sand using finite elements.
Computers and Geotechnics 36: 871–879.
Marcuson WF and Franklin AG (1983) Seismic design, analysis, and remedial measures to improve the stability of
existing earth dams – Corps of Engineers approach. Proceedings of the Seismic Design of Embankments and
Caverns. American Society of Civil Engineers, Reston, VA, USA.
Mononobe N and Matsuo H (1929) On the determination of earth pressure during earthquakes. Proceedings of the
World Engineering Conference 9: 177–185.
Mori T, Tobita Y and Okimura T (2012) The damage to hillside embankments in Sendai City during the 2011 off
the Pacific Coast of Tohoku earthquake. Soils and Foundations 52(5): 910–928 (special issue).
Noda S, Uwabe T and Chiba T (1975) Relation between seismic coefficient and ground acceleration for gravity
quay wall. Report of the Port and Harbour Research Institute 14(4). (In Japanese.)
Okabe S (1926) General theory on earth pressure and seismic stability of retaining walls and dams. Journal of the
Japanese Society of Civil Engineering 12(1): 123–134. (In Japanese.)
Richards R and Shi X (1994) Seismic lateral pressures in soils with cohesion. Journal of Geotechnical Engineering
120(7): 1230–1251.
Richards R, Elms DG and Budhu M (1990) Dynamic fluidization of soils. Journal of Geotechnical Engineering
116(5): 740–759.
Saran S and Gupta RP (2003) Seismic earth pressures behind retaining walls. Indian Geotechnical Journal 33(3):
195–213.
Saran S and Prakash S (1968) Dimensionless parameters for static and dynamic earth pressures behind retaining
walls. Indian Geotechnical Journal 7(3): 295–310.
Seed HB (1979) Considerations in the earthquake-resistant design of earth and rockfill dams Géotechnique 29(3):
215–263.
Seed HB and Whitman RV (1970) Design of earth retaining structures for dynamic loads. ASCE Specialty
Conference: Lateral Stresses in the Ground and Design of Earth Retaining Structures, Ithaca, NY, USA,
pp. 103–147.
Shukla SK, Gupta SK and Sivakugan N (2009) Active earth pressure on retaining wall for c–f soil backfill under
seismic loading condition. Journal of Geotechnical and Geoenvironmental Engineering 135(5): 690–696.
Tatsuoka F, Koseki J and Tateyama M (1996a) Performance of reinforced structures during the 1995 Hyogo-ken
Nanbu earthquake special report: earth reinforcement. Proceedings of the International Symposium on Earth
Reinforcement, IS Kyushu ‘96, Fukuoka, Japan, vol. 2 preprint, pp. 105–140.

363

Seismic Design of Foundations

Tatsuoka F, Tateyama M and Koseki J (1996b) Performance of soil retaining walls for railway embankments.
Soils and Foundations 36: 311–324 (special issue).
Tatsuoka F, Koseki J, Tateyama M, Munaf Y and Horii K (1998) Seismic stability against high seismic loads of
geosynthetics reinforced soil retaining structures. Proceedings of the 6th International Conference on
Geosynthetics, Atlanta, USA, vol. 1, pp. 103–142.
Terzaghi K (1950) Mechanisms of Landslides. Geological Society of America, Boulder, CO, USA.
Uchimura T, Koseki J, Tatsuoka F, Hayano K and Huang CC (2000) Damage investigation of retaining walls and
reinforced soil structures affected by the Chi-Chi Taiwan earthquake. Geosynthetics Technical Information
16(1): 17–22. (In Japanese.)
US Army Corps of Engineers (1982) Slope Stability Manual EM-1110-2-1902. US Army Corps of Engineers,
Washington DC, USA.

364

Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

Chapter 12
Liquefaction countermeasures
12.1. Introduction
Many innovative soil improvement techniques have been developed in the last 50 years to mitigate seismic
effects, including liquefaction. New techniques are still being developed when difficult circumstances are
encountered in the field, while available techniques are being refined to better suit different site conditions or
construction constraints.

There is no attempt to provide a comprehensive discussion of all available liquefaction countermeasures in

this chapter; rather, only commonly used methods are discussed in detail. References to more complete
descriptions of each method are provided. Because ground improvement technologies change rapidly and as
new techniques are developed and existing techniques are reﬁned and tested by actual earthquakes, the
relevant geotechnical literature should be constantly reviewed.

12.2. Liquefaction countermeasures

Before selecting and designing a ground improvement technique to mitigate liquefaction effects at a site, it is
necessary to understand the performance requirements of the improved ground and the structural system.
Note that in some countries, differences arise in the performance requirements between new and existing
structures.

The general philosophy for the design of ground improvement is to eliminate liquefaction and the associated
ground deformations (lateral spreading and settlement) or mitigate their effects to the extent needed to meet
the design performance required for each damage state of a structure.

Performance measures, comparison of the relative performance and the cost of different foundation systems
should be evaluated. Remedial measures against liquefaction can be classiﬁed into the following categories

■ to treat liquefiable soil so that liquefaction will not occur (e.g. ground improvement)
■ to strengthen structures or make them flexible to relieve the effects of liquefaction (e.g. structural
retrofitting).

Consideration should be given with regard to whether ground improvement is the best solution. Other
solutions, or means to mitigate the effects of liquefaction, can be implemented on their own or in conjunction
with ground improvement – the simplest of which is to avoid liqueﬁable sites where possible. Alternatively,
improved structural measures can be incorporated to reduce damage susceptibility due to liquefaction, and
improve resilience.

12.2.1 Ground improvement

Ground improvement methods are generally based on one or more of the following principles: (a) densiﬁ-
cation, (b) solidiﬁcation, (c) reinforcement and containment, (d) drainage, (e) increasing in situ stress and (f) soil
replacement. Table 12.1 summarises the most common liquefaction countermeasures. Brief descriptions of the

365

Seismic Design of Foundations

366
Table 12.1 Summary of the most common ground improvement methods

Seismic Design of Foundations

Method Principle of Description Most suitable soil Maximum effective Relative cost
improvement conditions/types treatment depth

Heavy tamping (dynamic ■ Densiﬁcation Repeated application of Cohesionless soils are 30 m (possibly deeper) Low
Seismic Design of Foundations

compaction) ■ Increased in high-intensity impacts at surface best, other types can

situ stress also be improved

Vibratory probe ■ Densiﬁcation Densiﬁcation by vibration; Saturated or dry clean 20 m routinely (ineffec- Moderate
■ Terraprobe ■ Increased in liquefaction-induced settlement sand; sand tive above 3–4 m depth);
■ Vibro-rods situ stress and settlement in dry soil under >30 m sometimes;
■ Vibro-wing overburden to produce a higher vibro-wing, 40 m
density

Vibro-compaction ■ Densification Densification by vibration and Cohesionless soils with >20 m Low to moderate
■ Vibroflot ■ Increased in compaction of backfill material of less than 20% fines
■ Vibro-composer system situ stress sand or gravel

Vibro-replacement stone Hole jetted into fine-grained soil Sands, silts and clays >30 m (limited by Moderate
and sand columns and backfilled with densely vibratory equipment)
■ Grouted compacted gravel or sand hole
■ Not grouted formed in cohesionless soils by
vibro techniques and compaction
of backfilled gravel or sand; for
grouted columns, voids filled
with a grout

Compaction piles ■ Densification Densification by displacement of Loose sandy soil; partly > 20 m Moderate to high
■ Reinforcement the pile volume and by vibration saturated clayey soil;
■ Increased in during driving; increase in the loess
situ stress lateral effective earth pressure
SolidificationDisplacement ■ Densification Highly viscous grout acts as a All soils except at Unlimited Low to moderate
(compaction grout) ■ Reinforcement radial hydraulic jack when shallow depths (low
■ Increased in pumped in under high pressure overburden pressures

Seismic Design of Foundations

situ stress may result in ground
heave)

Surcharge or buttress ■ Densiﬁcation The weight of a surcharge/ Can be placed on any Dependent on the size Moderate if
■ Increased in buttress increases the liquefaction soil surface of surcharge/buttress vertical drains are
situ stress resistance by increasing the used
effective conﬁning pressures in
the foundation

Mix-in-place piles and walls Lime, cement or asphalt Sands, silts, clays, and >20 m (60 m obtained in High
introduced through a rotating all soft or loose Japan)
auger or special in-place mixer inorganic soils

Jet grouting High-speed jets at depth Sands, silts and clays Unknown High
excavate, inject and mix a
stabiliser with soil to form
columns or panels

Particulate grouting ■ Solidiﬁcation Penetration grouting ﬁlls soil Medium to coarse sand Unlimited Lowest of the
■ Reinforcement pores with soil, cement and/or and gravel grout methods
clay

Chemical grouting ■ Solidiﬁcation Solutions of two or more chemi- Medium silts and Unlimited High
■ Reinforcement cals react in soil pores to form a coarser
gel or a solid precipitate

Drains ■ Drainage Relief of excess pore water Sand, silt and clay Gravel and sand >30 m; Moderate to high
■ Gravel pressure to prevent liquefaction depth limited by
■ Sand (wick drains have comparable vibratory equipment;
■ Wick permeability to sand drains); wick >45 m
■ Wells (for permanent gravel drains primarily used –
dewatering) sand/wick may supplement

(continued on next page)

Liquefaction countermeasures

367
368

Seismic Design of Foundations

Table 12.1 (Continued)

Method Principle of Description Most suitable soil Maximum effective Relative cost
Seismic Design of Foundations

improvement conditions/types treatment depth

them or relieve existing excess

pore water pressure; permanent
dewatering with pumps

Pressure injected lime Penetration grouting ﬁlls soil Medium to coarse Unlimited Low
pores with lime sand and gravel

Electrokinetic injection Stabilising chemical moved Saturated sands, silts Unknown Expensive
into and ﬁlls soil pores by and silty clays
electro-osmosis, or colloids
enter pores by electrophoresis

Root piles, soil nailing Small-diameter inclusions used All soils Unknown Moderate to high
to carry tension, shear and
compression

Blasting Shock waves and vibrations Saturated clean sand; >40 m Low
cause limited liquefaction, partly saturated sands
displacement, remoulding and silts after ﬂooding
and settlement to higher density
Liquefaction countermeasures

principles behind the improvement mechanism associated with each category are provided below, along with
references detailing the design of ground treatment schemes and methods to evaluate the effectiveness of soil
improvement. Further details concerning the speciﬁc treatment methods can be found in references on ground
improvement (e.g. Andrus and Chung, 1995; Cooke and Mitchell, 1999; Hausmann, 1990; JGS, 1998; Port
and Harbour Research Institute, 1997; Schaefer 1997; Xanthakos et al., 1994).

12.2.2 Structural enhancement

There are several methods directed to the reduction of structural damage even when liquefaction occurs.
These include

■ pile foundations to mitigate settlement and bearing capacity failure, tilt and uplift (buoyancy)
■ ductile piles to improve performance of structures on sites prone to lateral spreading
■ strengthening of the foundation (increase number of piles, use large piles or inclined piles)
■ flexible structures to cope with large total and differential settlements
■ control of ground deformation and structural performance by structural measures (rigidity of the
structure, rigid rafts, sheet piles to confine liquefiable material, geogrids or base isolation of
structures).

Structural strengthening has the following advantages: (a) some methods are more economical compared
with measures to prevent liquefaction occurrence; (b) some methods can be applied to existing structures.
However, the effectiveness of some of these structural measures must be adequately evaluated because design
methods have not been established and there has been little experience in applying this kind of remedial
measure.

12.3. Ground improvement methods

12.3.1 General
Excluding lowering of the groundwater table, there are various mechanisms by which ground improvement
methods treat soils to increase their resistance to liquefaction. These are

■ densiﬁcation
■ solidiﬁcation
■ reinforcement
■ drainage
■ increasing in situ stress
■ replacement.

Ground improvement techniques use one or a combination of these mechanisms to improve the resistance of
the ground to liquefaction and to improve seismic performance. Each improvement mechanism is described
in more detail below.

12.3.2 Densification
Densification or compaction methods involve rearranging the soil particles into a tighter configuration,
resulting in increased density. This increases the shear strength and liquefaction resistance of the soil, and
encourages a dilative instead of a contractive dynamic soil response. Densifying a loose sandy deposit with
vibration and/or impact has been used extensively, making it the most popular liquefaction countermeasure.

369

Seismic Design of Foundations

An increase in the soil density can be achieved through a variety of means. These include (JGS, 1998)

■ compaction by penetration – penetration of granular material (e.g. stone or sand columns or piles) into
the liqueﬁable deposit will push the surrounding soil, and result in a reduced void ratio, and therefore
an increase in the resistance of the soil to liquefaction
■ compaction by vibration – subjecting a loose sandy deposit to vibration energy will compact the soil
and increase its strength
■ compaction by impact energy – which can densify a loose granular deposit.

Densiﬁcation is a key improvement mechanism in dynamic compaction, rapid impact compaction, com-
paction, impact roller compaction, deep vibro-compaction, vibro-rod compaction, compaction piling, and
installation of stone or sand columns.

12.3.2.1 Stone columns

In the stone column method of ground improvement, columns of dense stone are compacted into the ground
in either a triangular or square grid across the site. The columns are typically spaced 1.5–4 m apart and have a
diameter of 0.6–1.2 m. The depth of improvement is typically 4–15 m, but soils as deep as 30 m have been
treated using this method in the past. Generally, the fill material consists of crushed coarse aggregates of
various sizes, with the particle size distribution prescribed in the design criteria. Depending on the con-
struction method, the installation of a grid of granular columns can improve liquefiable soil deposits not only
by densification but also by increasing the in situ lateral stress, replacing the liquefiable in situ soil with non-
liquefiable material, reinforcing the original ground with stiffer columns of fill materials and providing
drainage paths for the relief of excess pore water pressure.

Vibro-replacement stone columns are constructed using a vibroflot whose purpose is to create a dense column
of stone as well as to densify the surrounding natural soils. Either the wet top-feed process or the dry bottom-
feed process is used for construction. In the wet top-feed process, the vibrator penetrates to the design depth
by means of the vibrator’s weight and vibrations, as well as water jets located in the vibrator’s tip. The stone
(crushed stone or recycled concrete) is then introduced at the ground surface to the annular space around the
vibrator created by the jetting water. The stone falls through the annular space to the vibrator tip, and fills the
void created as the vibrator is lifted a metre or so at a time. The vibrator is then lowered, densifying and
displacing the underlying stone. The vibro-replacement process is repeated until a dense stone column is
constructed to the ground surface (Hayward Bakers, 2015). The dry bottom-feed process is similar except that
no water jets are used and the stone is fed to the vibrator tip through a feed pipe attached to or integrated into
the vibrator; this is illustrated in Figure 12.1. Predrilling of dense strata at the column location may be
required for the vibrator to penetrate to the design depth. Both methods of construction create a high-modulus
stone column that reinforces the treatment zone and densifies surrounding granular soils.

12.3.2.2 Sand compaction piles

Sand compaction piles (SCPs) are common in Japan and are similar to stone columns. Sand columns are
constructed by driving a sand-filled steel casing into the ground, after which the casing is then incrementally
withdrawn, and the deposited sand is compacted by re-driving to form a high-density sand column (JGS,
1998; Schaefer 1997). Figure 12.2 illustrates the procedure for installing SCPs. The sand backfill must have
specific properties, and sometimes may be difficult to source. Recent case studies indicate that metal slag,
crushed oyster shell, granulated fly ash, or recycled crushed concrete and asphalt are feasible alternatives to
sand (Kitazume, 2005).

370

Seismic Design of Foundations

Liquefaction countermeasures

Figure 12.1 Stages in vibro-replacement method using dry-bottom feed process. (a) Penetration: the vibro-probe
penetrates by vibration and with the aid of compressed air to the required depth; (b) installation: the stone column
is installed by adding gravel through the separate gravel pipe alongside the vibro-probe; (c) completion: the surface
is levelled and the surface compacted (Courtesy of Vibroﬂotation Group)

(a) (b) (c)

Figure 12.2 Procedure for installing SCPs (Courtesy of Fudo Tetra Corporation)

Positioning Insertion completed (Pile formation completed)

Transport Insertion Pile formation

1 6
Vibro-hammer
2 Pulling out Redrive
4
Sand 3 5
Casing depth Compaction Compaction

Enlarging
diameter

Sand pile

12.3.2.3 Dynamic Compaction

Dynamic compaction involves repetitively dropping a large weight from a significant height onto the ground,
causing the soil grains to rearrange and form a denser arrangement. Figure 12.3 illustrates the application of
dynamic compaction. Additionally, the impact of the dropped weight on the ground surface produces
dynamic stress waves, which can be large enough to generate significant excess pore water pressure in the
soils beneath the point of impact (Idriss and Boulanger, 2008). Dissipation of the excess pore water pressures
results in densification, accompanied by surface settlement. The drop height, weight and spacing vary,
depending on the ground and groundwater conditions.

Tampers are typically concrete or steel with a weight of 5–20 t, and dropped using crawler cranes from
heights of 12–40 m. With specialist equipment, drop weights of up to 170 t (Slocombe, 2004) and drop
heights of up to 114 m (Schaefer, 1997) have been used. Drop locations are organised in a grid pattern with a
spacing of 4–15 m. Compaction is carried out in a series of passes of different energy levels to treat different
layers within the depth of treatment. The ﬁrst pass targets the deeper layers with high energy tamping in a
relatively widely spaced grid pattern. Successive passes use lower energy levels and closer grid spacing to

371

Seismic Design of Foundations

Figure 12.3 Procedure for densifying soil through dynamic compaction (Courtesy of Vibro-Menard)

treat the intermediate and surface layers. Dynamic compaction is known to be fast and economic, especially
in treating large areas. However, it has obvious disadvantages due to the noise and vibration that are pro-
duced. Moreover, it is often effective only in the upper 5–10 m of the deposit, and it is less effective for soils
with high ﬁnes content.

12.3.2.4 Compaction grouting

In compaction grouting, a very stiff grout is injected into the soil such that it does not permeate the native soil
but results in coordinated growth of the bulb-shaped grout that pushes and displaces the surrounding soil
(Figure 12.4). Typically, the grout consists of a soil–cement–water mixture with sufﬁcient silt sizes to provide
plasticity, together with sand and gravel sizes to develop internal friction (Welsh, 1992). Note that the
strength of the grout is unimportant because the purpose of the technique is to densify the surrounding soil by
displacement. Since the technique involves the pressurised injection of grout into the soil deposit using small-
scale, manoeuvrable and vibration-free equipment, there is minimum disturbance to the structure and sur-
rounding ground during implementation. Additional advantages include relatively little site disruption and

Figure 12.4 Procedure for compaction grouting (Courtesy of Sanshin Corporation)

Drilling Grouting Step-up Step-up Step-up

grouting grouting grouting

4
3 3
2 2 2
1 1 1 1

1 step (h = 0.33 m)

372

Seismic Design of Foundations

Liquefaction countermeasures

the ability to work in a constricted space, resulting in greater economy. However, it has some disadvantages;
for example, stabilisation of near-surface soils is generally ineffective due to the fact that the overlying
restraint is small, and the grouting pressures can heave the ground surface rather than densify the soil.

12.3.3 Chemical modiﬁcation/solidiﬁcation

Increase in liquefaction resistance can also be obtained through a more stable skeleton of soil particles. This
can be achieved by solidiﬁcation methods, which involve either mixing cementitious or other additives into
the soil or ﬁlling the voids with a reagent, resulting in the soil particles being bound together. This will
prevent the development of excess pore water pressure, forestalling the occurrence of liquefaction. Another
possible effect is that the sandy soil is given cohesion by the stabiliser, and thus the ground can resist the
external force by its increased shear strength (due to the added cohesion). Typical methods include deep
mixing, jet grouting and permeation grouting techniques.

Solidification techniques can generally be used in a wide range of soil types, including those with high fines
content. They are advantageous because the installation methods are relatively quiet and induce relatively
small vibrations as compared with compaction methods. These are important considerations for the
improvement of sites with adjacent infrastructure or inhabitants that could be affected by noise and vibration
from densification techniques. The induced horizontal earth pressures are smaller than with compaction
methods, and are larger than with drainage methods. Their disadvantage is the relatively high cost–benefit
ratio as compared with compaction and drainage methods.

12.3.3.1 Deep soil mixing

The deep-mixing method achieves liquefaction remediation by agitating and mixing stabilising material such
as cement in sandy soil and solidifying the soil. A variety of plant has been developed for soil mixing. For
mass stabilisation, a rotating drum cutter attached to an excavator arm can be used to mix the soil with the
stabilising agent. This method is generally limited to treatment depths of about 6 m. Other methods use
rotating augers or blades attached to rods to mix soils in vertical columns up to depths of 30 m or more.
Cutting heads have also been attached to directional drilling plant to mix soils in horizontal beams below
existing structures. The soil–cement materials can have a wide range of unconﬁned compressive strengths,
depending on the homogeneity of the mixing, the degree of compaction imparted by the mixing technique,
the amount of cementitious material used and the in situ soil characteristics (Kitazume and Terashi, 2013).

Deep soil-mixed columns can be applied over a whole area of liqueﬁable soil (block type) or over partial
sections (wall, pile or lattice type), as indicated in Figure 12.5. In the block type, the mix columns overlap
each other, and the whole layer is improved due to the increased stiffness. In partial improvement, unim-
proved sections remain and may liquefy; however, depending on the shape of the installation, the mix
columns are effective in preventing liquefaction by restraining the shear deformation of the ground through
reinforcement and conﬁnement. Soil–cement-mixed materials have low ductility, and the layout of the
columns needs to be designed to guard against a brittle-type failure.

12.3.3.2 Jet grouting

In jet grouting, high-pressure jets of air, water and/or grout are injected into the native soil in order to break up
and loosen the ground and mix it with a thin slurry of cementitious materials. In essence, it is not truly
grouting but rather a mix-in-place technique to produce a soil–cement material. Depending on the application
and the soils to be improved, different kinds of jet are combined by using a single-fluid system (slurry grout jet
only), a double-fluid system (slurry grout jet surrounded by an air jet) or a triple-fluid system (water jet
surrounded by an air jet, with a lower grout jet). The process can construct grout panels or full columns, or
anything in between (partial columns), with a specified strength and permeability.

373

Seismic Design of Foundations

Figure 12.5 Basic deep-mixing treatment patterns

Block Wall Lattice Pile

Tangent circle Tangent circle Tangent circle Tangent circle

block wall lattice lapping block

Construction of jet grout columns involves drilling to the base of the column then mixing cement slurry into
the soil in situ with rotating high-pressure jets that are located just above the drill head as the drill string is
brought to the surface; the process is illustrated in Figure 12.6. The double-fluid process is capable of pro-
ducing larger-diameter but generally weaker columns compared with the single-fluid process. Jet grouting
can be carried out using comparatively compact plant, and thus it can be used to create continuous under-
ground walls or individual columns to treat the soil below existing structures where liquefaction remediation
is difficult with other methods.

Figure 12.6 Jet grouting process (Courtesy of Hayward Baker)

Jetting, forming
Drilling Jetting test Completion
a column

374

Seismic Design of Foundations

Liquefaction countermeasures

12.3.4 Reinforcement and containment method

When saturated sand deposits are sheared during seismic loading, excess pore water pressure is generated,
reducing the stiffness and strength of the soil and increasing strains. The aim of reinforcement is to reduce
shear deformation in the ground during an earthquake, to mitigate the development of excess pore water
pressures and, at the same time, provide support to the overlying structure. Reinforcement can also be used to
prevent excessive deformation of structures by restraining the lateral ﬂow of the ground after liquefaction.
The principle behind this is illustrated in Figure 12.7.

Reinforcement of the ground can be achieved by underground diaphragm walls, sheet piles or lattice-shaped
walls using deep-mixing techniques or through installation of a grid of stiff columnar elements such as
concrete, timber or steel piles, or soil–cement mixed or jet grout columns. One disadvantage of reinforcement
techniques is that there are no simple methods to verify the effectiveness of the reinforcement to mitigate
shear strains that would cause a substantial reduction in the development of excess pore water pressures.

12.3.5 Drainage
Excess pore water pressure generated by cyclic loading can be reduced and/or dissipated by installing a
permeable drain within the deposit. The methods employed can be divided into two categories, depending on
the material used for the drain: gravel drain and artiﬁcial drain methods. These methods rely on two
mechanisms to reduce damage due to liquefaction

■ delaying the development of excess pore water pressure due to earthquake shaking
■ preventing the migration of high excess pore water pressure from untreated liqueﬁed zones into
non-liqueﬁed areas (e.g. underneath the structure), to prevent secondary liquefaction caused by
pore water pressure redistribution.

Gravel drains are typically installed either as column-like drains in a closely spaced grid pattern or as backﬁll
around underground structures to control the levels of the maximum excess pore water pressure ratio during
earthquake shaking. They can also be installed as wall-like or column-like perimeter drains at both sides of

Figure 12.7 Principle of reinforcement and containment: (a) suppression of shear deformation in the ground
during an earthquake; (b) suppression of the lateral ﬂow of the ground after liquefaction

Structure
(building/dam)

Free-field
Diaphragm
p g wall
Large shear deformation

Suppress
lateral flow
Small shear
deformation

Earthquake
motion

375

Seismic Design of Foundations

Figure 12.8 Process of installing a gravel drain (Courtesy of Fudo Tetra Corporation)

Motorized
1 vibration machine Completed drain

2 4

Gravel

densified (treated) zones or to isolate the high excess pore water pressure from liquefied areas. In the
installation of gravel drains, a casing with an auger inside is drilled into the ground down to the specified
depth. Crushed stone is then discharged into the casing, and the gravel drain is formed by lifting the casing
pipe. This is illustrated in Figure 12.8.

Artiﬁcial drains can be made of geosynthetic composites or piles with drainage functions. They have not only
a drainage effect but also a reinforcing effect. For example, a plastic drain consists of a plastic perforated pipe
drainage skeleton wrapped in geofabric to prevent clogging from soil particles. These can be easily installed;
however, close spacing is usually required due to the limited capacity of each drain.

12.3.6 Increasing in situ stress

12.3.6.1 Increase in the lateral stress
The initial stress state of the soil affects its liquefaction resistance. Increasing the lateral stress in the soil
reduces the initial shear stress in the soil, and thus increases its resistance to liquefaction. Installation of stone
columns, SCPs, compaction piles and compaction grouting may lock in additional lateral stresses to the soils,
reducing the shear stress and its potential to liquefy.

12.3.6.2 Preloading
Preloading involves temporarily loading the soil to consolidate it and increase the lateral stress, thus making it
less susceptible to liquefaction. Preloading is usually carried out by constructing an embankment over the site
and leaving it in place until the soils are consolidated. This method is most suited to silty soils but will also
improve sands. An advantage of this method is the low noise and vibration compared with compaction

376

Seismic Design of Foundations

Liquefaction countermeasures

methods, but a considerable height of preload of soil may be needed for sufficient improvement, and it can be
difficult to fit the side within the boundaries and treat the full footprint. An alternative to the construction of an
embankment is the vacuum consolidation method developed for compressive soils. This method involves
vertical drains associated with the installation of a temporary liner and vacuum pumps. Eighty per cent of
vacuum is commonly achieved, equivalent to 4.0 m of embankment.

12.3.6.3 Permanent dewatering

Lowering the groundwater table makes the soil above the water table unsaturated, resulting in an increased
effective stress for soils below the water table. In addition, if the water table is reduced to a level below the
liqueﬁable soil layer, liquefaction is prevented because the absence of water makes the build-up of excess
pore water pressure impossible (Cox and Grifﬁths, 2010). For this method, it is always necessary to maintain
the lower groundwater level for the entire design life of the structure, and this may result in higher operating
and maintenance costs associated with continuous pumping. Thus, this technique is typically limited to short-
term applications.

12.3.7 Replacement
The replacement method is a liquefaction mitigation technique that involves the removal of the in situ
liquefiable material, and replacement with a non-liquefiable material. Previous case studies have indicated the
grain size distributions of soils that are susceptible to liquefaction; these are indicated in Figure 7.1. It is
apparent from the figure that liquefiable soil is limited to a material with a grain size distribution of a given
range. Clay, well-compacted gravel or soil mixed with cement are non-liquefiable materials and can be used
for replacement in liquefaction remediation.

The advantage of the replacement method is that it provides a high degree of confidence in the final
product and uses construction equipment and practices that are widely available and easily tested (Idriss
and Boulanger, 2008). The method is a beneficial anti-liquefaction measure for areas with a shallow
liquefiable layer, or where vibration and other problems from the remedial work would not be tolerable.
However, this method is only feasible when relatively small volumes of liquefiable material are involved
at shallow depths. The replacement soils can be premixed with a stabilising material such as cement in
advance of placing the treated soil at the site (JGS, 1998). This is recommended for constructing new
reclaimed land or in filling behind caissons, allowing both reclamation work and anti-liquefaction
measures to be implemented at the same time, to create stable ground.

12.4. Analysis and design of typical ground improvement methods

This section presents some of the recommended design practice for the most common methods of ground
improvement. Many of the methods are empirical in nature, and the analysis and design methodologies are
constantly being upgraded as more experience is obtained regarding the method. As such, the practice
outlined for each method is not comprehensive and detailed; rather, emphasis is given to the simpliﬁed
approaches available, with sufﬁcient information for designers to be able to design a particular ground
improvement system. Readers are referred to the references provided for further details.

12.4.1 Densiﬁcation method

As mentioned earlier, the principle of the densification method is to increase the resistance of the ground to
liquefaction by reducing the volume of voids between the particles, making it denser. In addition, there is a
consequent increase in the stress state of the deposit. Figure 12.9 illustrates the principle employed in most
densification methods that involve the installation of replacement materials into the native ground, such as
sand piles, stone columns and compaction grout. In the figure, a unit volume of the sand particles is con-
sidered and e0 and e1 represent the initial and target void ratios, respectively. In effect, the replacement soil

377

Seismic Design of Foundations

Figure 12.9 Principle of soil improvement by the densiﬁcation method

e1 Void
e0

∆e Sand pile

Sand
1
particles

∆e e0 − e1
Replacement ratio as = =
1 + e0 1 + e0

reduces the void ratio of the native soil, leading to densiﬁcation. The amount of soil replacement is expressed
as the replacement ratio as

De e −e
as = = 0 1 (12:1)
1 + e0 1 + e0

Note that the equation indicates that the volume of the void is reduced by forcing fill/replacement material of
volume as into a unit volume of subsoil profile, assuming that the total volume of soil does not change after
densification.

A typical flowchart employed in the design of these densification methods is illustrated in Figure 12.10,
which employs void ratio changes. Although the penetration resistance shown is derived from the standard
penetration test (SPT), other indices, such as the cone penetration test (CPT) qc value or Vs can also be
employed. Typically, the post-improvement or target penetration is derived from liquefaction potential
evaluation and subsequent deformation analysis; that is, depending on the project performance requirement,
the required level of improvement is estimated, which is expressed in terms of, for example, the target SPT
N value. To estimate the initial and target void ratios, available correlations are typically employed; for
example, from the correlation between the SPT N value and the relative density Dr of sand. Once these are
known, Equation 12.1 is used to estimate roughly the amount of fill replacement required.

Knowing the replacement ratio, the spacing required as well as the cross-sectional area of the sand pile
(or stone column) are calculated. Typically, for a given as, many combinations of the installation spacing
and the diameter of sand piles exist. However, there is limit to the volume of fill material that can be
placed on the ground as well as a limit on the densification effect on the sand between the piles. In most
cases, these are controlled by the equipment used in the installation process; hence, the diameter of the
sand pile is typically set. Thus, considering the plane cross-section, the pile spacing can be determined
depending on the layout adopted. Figure 12.11 provides the layout for sand piles (or stone columns, grout
piles, etc.) and the corresponding spacing. In the figure, As and A correspond to the area of the sand pile
and the area of the original ground for one sand pile, respectively, while x, x1 and x2 are the pile spacing.

378

Seismic Design of Foundations

Liquefaction countermeasures

Figure 12.10 Flowchart for evaluating densiﬁcation based on the void ratio

Pre-improvement Post-improvement
(original) SPT N value (target) SPT N value

N–Dr–e relationship

Initial void ratio e0 of Target void ratio e1

the native soil of the improved soil

Estimate the replacement ratio

as required

Implementation of the
Estimate the spacing and cross-
densification method
sectional area of the compacted piles
and recording

Post-improvement verification test

and final decision

Figure 12.11 Typical pile layout and required spacing for each case

Square pattern Trapezoidal pattern

x2
x

x1
As As x
as = = As As
as = =
A x2
A x1 x2

Triangular pattern x
As 2 As
as = =
A 3 × x2

379

Seismic Design of Foundations

In the above procedure, only the effect of volume reduction related to the volume of the replacement material
is taken into account; the compaction effect caused by vibration during installation is not considered. It is
therefore practical to consider that Equation 12.1 is a rough estimate of the required amount, and a post-
improvement veriﬁcation test, using an appropriate in situ testing method, is required to check if the level of
improvement is reached.

Note that the degree of densiﬁcation that can be achieved using this method is affected by many factors,
including (Mitchell and Katti 1981)

■ the soil characteristics (grain size distribution, ﬁnes content, etc.)

■ the degree of saturation and location of the groundwater table
■ the relative density of the soil before densification
■ initial stresses in the soil layer before densification (effective vertical stress, etc.)
■ the skeleton structure and cementation before densification
■ the distance from the source of vibration during densification
■ the properties of fill materials
■ the characteristics of densification methods (types of equipment used, vibration capacity of equipment,
installation method, expertise of the operator, etc.).

Hence, a simple theoretical approach for the design is not easy to implement. The design method adopted is
typically based on case studies and recent test programmes performed at various test sites.

12.4.1.1 SCP method

While the method outlined above requires estimates of the initial and target void ratios in order to determine
the required replacement ratio, current SCP design procedures make use of empirical charts established from
previous work records. These make use of case history data where reliable data on the replacement ratio, the
characteristics of soils, and the SPT N values of the subsoil proﬁle before and after densiﬁcation.

One of the procedures, applicable for loose sands with a fines content Fc < 20%, employs the design charts
shown in Figure 12.12, which make use of the relationship between SPT N values of the original ground and
N values measured between piles (see Figure 12.12(a)) or at the centre of the sand pile (see Figure 12.12(b))
after densification. The required replacement ratio is then read from the charts. In most cases, the N value after
improvement is taken between the piles, at the location of minimum effect. For soils with a higher Fc,
appropriate charts and procedures have been developed. Note that as Fc increases, a high SPT N value is
difficult to attain; hence it may be better to use other methods to improve such soils.

12.4.1.2 Stone column method

Similar to the SCP method, the vibro programme in installing stone columns would be designed to achieve a
certain minimum penetration resistance halfway between the stone columns. Typically, a square or triangular
grid pattern is employed to achieve the desired improvement.

Figure 12.13 shows the plots of the approximate variation of relative density (i.e. penetration resistance of
sands) with the tributary area per compaction probe/stone column (Barksdale and Bachus, 1983) that are used
for preliminary designs. The area replacement ratio is deﬁned as the area of the stone column to the tributary
area per stone column (as indicated in the ﬁgure). Note, however, that the curves in this chart are not a

380

Seismic Design of Foundations

Liquefaction countermeasures

Figure 12.12 SCP design charts applicable to sandy ground with Fc < 20% (Fudo Construction, 1971)

40 40
N1 Np
15
SPT N value between piles after improvement, N1

SPT N value at core of pile after improvement, Np

as = 0.20 0.
15 10 as = 0.20 10
0. 0. 0. 05
0.
30 30
05
0.

0
N
=

0
N
1
N

=
1
N
20 20

as = 0.025–0.075 as = 0.025–0.075
0.075–0.125 0.075–0.125
10 0.125–0.175 10 0.125–0.175
0.175–0.225 0.175–0.225
N0: original N value N0: original N value
N1: N value between piles Np: N value at core of pile
after improvement after improvement
as: replacement ratio as: replacement ratio
0 0
0 10 20 30 0 10 20 30
Original SPT N value, N0 Original SPT N value, N0

(a) (b)

Figure 12.13 Variation of the relative density with the tributary area (Barksdale and Bachus, 1983)

90
Note: 1 ft2 = 0.093 m2
Relative density: %

Ac
Ar =
A
70 Tributary area, A

Stone column area, Ac

50
20 60 100 140
Area per compaction probe: ft2
Silty sand Uniform fine to Well-graded,
(5–15% silt) medium sand (clean) clean sand

381

Seismic Design of Foundations

Figure 12.14 Relation between SPT N values before and after the installation of stone columns (Baez, 1995)

Post-normalised SPT N value: blows/ft

15 Ar = 20%
Ar = 15%
10
Ar = 10%
Ar = 5%
5 Fc < 15%
1 ft ~– 0.3 m
0
0 2 4 6 8 10 12 14 16 18 20
Pre-normalised SPT N value: blows/ft

function of initial conditions. Using case history data, Baez (1995) made improvements, and proposed the
chart shown in Figure 12.14 to estimate the required replacement ratio for known normalised SPT N values
before and after improvement using stone columns.

Similar to the SCP method discussed above, these empirical charts were developed considering densification
criteria only; more refined approaches that consider drainage and stiffness benefits from the installation of
stone columns have also been developed. For example, when considering stress concentration criterion (i.e.
by assuming that the distribution of stresses within the improved zone varies according to the stiffness of the
individual elements, it follows that stresses would be concentrated within the stiffer stone column material),
Baez and Martin (1993) recommended the use of a stress-based approach. In addition, the influence of the
drainage capabilities of stone columns have been studied by Baez (1995), who concluded that the Seed and
Booker (1977) model is useful if the allowable maximum pore pressure ratios are maintained below 0.6.
Priebe and Grundbau (1998) proposed design charts for the application of stone columns to mitigate
earthquake-induced liquefaction based on the understanding that the loads taken by the stone columns from
the soil do not contribute to liquefaction.

12.4.2 Dynamic compaction

The dynamic compaction technique involves high-energy impacts to the ground surface by systematically
dropping heavy weights of 5–35 t from heights ranging from 10 to 40 m to compact the underlying ground
using heavy crawler cranes (Lukas, 1995). Sand deposits densiﬁed by dynamic compaction are more resistant
to liquefaction, and have performed well during earthquakes. On the other hand, silty sand deposits appear to
perform well when improved by this technique supplemented with wick drains (Dise et al., 1994).

For both sands and silty sand sites, there are at present no detailed analytical procedures available to analyse
the effects of field dynamic compaction operational parameters and soil conditions to determine the densi-
fication and the degree of improvement achievable in the field. Current practice for liquefaction screening,
mitigation design, suitability assessment, and the determination of optimum field operation parameters for
dynamic compaction rely mainly on field pilot tests, and past experience based on case histories (Baez, 1995;
Lukas, 1995).

382

Seismic Design of Foundations

Liquefaction countermeasures

It is known that the level of compaction required to improve the ground to resist liquefaction for a given
earthquake (in terms of the impact energy, the spacing between impacts, the number of repeated impacts per
location and the time cycle between impacts) depends on the initial density of the soil, compressibility,
hydraulic conductivity (and silt content), and provisions available for drainage of excess pore pressures
induced during ground improvement works. Thevanayagam et al. (2006) made use of this framework to
develop energy-based liquefaction screening guidelines, and design methods for ground improvement to
resist liquefaction.

In the development of the guidelines, Thevanayagam et al. (2006) conducted numerical simulations to obtain
the relationship between pre- and post-improvement densities for various uniform soil sites containing clean
sands to non-plastic silty soils supplemented with or without wick drains. For all the simulations, the impact
grid pattern was assumed to be as shown in Figure 12.15. It involved three phases of impact (primary,
secondary and tertiary) at the grid locations shown in the ﬁgure. The energy per impact WH, the impact grid
spacing S, the total number of impacts per grid point during each phase NI, the wick drain spacing Sw, the
wick drain size and the time cycle between impacts T were varied for each simulation. The groundwater level
was assumed to be at 2 m depth from the impact surface.

Two sets of results are presented in Figure 12.16. For these examples, the size of the wick drains was assumed
to be 100 mm × 5 mm with an equivalent diameter of 5 cm pre-installed at 1.5 m spacing in a rectangular
pattern. The number of impacts per grid location and the time cycle between impacts were set at NI = 12 and
T = 2 min, respectively. Figure 12.16(a) shows the pre- and post-improvement (N1)60,cs (the normalised SPT
N value for clean sand) profiles for two uniform soil deposits with pre-improvement (N1)60,cs = 7.5 and 16,
respectively, and an impact grid spacing of S = 15 m. The first two figures are for soils with a coefficient of
permeability k of 10–7 m/s, and the last two figures are for k = 10–8 m/s. Each curve in these figures refers to
the pre-improvement profile and post-improvement profile for a different energy per impact WH of 100, 250,
500 and 750 t m. Figure 12.16(b) is for an impact grid spacing of S = 12 m for the same WH values. Additional
design charts are presented by Nashed (2005).

Figure 12.15 Impact grid pattern adopted for dynamic compaction (Thevanayagam et al., 2006)

S
Primary phase
Secondary phase
Tertiary phase
Wick drain

383

Seismic Design of Foundations

Figure 12.16 Pre- and post-improvement (N1)60cs values for dynamic compaction: impact grid spacing of
(a) S = 15 m (WH = 100–750 Mg-m); and (b) S = 12 m (WH = 100–500 Mg-m) (Thevanayagam et al., 2006)

(N1)60,cs (N1)60,cs (N1)60,cs (N1)60,cs

0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40
0 0 0 0
2 2 2 2
4 4 4 4
Depth: m

Depth: m

Depth: m
6 6 6 6
8 8 8 8
10 10 10 10
12 12 12 12
14 14 14 14
16 16 16 16
k = 10–7 m/s, pre-(N1)60,cs = 7.5 k = 10–7 m/s, pre-(N1)60,cs = 16 k = 10–8 m/s, pre-(N1)60,cs = 7.5 k = 10–8 m/s, pre-(N1)60,cs = 16

Pre Post 750 Post 500 Post 260 Post 100

(a)
(N1)60,cs (N1)60,cs (N1)60,cs (N1)60,cs
0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40
0 0 0 0
2 2 2 2
4 4 4 4
Depth: m

Depth: m
Depth: m

(b)

12.4.3 Vertical drains

Vertical drains operate by providing preferential drainage paths that enable accumulating pore pressures to
dissipate ideally before the surrounding soil reaches a state of initial liquefaction. It is common to stipulate a
maximum excess pore pressure ratio ru of 0.25–0.5 (Brennan and Madabhushi, 2002), and Iai and Koizumi
(1986) recommend a maximum ru of 0.5.

The most common design criteria used in the design of vertical drains is the chart-based approach of Seed and
Booker (1977), shown in Figure 12.17. In the ﬁgure, 2a and 2b are the drain diameter and centre-to-centre
spacing between drains, respectively, Neq is the equivalent number of cycles, Nl is number of cycles to
liquefaction and Td is a non-dimensional time factor

k td
Td = (12:2)
mv gw a2

384

Seismic Design of Foundations

Liquefaction countermeasures

Figure 12.17 Design charts for vertical drains (Seed and Booker, 1977; reproduced with permission from the
publisher, ASCE)

1.0 Neq Neq

=1 =2
Nl Nl
0.8
ktd
Td =
0.6 mvγw a2

2
Average maximum pore pressure ratio, (umax/σv)ave

5
10
5

0.4
10

25 25

50
10
50
10

0.2
0
0

Td = 200 Td = 200
0
1.0 Neq Neq
=3 =4
Nl Nl
0.8
2
2

5
0.6
10
5

25
10

0.4
50
25

10
50

0
10

0.2
0

Td = 200 Td = 200
0
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
a/b a/b

where k is the coefﬁcient of permeability of the soil, mv is the coefﬁcient of volumetric compressibility of the
soil, td is the duration of earthquake shaking and gw is the unit weight of water (9.81 kN/m3). The analysis is
based on separating pore pressure generation and dissipation according to the following equation, with a
suitable term used to model pore pressure generation.

∂u ∂u
= C ∇2 u + (12:3)
∂t ∂t general

385

Seismic Design of Foundations

where u represents the excess pore pressure and C is a coefﬁcient of consolidation in the appropriate direction.
Equation 12.3 is commonly solved according to a time-marching ﬁnite-difference scheme. An achievable
drain size is selected, then an inter-drain separation distance chosen to prevent the excess pore pressure
reaching a certain level during the design event. If dissipation balances generation, equilibrium is maintained
until the generation stops.

Much follow-up work (as reviewed by Boulanger et al., 1998) has highlighted a major shortcoming of Seed
and Booker’s charts, in showing that the permeability of a drain is a major inﬂuence. Their assumption of
inﬁnitely permeable drains is overly simplistic for design purposes, and designers must update design pro-
cedures, at least by using charts that consider this parameter. For example, the charts developed by Onoue
(1988), which takes into account well resistance, are shown in Figure 12.18. The well resistance Lw is cal-
culated as
2
32 k H
Lw = (12:4)
π2 kw dw

where k and kw are the permeability coefﬁcients of the natural sand and drain, respectively, while H is the
height of the drain and dw is the diameter of the drain. In Figure 12.18, de is the centre-to-centre spacing
between drains and ru max is the overall average value of the maximum pore pressure.

In the design of drains, it is necessary to select a suitable drain material that has a coefﬁcient of permeability
substantially larger than the in situ soils. Based on studies done by various researchers, the Dutch Port and
Harbour Research Institute (1997) recommends that any of the following criteria may be used in the selection
of materials for a drain

■ d15/D85 < 5
■ d25/D75 < 9.5 or d15/D85 < 6.4
■ d15/D85 < 9

where d15 is 15% of the diameter of the drain material (mm) and D85 is 85% of the diameter of natural soil
(similar deﬁnitions for d25 and D75).

12.5. Worked examples for typical methods

12.5.1 Densiﬁcation method
12.5.1.1 Example
A 3 m-thick well-graded clean sand layer is overlain by a 3 m-thick clayey silt layer. It has an average
normalised SPT N value of (N1)60 = 15. The groundwater table is located at a depth of 1.2 m. Liquefaction
evaluation using the NCEER method (Youd et al. 2001; see also Chapter 7) showed that for a design
earthquake moment magnitude Mw of 7.5 and a peak ground acceleration (PGA) of 0.3g, the sand layer will
liquefy. In order for it not to liquefy (i.e. liquefaction safety factor FL = 1.0), the SPT N value should be
increased to a level equal to (N1)60 = 26.

386

Seismic Design of Foundations

Liquefaction countermeasures

Figure 12.18 Relationships between well resistance and the spacing ratio for vertical drains (Onoue, 1988;
reprinted with permission from the author)

de
dw
re
rw
Permeable r
1

∂u = 0
∂z Z
k
Kw
H

∂u = 0
∂z

Impermeable

Td = 2 Td = 10 Td = 50
Td = 5 Td = 25 Td =100 Td = 200 Td = 2 5 10 25 50 100 200
0.6 0.6
Spacing ratio, rs (=dw/de)

Spacing ratio, rs (=dw/de)

400
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2 ru max
ru max 0.3
0.1 0.3 0.1 4 ktd
0.4
0.5
4 ktd 0.4 Td =
rN (=Neq/Nl) = 1, Td = m γ d 2 0.5 rN (=Neq/Nl) = 3, mvγ wdw
2 0.6
v w w 0.6
0.1 0.2 0.5 1 2 5 10 20 50 0.1 0.2 0.5 1 2 5 10 20 50
Coefficient of well resistance, Lw Coefficient of well resistance, Lw

Td = 2 5 10 25 50 100 Td = 2 5 10 25 50 100 200

0.6 0.6
400
200
Spacing ratio, rs (=dw/de)

Spacing ratio, rs (=dw/de)

0.5 0.5
400

0.4 0.4
0.3 0.3

0.2 0.2 ru max

ru max 0.3
0.3 0.4
0.1 0.4 0.1 0.5
4 ktd 0.5 4 ktd 0.6
rN (=Neq/Nl) = 2, Td = rN (=Neq/Nl) = 4, Td = m γ d 2
mvγ wdw2 0.6 v w w

0.1 0.2 0.5 1 2 5 10 20 50 0.1 0.2 0.5 1 2 5 10 20 50

Coefficient of well resistance, Lw Coefficient of well resistance, Lw

387

Seismic Design of Foundations

12.5.1.2 Requirement
Determine the area replacement ratio, and the spacing of

(a) SCPs
(b) stone columns.

12.5.1.3 Solution

(a) If SCPs are used, the required replacement ratio as is determined from Figure 12.12(a). While the
SPT N values indicated are not normalised with respect to the overburden pressure, it is possible to
use this chart as long as they represent the same depth. Hence, for original N0 = 15 and improved
N1 = 26, the required as is read from the chart as 0.14 or 14%. Typically, the diameter of an SCP
is set by the equipment and construction, say 0.70 m. Assuming a square grid pattern, then from
Figure 12.11

As As ð π / 4Þð0:7 mÞ2
as = = = = 0:14 x = 1:6 m (12:5)
A x2 x2
(b) If a stone column is used, the required replacement ratio as is determined from the chart of Baez (1985)
shown in Figure 12.14. For pre-improvement normalised (N1)60 = 15 and post-improvement normalised
(N1)60 = 26, the required as is read from the chart as 6%. On the other hand, if Figure 12.13 is used, the
post-improvement relative density is required. Using the correlation proposed by Tokimatsu and Seed
(1987), (N1)60 = 26 corresponds to Dr = 75%. From the chart, this corresponds to as = 6.2% – 8.4%, or
say ≈7%. Assuming a triangular grid pattern and a stone column diameter of 0.76 m, then

As 2 A 2 ð π/ 4Þð0:76 mÞ2
as = = pffiffiffi 2s 0:07 = pffiffiffi x = 2:7 m (12:6)
A 3 x 3 x2

Generally, larger replacement ratios are required in the SCP method to provide the same level of improvement
as the stone column method.

12.5.2 Gravel drain design

12.5.2.1 Example
Consider the same clean sand layer discussed in Section 12.5.1. Additional data for the clean sand include the
following: permeability, k = 1.0 × 10–5 m/s; coefﬁcient of volumetric compressibility, mv = 2.9 × 10–5 kN/m2;
and effective stress at the centre of the layer, s 0v = 52.4 kPa. Additional earthquake parameters are as follows:
duration of shaking, td = 40 s; number of representative cycles, N = 15 (corresponding to an Mw = 7.5
earthquake); and the number of cycles to liquefaction, Nl = 5 cycles (typically read from the liquefaction
resistance curve from undrained cyclic triaxial tests).

12.5.2.2 Requirement
Assuming that the gravel drain has permeability k = 0.4 m/s, and height H = 3 m, determine the drain spacing
required using

(a) the Seed–Booker approach

(b) the Onoue approach.

388

Seismic Design of Foundations

Liquefaction countermeasures

12.5.2.3 Solution
Considering a gravel drain diameter of 0.76 m, and the input earthquake duration, the number of represen-
tative cycles, the initial effective stress, the number of cycles to liquefaction, the soil permeability and the
coefﬁcient of volume compressibility, the following are obtained

k td 1 10−5 m/ s 40 s
Td = = ≈ 10 (12:7)
mv gw a2
2:9 10−5 m2 / kN (9:81 kN/ m3 )(0:76 m/ 2)2

Neq 15
= =3 (12:8)
Nl 5

Now, set the limiting maximum pore water pressure ratio between the drains to be ru = 0.5 (Iai and Koizumi,
1986).

(a) Using the Seed–Booker approach (assuming the drain has inﬁnite permeability): reading from
Figure 12.17 (i.e. for Neq/Nl = 3, Td = 10), a/b = 0.36, from which b = 1.05 m. Thus, the required
(centre-to-centre) spacing is 2b = 2.1 m.
(b) Using the Onoue approach (which considers well resistance)
2
32 k H 32 1 10−5 m/ s 3m 2
Lw = = = 1:2 10−3 (12:9)
π2 kw dw π2 0:4 m/ s 0:76 m

Reading from Figure 12.18 (i.e. for Neq/Nl = 3, Td ≈ 10, the spacing ratio dw/de = 0.36, so de = 2.1 m.

Thus, it would appear that the required drain spacing corresponds to that from the Seed–Booker approach,
even if well resistance is considered. This is because the effective height of the drain is quite short, and
therefore the pore water pressure generated by the earthquake would easily be dissipated, almost similar to
the assumption of an inﬁnitely permeable drain.

12.6. Case studies: ground improvement method application

12.6.1 Haneda Airport (Tokyo, Japan)
An example application of compaction grouting in an open unrestricted space was discussed by Zen et al.
(2000, 2001). In order to improve the seismic resistance of the foundation ground beneath the new Runway B
of the Tokyo International Airport in Haneda (Tokyo, Japan), the SCP method was used to densify the
ground. To prevent any disturbance to the airport operation brought about by the large equipment, the ground
treatment was performed from night to dawn (23:00–5:00) when the airport was not operational. However,
since the SCP method requires large holes, which would be difficult to restore easily in time for the
resumption of flight operation the following morning, a special method was necessary to treat the ground at
the intersection of the new Runway B with the existing Runway A without disrupting air traffic. For this
purpose, compaction grouting was implemented at the said intersection.

389

Seismic Design of Foundations

Initially, a test programme was performed as part of the initial production grouting programme to evaluate the
effectiveness of the grouting and to reﬁne the guidelines for production. In the test programme, four cases
were investigated:

■ case 1: replacement ratio as = 10%, bottom-up approach

■ case 2: as = 15%, bottom-up approach
■ case 3: as = 20%, bottom-up approach
■ case 4: as = 15%, top-down (ground level (GL) – 5 m and above) and bottom-up (GL – 5 m and
below) approaches.

Note that the replacement ratio as, sometimes referred to as the grout take, is the volume of grout injected per
unit volume of in situ soil. In all cases the grout columns were arranged in triangular patterns with 1.5–1.7 m
pitch. For the test programme, the target depth was GL – 2.8–11.8 m (total depth = 9.0 m), and the average
SPT N value within the target zone was less than 15, as shown in Figure 12.19(a). The in situ soil had a ﬁnes
content Fc of between 10 and 35%. SPTs and CPTs (using a radioisotope cone penetrometer) were performed.
The results of the SPTs and CPTs, which were conducted primarily at treatment grid centres, are shown in
Figure 12.20, where the relations between the SPT N value and the CPT tip resistance qc before and after the

Figure 12.19 Typical soil proﬁles and vertical distributions of SPT N values: (a) test programme (case 2); and (b) main
programme (site 15) (Zen et al., 2001)

Treatment SPT N value Target SPT N value

GL: m depth 0 10 20 30 40 50 GL: m depth 0 10 20 30 40 50
0 0
Road
base
3.46
as = 12%

–5 Bs
as = 15%

As0
–5
As0 Ac1
as = 15%

(sand) –10
as = 16%

As1

–15 Ac2

Ac1
as = 9%

–10 Acs2
(clay)

Ac3

–20

(a) (b)

390

Seismic Design of Foundations

Liquefaction countermeasures

Figure 12.20 Relation between (a) the SPT N values; and (b) the CPT tip resistance qc before and after treatment
(Zen et al., 2000)

40 300

CPT tip resistance after treatment qc: MN/m2

35
250
30 Case 3
SPT N value after treatment

200
25
Case 2
20 Case 1 150
Case 2
Test
15 programme Case 3
Case 4 100
Borehole A
10 Borehole B
Case 1 (as = 10%)
Main Case 1 Case 2 (as = 15%)
No. 31 50
programme Case 3 (as = 20%)
5 No. 32
No. 33
No. 34
0 0
0 5 10 15 20 25 30 35 40 0 50 100 150 200 250 300
SPT N value before treatment CPT tip resistance before treatment qc: MN/m2
(a) (b)

treatment are presented. It can be seen from Figure 12.20(a) that the SPT N values increased in all cases.
Furthermore, qc after treatment generally showed an increase in value, with the effect more predominant with
higher replacement ratios, as shown in Figure 12.20(b).

After confirming the effectiveness of the compaction grouting method in densifying loose liquefiable soils,
the main compaction programme was implemented for the target site itself. Using the refined guidelines for
production, replacement ratios as = 12%, 15%, 16% and 9% were selected and applied at appropriate depths,
since the SPT N values varied from layer to layer. An example of the soil profile at site 15 and the corre-
sponding as for each depth are shown in Figure 12.19(b). The average target depth for implementation within
the whole area was the liquefiable layer between GL – 3.0 and 18.4 m (total depth = 10 m). In the main
grouting programme, a top-down bottom-up approach was employed.

A comparison of the SPT N values recorded before and after the grouting implementation is shown in Figure
12.20(a), which includes results for both the main grouting programme (boreholes A and B and sites 31–34)
and the test programme (cases 1–4). It can be seen that the SPT N values increased for all sites, indicating an
increase in the relative densities of the ground. Similar trend was observed for the CPT test results, as
illustrated in Figure 12.20(b).

In addition, in-hole lateral load tests using self-boring-type equipment were performed at five sites during the
test programme. These tests include case 2 before treatment and cases 1–4 after treatment. In addition, lateral
load test was performed at the case 2 site 1.5 years after grouting implementation and at site 2 immediately
after the main grout programme. The test results are plotted in Figure 12.21, which shows the relation
between SPT N values and coefficient of earth pressure at rest K0, defined as the ratio between the horizontal

391

Seismic Design of Foundations

Figure 12.21 Relation between SPT N value and earth pressure coefﬁcient, K0 (after Zen et al., 2001)

3.5
Before (case 2)
After (case 1, as = 10%)
3.0 After (case 2, as = 15%)
After (case 3, as = 20%)
After (case 4, as = 15%)
Earth pressure coefficient, K0 2.5 + 1.5 years after (case 2, as = 15%)
After (No. 2, as = 16%)

2.0

+
1.5
+ +
+
1.0

0.5

0
0 10 20 30 40
SPT N value

effective stress s0h to the effective vertical stress s0v . It can be seen from the ﬁgure that prior to grouting
implementation, K0 values were about 0.5; however, after grouting, K0 values increased to 0.6–2.9.
Speciﬁcally, the results of tests for case 2 conducted 1.5 years after grouting implementation showed that
the K0 value remained higher than 1.0.

As far as the liquefaction potential of the site was concerned, evaluation procedures were performed based on
the technical standards for port and harbour facilities in Japan (Port and Harbour Research Institute, 1997). In
this procedure, seismic response analyses were performed, and the zone to which a soil layer belongs to is
determined from the equivalent SPT N value and the equivalent acceleration. A check is made on whether the
data points plot within zone IV, corresponding to the zone where liquefaction will not occur. Figure 12.22(a)
plots the relation between the equivalent N value and equivalent acceleration. In the seismic response
analyses, two types of baserock input motions were employed, namely the Hachinohe record (H wave) and
the Oofunato record (O wave), two strong motion recordings obtained at separate port facilities in Japan
during past earthquakes. The amplitude of the peak acceleration was set to 390 cm/s2. It can be seen from
Figure 12.22(a) that although most of the data points plot in zone IV (70 from test programme and 119 from
main programme), some data points lie within zone III (four from the test programme and 23 from the main
programme). According to the technical standards, the liquefaction potential in zones II and III can be
determined by conducting further evaluations based on cyclic triaxial tests. For this purpose, undisturbed soil
samples were obtained at three test sites (cases 1–3) and two main sites (Nos. 1-1 and 2) before and after the
grouting operation, and cyclic triaxial tests were performed. Some data points obtained from the test pro-
gramme that lie within zone IV were also included in the evaluation. Based on the test results, the factor of
safety against liquefaction FL was evaluated, and the results are shown in Figure 12.22(b). In the evaluation,

392

Seismic Design of Foundations

Liquefaction countermeasures

Figure 12.22 (a) Relation between the equivalent acceleration and the equivalent SPT N value; and (b) comparison
between the factor of safety against liquefaction FL before and after treatment (Zen et al., 2001). Note: H wave:
Hachinohe Record; O wave: Oofunato record

50 0
Test (H wave)
45 Test (O wave) 2
No. 31 (H wave)
40 No. 31 (O wave) 4 +
Equivalent SPT N value

No. 32 (H wave)
35 ×× 6
No. 32 (O wave)
No. 33 (H wave) 8
30 + –

Depth: m
× ×+× × No. 33 (O wave) Liquefaction No liquefaction
25 ×– × +– No. 34 (H wave) 10
–× +–
No. 34 (O wave) ×
20 IV 12
× No. A (H wave)
+ – ×
No. A (O wave)
15 III + – No. B (H wave)
14 Before (H wave) Before (O wave) Case 2

II Test After (H wave) After (O wave) Case 1

10 + No. B (O wave) 16 programme After (H wave) After (O wave) Case 2
I
After (H wave) After (O wave) Case 3
5 18 Main + After (H wave) – After (O wave) No. 1-1
programme
After (H wave) × After (O wave) No. 2
0 20
0 100 200 300 400 500 600 0 0.5 1 1.5 2
Equivalent acceleration: cm/s2 Factor of safety against liquefaction, FL
(a) (b)

K0 = 0.5 was set for data points before grouting, while K0 = 1.0 was set for data points after grouting, as
suggested by Figure 12.21. Before treatment, most of the sites are liqueﬁable (FL < 1.0), as noted in Figure
12.22. However, after treatment, it can be seen that, except for three data points, liquefaction will not occur
within the foundation ground.

The compaction grouting programme was implemented in a section of Tokyo International Airport in Haneda
with a total area of 4500 m2. Over all, the programme involved 1792 grout columns, requiring a total of
6700 m3 of grout materials. During the grouting programme, the ﬂight operation in the airport was not
disturbed, since grouting was conducted at night when the airport was not operational. Furthermore, the grout
holes were only 7 cm in diameter and could easily be covered, so the holes did not hinder runway operation
during the resumption of ﬂights.

Following the 2011 Tō hoku earthquake, the reclaimed lands in Tokyo were subjected to ground accelerations
with a peak value of about 0.16g lasting for more than 3 min. Clearly, liquefaction did not occur at the airport,
and the earthquake barely affected airport operation, with ﬂights resuming the very next day after the
earthquake (Orense et al., 2012).

12.6.2 Meriken Park Hotel (Kobe, Japan)

Many areas, especially reclaimed land on the waterfront, suffered extensive liquefaction during the 1995
Hyogoken Nambu (Kobe) earthquake. There were a few buildings in the area for which liquefaction remedial
measures had been implemented prior to the earthquake. One case is the 14-storey Meriken Park Hotel
building on a pier in Chuo-ku, Kobe City. The soil proﬁle at this location consists of 10–12 m of soft

393

Seismic Design of Foundations

reclaimed sand and gravel layers over the seabed, which in turn consists of alternating layers of clay, sand and
gravel. The building, which was under construction when the earthquake occurred, was 60 m high and had a
footprint measuring 134 m × 59 m. It was supported on cast-in-place concrete piles 2.5 m in diameter and
33 m long (Namikawa et al., 2007; Tokimatsu et al. 1996).

As a liquefaction countermeasure, the foundation ground was improved using the deep cement mixing
(DCM) method. The cement-treated soil walls were produced by mixing the original soil with cement slurry
in overlapping columns with a diameter of 1 m. The treatment ratio (defined as the ratio of the improved soil
area to the whole area) was approximately 20%. The unconfined compressive strength of the improved soil
samples after curing for about 6 weeks was 4–6 MPa (Suzuki et al. 1996). The section and plan diagrams are
shown in Figure 12.23. The lattice-shaped ground improvement was applied because of the potential for
liquefaction of the upper loose fill material with a thickness of 10–12 m. The depth of ground improvement
that surrounds the cast-in-place piles extends to a depth of about 16 m.

Figure 12.23 Section and plan view of the superstructure and improved soil grid (Namikawa et al., 2007; reprinted
with permission from the author)

Liquefiable
sand
GL 0.0: m
GL – 1.2: m 1FL (GL + 0.7: m)

GL – 12: m

Seabed soil
GL – 30: m

Improved ground
Section view Pile
7.2 6.4 7.2
59.2: m
17.6

P5
7.2 6.4 7.2

16.0 6.4 9.6 6.4 9.6 6.4 9.6 6.4 9.6 6.4 9.6 6.4 9.6 6.4 9.6 6.4
134.4: m

Plan view Unit: m

394

Seismic Design of Foundations

Liquefaction countermeasures

Figure 12.24 View of the lattice-shaped DCM wall (Courtesy of Takenaka Corporation)

During the earthquake, the quay walls on the west, south and east of the building moved laterally by 1, 2 and
0.5–0.6 m, respectively; they also settled by 0.4–0.6, 0.5–0.7 and 0.2–0.3 m, respectively (Tokimatsu et al.,
1996). However, Namikawa et al. (2007) reported that there was no evidence of liquefaction occurring at the
unimproved soil within the improved soil grids. Moreover, the head of the cast-in-place pile supporting the
building was intact. Needless to say, the building survived the earthquake without damage to its pile foun-
dations and without undergoing differential settlement. Such good performance can be attributed to the DCM
walls that surrounded the piles, as shown in Figure 12.24. The DCM walls were able to withstand the lateral
forces from both the superstructure and liqueﬁed soils during the earthquake.

12.7. Case studies: seismic performance of improved-ground sites

Events as far back as the 1964 Niigata earthquake (Japan) and, more recently, the devastating 2011 East Japan
earthquake indicate that improved sites suffer less ground deformation and subsidence than unimproved
areas. Hausler and Sitar (2001) compiled over 90 case studies on the performance of improved sites from 14
earthquakes in Japan, Taiwan, Turkey and the USA covering a wide range of improvement methods. The
field case studies encompass a wide range of improvement methods, from conventional densification methods
such as SCPs to less common lateral restraint-based methods such as sheet pile walls or deep soil mixing
grids. The data collected are summarised in Table 12.2, which indicate that improved sites generally per-
formed well. However, about 10% of the sites required significant post-earthquake remediation, repair or
demolition. Unacceptable performance of ground improvement resulted from excessive ground deformation
due to severe lateral spreading or because of an insufficient depth of remediation.

In addition to the seismic performance of improved ground sites discussed in the previous section, case
studies of the performance of other ground improved methods during past severe earthquakes are discussed
below.

395

Seismic Design of Foundations

Table 12.2 Performance of various ground improvement methods (Hausler and Sitar, 2001)

Method Performance Average increase in (N1)60

(acceptable/unacceptable)

Densiﬁcation through vibration and compaction

Sand compaction piles 26/5 11
Deep dynamic compaction 15/0 5
Vibro-rod/vibroﬂotation 11/6 13
Stone columns 7/1 8
Preloading 5/0 5
Compaction grouting 1/1 —
Timber displacement piles 1/0 —

Dissipation of excess pore water pressure

Gravel drains 5/0 7
Sand drains 5/0 9
Wick or paper drains 2/0 —

Restraining effect through inclusions

Deep soil mixing 4/1 —
Diaphragm walls 0/1 —

Stiffening through chemical or cement addition

Jet grouting 5/0 —
Chemical grouting 1/0 —

—, not applicable

12.7.1 Performance of SCP-improved ground

The SCP method is one of the most popular densiﬁcation methods in Japan. For sandy ground, the principle
of the method is to reduce the void ratio by injecting well-compacted sand piles through hollow steel casings
driven into the ground. The SCP method can be implemented in two ways: vibratory and non-vibratory SCP.
The two methods are applicable both onshore and offshore. The total sand pile length accumulated up to the
present time has reached 400 000 m using the vibratory method and 10 000 m by the non-vibratory method
(Harada et al., 2015).

The effectiveness of compaction by the SCP method has been confirmed by damage investigations of past
earthquakes, and many case studies have been reported. Figure 12.25 shows the epicentre locations and
characteristics of the 1964 Niigata earthquake and nine other large-scale earthquakes, including the 2011
Tō hoku earthquake off the Pacific Coast, as well as information on the performance of SCP-improved ground
during these earthquakes. As shown in the figure, there have been no reports of major disruption to structures

396

Seismic Design of Foundations

Seismic Design of Foundations
Figure 12.25 Case studies of the performance of the SCP method in past earthquakes (Modiﬁed from Ohbayashi et al., 1998)

6 Hokkaido Toho-oki Earthquake (October 1994; Mw = 8.1)

5 Hokkaido Nansei-oki Earthquake (July 1993; Mw = 7.8)
Effectiveness confirmed of SCP method at site
Effectiveness of SCP improvement for storage tank site at where it was adopted for restoration work after
Hakodate confirmed (Fudo, 1993) the 1993 Kushiro-oki earthquake (Nishikawa et al., 1995)
3 Nihonkai Chubu Earthquake (May 1983; Mw = 7.7) 4 Kushiro-oki Earthquake (January 1993; Mw = 7.8)
Effectiveness of SCP improvement for Effectiveness confirmed of SCP and gravel drain
Okitate storage tank site confirmed (JSCE, 1986) method used for Kushiro's West Harbor (Iai et al., 1993);
1 Niigata Earthquake (June 1964; Mw = 7.5) SCP used in restoration work of the Kushiro and
Tokachi Rivers (Sasaki et al., 1995)
Effectiveness of vibrofloatation for tank
at Ohse confirmed (Watanabe, 1966)
11 Nemuro-Hando-oki Earthquake (November 2004; Mw = 7.1)
8 Hyogo-ken Nambu Earthquake (January 1995; Mw = 7.2)
10 Tokachi-oki Earthquake (September 2003; Mw = 8.0)
Earthquake-resistant effect of ground
Effectiveness of vibratory SCP improvement for quaywall
improvement confirmed (Yasuda et al., 1996)
at Hanasaki Port confirmed (Iida et al., 2005)
7 Sanriku Haruka-oki Earthquake (December 1994; Mw = 7.5)
Effectiveness of SCP improvement confirmed
at Hachinohe storage tank site (JGS, 1996)
12 Off the Pacific Coast of Tohoku Earthquake (March 2011; Mw = 9.0)
Effectiveness of SCP vibratory and non-vibratory SCP improvement for
residential area at Urayasu and Koto-ku confirmed (Yasuda et al., 2012)

2 Miyagi-ken-oki Earthquake (June 1978; Mw = 7.4)

9 Tottori-ken Seibu Earthquake (October 2000; Mw = 7.3) Effectiveness of SCP improvement confirmed at storage tank site near
Effectiveness of non-vibratory SCP improvement for Ishinomaki fishing port (Ishihara et al., 1980)
embankment at Abe Sanryu line confirmed (Fudo, 2000)
Liquefaction countermeasures

397
Seismic Design of Foundations

erected on compacted ground, such as by the vibratory and non-vibratory SCP methods, thus conﬁrming in a
qualitative way the effectiveness of compaction-type ground improvement techniques. The detailed infor-
mation for each case is described in the respective reference indicated. The tamper method was adopted for
restoration work on road foundations after the 1983 Nihonkai Chubu earthquake.

Some examples of SCP application in Japan, and their performance after major earthquakes are discussed
below (Kinoshita et al., 2012).

12.7.1.1 Improvement effectiveness in port facility structures (1993 Kushiro-oki

earthquake)
Figure 12.26 shows a standard section of the Kushiro West Port that was affected by the 1993 Kushiro-oki
earthquake (Yamada and Saito, 1990). At this location, countermeasures to resist earthquakes, mostly
compaction by the vibratory SCP method, were implemented in the ground behind the quay walls. In areas
adjacent to steel structures, the gravel drain method was adopted to avoid any negative effects of vibration or
displacement from the improvement work. SPT N values in areas between sand piles in the compaction-
improved areas were around 20–30. The maximum horizontal acceleration recorded at the Kushiro Port
Construction Ofﬁce, located about 1.5 km from the quay walls, was 0.48g, but no damage due to the
earthquake was observed, and port activities resumed the day following the earthquake. However, in
unimproved ground within the same wharf area, large cracks appeared in quay walls (approximately 10 cm
wide with 20 cm vertical offset), and the quays could not be used.

Figure 12.26 Standard section of revetment at Kushiro West Port (Courtesy of Fudo Tetra Corporation)

30.0
Mooring post
20.0
Curb
1.2 Gravel mat
Fender
t = 0.3
+3.0
HWL +1.5 +1.8 +1.5 +2.2 +1.5
HWL
l = 8.0: m

(highest water level)

Rubble −6.0
Natural
ground −7.5
Steel sheet pliles

−7.5
Steel sheet pile SKY-41 t = 16, l = 24.5

Type IV

−12.6
1.4
−14.6 4 @ 1.4 = 5.6 4 @ 1.7 = 6.8
(φ 400) (φ 700)

– 22.7 Gravel drains Sand compaction piles

Sand compaction piles

398

Seismic Design of Foundations

Seismic Design of Foundations
Figure 12.27 Standard section of river embankment along the Kushiro River (Courtesy of Fudo Tetra Corporation)

Point of post-construction
investigation

1.5: m
Sand compaction piles
e φ 0.7
lin
re
nt Elevation: m 10
Ce 1.5: m
Mat t = 1.0: m

2.000 2.000
5
+3.5 1:2
1:2
+2.5

–2.5 –2.5

–5.0 –5

52.5: m
–10
Liquefaction countermeasures

399
Seismic Design of Foundations

12.7.1.2 Improvement effectiveness in river embankments (1993 Kushiro-oki earthquake)

The 1993 Kushiro-oki earthquake caused serious damage to many parts of eastern Hokkaido. River
embankments, particularly those along the Kushiro River, suffered damage that included lateral cracking,
cross-cracking, slope collapse and cave-ins. The Kushiro River embankments suffered intermittent collapse
over a section of several hundred metres, attributed to liquefaction in the alluvial sand layer and in the
embankment itself below groundwater level.

In the restoration work, the vibratory SCP method was adopted for the ﬁrst time for foundation ground
improvement in the restoration of river embankments. Figure 12.27 shows a standard section (Sasaki et al.,
1995). One year and nine months after the restoration work was completed, the Hokkaido Toho-oki earth-
quake occurred, and again the region suffered massive seismic impact. At the locations where the SCP
method had been used there was no damage, but cracking reoccurred at locations where the embankment had
been restored after the Kushiro-oki earthquake only by compact-tamping, because the cracking at that time
had been light.

12.7.1.3 Improvement effectiveness in a storage tank facility (1993 Hokkaido Nansei-oki

earthquake)
In the 1993 Hokkaido Nansei-oki earthquake a remarkable difference was observed between the areas of
improved ground, including the vibratory SCP-improved tank base ground, and the adjacent unimproved
areas. Figure 12.28 shows the locations of sand boil marks in the improved tank base ground and the adjacent
non-improved ground areas. The tanks themselves, installed on ground that had been improved against
liquefaction using the SCP method, did not suffer damage, but evidence of sand boiling could be observed at a
distance of approximately half of the improvement depth from the improved section.

Figure 12.28 Standard section and plan of the storage tank facility (Courtesy of Fudo Tetra Corporation)

~1.5: m
Trace of sand boil
13.00: m

Crack
SCP 1.8: m

1.8: m
0

9.00: m 18.00: m 9.00: m

36.00
φ
1.8: m

Crack

Trace of sand boil

400

Seismic Design of Foundations

Liquefaction countermeasures

12.7.1.4 Improvement effectiveness in buildings (1995 Hyogo-ken Nambu earthquake)

Large-scale liquefaction occurred in the two artiﬁcial islands of Port Island and Rokko Island in the Kobe Port
area as a result of the 1995 Kobe earthquake. Various types of ground improvement were undertaken for
structures on both islands, as measures to accelerate consolidation settlement in the alluvial clay layer due to
excess weight, and to increase the bearing capacity of the landﬁll soil layer. In the improved sections, there
was little damage caused by settlement and other effects as compared with the damage in unimproved areas.
Figure 12.29 shows the measured post-earthquake settlements of the buildings with pile foundations in the
areas of improved ground. On both islands, 40–50 cm settlement occurred in the unimproved areas, but in the
areas improved by compaction methods, including SCPs, the settlement was negligible (Yasuda et al., 1996).

In addition, six cases of subsidence measurement of buildings on spread foundations located on the artiﬁcial
islands were conducted before and after the earthquake (Kakurai et al., 1996). Figure 12.30 shows the cross-
sections of these buildings. The foundation ground beneath buildings a–e were improved by compaction,
while no improvement was implemented under building f. Also indicated in Figure 12.30 are the total bearing
load from each of the buildings, the amount of absolute settlement measured and the improvement ratio. As
shown in the ﬁgure, the settlement following the earthquake of building f with unimproved foundation
ground was about 20 cm whereas the amount of settlement in buildings a–e was of the order of several
centimetres, depending on the extent of improvement. The difference is clearly illustrated in Figure 12.31.

Figure 12.29 Relative settlement and methods of ground improvement (buildings with pile foundation) (Yasuda
et al., 1996; reprinted with permission from the author)

No. of sites measured

0 5 10 15
100

90 Port Island (PI)

Rokko Island (RI)
80
Average in PI
70 Average in RI
Subsidence: cm

0
Untreated Preloading Sand drains Sand drains Rod (vibro) Sand compaction
plus preloading compaction piles

401

Seismic Design of Foundations

402

Seismic Design of Foundations

Figure 12.30 Absolute settlement and SCP-improved area (buildings on a spread foundation); KP, Kobe Port construction datum level (Courtesy of Fudo
Tetra Corporation)

Building c
Seismic Design of Foundations

Building b
Building a
60 kN/m2 130 kN/m2
GL 40 kN/m2 KP + 9.0 m GL KP + 9.0 m GL KP + 6.3 m
(1.9 cm) (4.5 cm) (4.2 cm)
SCPzone SCPzone SCPzone
-19 -17 -18
–10 (as = 10%) KP + 1.0 m (as = 10%) KP + 3.0 m -23
(as = 8%)
-26 Fill
–20 Fill -25 Fill
-36 Alluvium clay
–30

Depth: m
-41 Alluvium clay Alluvium clay
–40 -45 Diluvium layer
Diluvium layer
Diluvium layer
Building d
50 kN/m2 Building f
Building e (12.6 cm)
110 kN/m2 GL 60 kN/m2
GL KP + 7.2 m GL KP + 5.2 m KP + 6.0 m
(2.0 cm) (20.3 cm)
KP + 3.0 m
SCPzone -11 SCPzone (as = 6%)
–10 -20 KP + 3.0 m
-24 (as = 12%) -21 -22 Fill
Fill
–20 Fill
Alluvium clay -35 Alluvium clay
–30 -41 Alluvium clay -38

Depth: m
–40 Diluvium layer
Diluvium layer
Diluvium layer
Note: GL indicated is from boring log
Liquefaction countermeasures

Figure 12.31 Building with foundation ground (a) not improved and (b) improved by the SCP method (Courtesy of
Fudo Tetra Corporation)

(a)

(b)

403

Seismic Design of Foundations

12.7.1.5 Improvement effectiveness in buildings (2001 To ̄ hoku Paciﬁc earthquake)

Following the 11 March 2011 Tō hoku Paciﬁc earthquake, liquefaction was observed in many areas adjacent
to Tokyo Bay, about 390 km from the epicentre. Manholes were uplifted, grounds settled, and buildings and
bridges were damaged as a result of liquefaction.

A medical centre building is located in reclaimed land along the Tokyo Bay, as shown in Figure 12.32(a). The
building is ﬁve storeys high and supported by piles. After the liquefaction assessment, it was judged that there

Figure 12.32 (a) Plan view of the SCP-improved area for the Tokyo Bay medical centre building (Kinoshita et al.,
2012); (b) layout of compacted piles (Courtesy of Fudo Tetra Corporation)

Non-vibratory SCP method

(using gravel)
Improved area

Location of building

Non-vibratory SCP method (using sand)

(a)
1500 mm

1500 mm

Extra
improved area
Improved area Improved ratio
Unimproved area (building) as = 17.1%

Building base

Gravel pile
Sand pile
(mat: gravel)
(mat: sand)

(b)

404

Seismic Design of Foundations

Liquefaction countermeasures

Figure 12.33 Successful ground improvement in the Tatsumi area (Courtesy of Fudo Tetra Corporation)

Improved area
Unimproved area

is a high potential for liquefaction and, consequently, the non-vibratory SCP method was used at this site as a
countermeasure against liquefaction (Nunokawa et al., 2012). The improvement speciﬁcation was a square
arrangement with a pitch of 1.5 m (as = 16.7%) and the length of the piles was 12 m (see Figure 12.32(b)). In
the surrounding area of the improved site, gravel rather than sand was used to dissipate the excess water
pressure from the liqueﬁed area. As shown in Figure 12.33, although liquefaction-induced damage was
observed outside the improved area, no damage was seen within the improved area.

12.7.2 Performance after the 2010–2011 Canterbury Earthquake Sequence (CES)

Prior to the 2010–2011 CES in New Zealand, soil improvement had been implemented at some sites in
Christchurch to mitigate the risk from liquefaction, especially at sites of multi-storey buildings, large-scale
municipal facilities and bridge approaches. The most common method had been vibro-replacement/vibro-
displacement stone columns. Following the CES, Wotherspoon et al. (2014) reviewed the performance of
some improved ground sites, and concluded that most of the remediated sites generally performed well.
Although the adjacent untreated ground showed evidence of liquefaction, the structures constructed on
remediated sites generally did not undergo signiﬁcant damage.

An example of a site improved by the ground improvement method is the foundation ground underneath the
AMI Stadium, located southeast of the Christchurch central business district (CBD). Built originally in 1880,
the stadium underwent several renovations and expansions. In 2000, some portions of the stadium were
demolished and the west stand, now called the Paul Kelly Motor Company Stand, was constructed. In addition,
as part of a redevelopment for the 2011 Rugby World Cup, the east stand was demolished and replaced with
the new Deans Stand. The soil proﬁle beneath these two stands consists of stiff silts to a depth of 3–5 m,
overlying medium-dense to dense sand to a depth of 16–20 m. Below the sand layer are soft silts to between
23 and 25 m depth, at which point the stiff Riccarton Gravels are encountered (Tonkin & Taylor, 2011).

405

Seismic Design of Foundations

During the expansion works, both the east stand (Deans Stand) and the west stand (Paul Kelly Stand) were
renovated such that they were supported by stiff and well-connected ground beams. Moreover, the foundation
ground underneath these stands was improved with stone columns to reduce the potential for liquefaction-
induced ground damage. Stone columns reduce the effects of liquefaction by limiting the extent and severity
of liquefaction that can occur, and by reinforcing the ground to limit foundation settlement and maintain
bearing capacity.

The stone columns under the Paul Kelly Stand are 600 mm in diameter and arranged in a triangular grid
pattern with 1.8 m centre-to-centre spacing. These extend to a depth of 8 m, covering an area of 12 500 m2.
Under the Deans Stand, the stone columns are 900 mm in diameter and arranged in a triangular grid
pattern with 2.7 m centre-to-centre spacing. These extend to 9 m depth (Tonkin & Taylor, 2011). The north
and south stands (called the Hadlee and Tui Stands, respectively) are built on shallow foundations without
ground improvement works. The layout of the stadium and the locations of these stands are depicted in
Figure 12.34(a).

Following the February 2011 Christchurch earthquake, the stadium suffered extensive damage as a result of
soil liquefaction and high levels of ground shaking (see Figures 12.34(b) and 12.34(c)). The Hadlee Stand
suffered severe structural damage, and was recommended for demolition while the Tui Stand, constructed on
a ﬁll platform to raise its level, suffered less severe structural damage. The Deans and Paul Kelly Stands, both
with foundation ground improved by stone columns, suffered structural damage from differential settlement

Figure 12.34 AMI Stadium: (a) aerial view following Christchurch earthquake (adapted from CERA 2013); (b) ejecta
outside stadium following the Christchurch earthquake; and (c) ejecta and blistering of playing turf inside stadium

Hadlee stand N
Paul Kelly stand

Deans stand

Tui stand

(a)

406

Seismic Design of Foundations

Liquefaction countermeasures

Figure12.34 (Continued)

(b)

(c)

and ground shaking. The Paul Kelly Stand settled by up to 400 mm, with variation in settlements across the
structure of approximately 80–100 mm. There was no visible ejected material within the stadium footprint, as
the thick foundation slab supporting the structure prevented this from occurring. The Deans Stand settled by
generally the same amount, but with much larger variations in settlements across the structure of up to
300 mm. The difference in the differential settlement was likely due to the difference in foundation systems,
with the stiff slab beneath the entire footprint of the Paul Kelly Stand more resistant to differential settlement
compared with the ground beams beneath the Deans Stand.

407

Seismic Design of Foundations

The stadium was subjected to shaking of the order of 0.5g, which equates to a magnitude-weighted (M 7.5)
value of 0.37g, which is greater than the design ultimate limit state PGA of 0.32g for an Importance Level (IL)
3 structure such as this stadium based on the code requirements at that time.

Figure 12.35 provides an indication of the change in the CPT tip resistance qc between stone column
installation locations before and after ground improvement beneath the Deans and the Paul Kelly Stands. The
two parts of the ﬁgure indicate that improvement was more successful beneath the Paul Kelly Stand. The most
improvement beneath the Deans Stand was in the clean sands between 5.5 and 9 m depth, with no
improvement in the surface silts to 3 m depth, and a small amount of improvement in the silty sands between
these two layers. The ground improvement was more successful in the silty sands beneath the Paul Kelly
Stand, as evidenced by the higher tip resistance over the entire 8 m of stone column installation (Tonkin &
Taylor, 2011).

Figure 12.35 Summary of CPT tip resistance before and after ground improvement: (a) beneath the Deans Stand in
2008; (b) beneath the Paul Kelly Stand in 2001 (Tonkin & Taylor, 2011). The shaded portion represents the
interquartile range of the CPT data from both locations

20
Range of post-
18
CPT tip resistance, qc : MPa

improvement CPTs
16
14
Required post- improvement
12
qc in ground improvement
10 specification
8
6
4 Range of pre-
2 improvement CPTs

0
0 1 2 3 4 5 6 7 8 9 10
Depth below ground level: m
(a)
25
CPT tip resistance, qc : MPa

Range of post-
20 improvement
CPTs
15

5
Range of pre-
improvement CPTs
0
0 1 2 3 4 5 6 7 8 9 10
Depth below ground level: m
(b)

408

Seismic Design of Foundations

Liquefaction countermeasures

Wotherspoon et al. (2014) reported that there were still liquefaction-related ground failure and/or intolerable
settlements that occurred at some facilities where ground treatment was used. They concluded that these are
likely the result of the following factors

■ ground shaking exceeded the design levels

■ some sites contained liquefiable silty strata that were not effectively densified during treatment,
resulting in liquefaction of the soil in between the stone columns
■ some ground improvement works (e.g. stone columns) did not extend through the entire liquefiable
layer, meaning that liquefaction occurred in the soil layers beneath the base of the treated zone.

Except for the AMI Stadium and other sites near the CBD whose foundation grounds were improved prior to
the CES, most of the improved sites reviewed were subjected to very low PGAs and were not sufﬁciently
tested (Wotherspoon et al., 2014). It is also of note that several improvement methods were applied following
the 2010 Darﬁeld earthquake, and these sites were put to test by the succeeding large aftershocks. Overall, the
mitigation measures have performed successfully, and their performance provides immediate insights on the
anticipated foundation response during future large events.

12.8. Methods to strengthen shallow foundations

Recent large-scale earthquakes have highlighted the destructive potential of soil liquefaction and ground
failure to residential houses and small-scale structures supported by shallow foundations. For example, the
shaking from the 2010–2011 Canterbury earthquake events induced liquefaction of varying degrees in the
Canterbury region, with ground subsidence as a result of liquefaction-induced differential settlement, lateral
spreading and ejection of liqueﬁed materials (water and sand boils) as the principal ground deformation
modes that caused damage to residential buildings in the region. Following the major events of the CES, a
survey of land damage and housing foundation damage was undertaken as part of the coordinated response
by agencies of the New Zealand government. About 8000 residential houses have been abandoned following
the earthquakes, while 20 000 houses were severely affected and an additional 60 000 houses were affected
by the liquefaction (van Ballegooy et al., 2014). Typical damage observed to house foundations is depicted
in Figure 12.36, while examples of damage to slabs on grade, mostly due to lateral spreading, are shown in
Figure 12.37.

Representative types of shallow foundations include

■ isolated pads
■ strip/beam footings (where the continuity of foundation elements ensures the integrity of a structure
subject to differential ground movements)
■ pad and tie beam foundations
■ mat or raft foundations (e.g. continuous structural slabs spanning between columns and walls that can
have sufﬁcient strength and stiffness to behave essentially as a rigid body when accommodating
differential ground movements).

Settled footings may be the result of liquefaction or the soil response at depth, or simply have been overloaded
by the earthquake-induced axial loads. Basements can be exposed to high uplift pressures generated in
liqueﬁed sands or in loose gravels. This can result in vertical displacement as well as damage to the basement

409

Seismic Design of Foundations

Figure 12.36 Observed deformation mode/damage to residential houses (DBH, 2011)

Uniform settlement Tilt settlement Lateral stretching

δ δ δ

Hogging Sagging or dishing Racking/twisting

δ
Back Front
δ δ

Figure 12.37 Typical damage observed in house foundations, which were not designed to carry lateral loads,
following the Christchurch earthquake

410

Seismic Design of Foundations

Liquefaction countermeasures

ﬂoor, depending on the construction as a raft or slab between footings. Uplifted basements, particularly those
on gravels rather than liqueﬁed sands, may have large voids below them. Basement walls may have been
subjected to lateral earth pressures much higher than the normal static loading.

Strengthening shallow foundations to resist liquefaction-induced ground deformations can be done in two
ways: (a) by improving the ground adjacent to the footings and (b) designing the footings to resist differential
settlement and lateral stretching.

12.8.1 Ground improvement adjacent to shallow foundations

If the soil adjacent to a footing is remediated such that liquefaction will not occur, then direct damage to
the foundation can be avoided. The liquefaction countermeasure techniques discussed in Section 12.3 are
generally applicable to remediate shallow ground for residential houses, although in some cases (e.g. the
solidiﬁcation technique or compaction grouting) the cost may be prohibitive.

Following the CES, the New Zealand Earthquake Commission (EQC) sponsored a ground improvement
programme (GIP) for the purpose of investigating the efﬁciency of various cost-effective and practical
shallow ground improvement methods for repairing and strengthening existing residential properties to
improve their resistance to future liquefaction-induced damage. Within the GIP, eight different ground
improvement methods were trialled: rapid impact compaction, rammed aggregate piers, horizontal soil-mixed
beams, low-mobility grout, the reinforced gravel raft, the reinforced soil–cement raft, driven timber poles and
resin injection (EQC 2015). These methods are illustrated in Figure 12.38. The methods all aim to thicken
and/or stiffen the near-surface soil layers to reduce liquefaction vulnerability.

The improved grounds were subjected to extensive in situ cyclic testing (employing vibroseis T-Rex testing
(using a truck-mounted vibrating machine) and blast-induced liquefaction testing by detonating explosive
charges in the ground) to liquefy the surrounding soils, along with standard in-situ investigation methods such
as CPTs, dilatometer tests and cross-hole geophysical testing.

The results showed that shallow ground improvements do not signiﬁcantly reduce ground surface subsidence
as a result of the liquefaction of the underlying soil layers, but they do improve crust rigidity and reduce the
differential ground surface subsidence that damaged buildings on top of the improved ground. The rapid-
impact compaction and rammed-aggregate piers ground-improvement methods work well in building thicker
non-liquefying crusts, reducing liquefaction vulnerability. Stiff soils (stiff surface crust) behave more rigidly
compared with less-stiff crusts, reducing the likelihood of differential ground surface subsidence (undula-
tions, tilt and differential settlement). The 1.2 m-thick shallow reinforced soil–cement and reinforced gravel
rafts work well in improving crust rigidity. Driven-timber poles do not prevent liquefaction triggering in the
near-surface soils but they do help to redistribute the weight of the house and make the liquefaction-induced
ground surface subsidence more uniform, which means a reduction in differential ground surface subsidence.

Generally, the results from the GIP indicated the importance of crust thickness for the performance of res-
idential dwellings, conﬁrming observations made following the CES. The outcomes from the trial programme
have made a signiﬁcant impact to the repair and rebuild of residential structures in the Canterbury region.
Following the successful GIP, cheaper ground improvement solutions to mitigate the risks of liquefaction in
residential lands have been endorsed.

411

Seismic Design of Foundations

Figure 12.38 Various methods trialled in the GIP conducted by EQC (2015): (a) rapid-impact compaction;
(b) Geopier Rammed Aggregate Pier system; (c) driven-timber poles; (d) low-mobility grout and resin injection;
(e) reinforced gravel raft and reinforced soli–cement rafts; (f) horizontal soil-mixed beams

Non-liquefiable Non-liquefiable
crust crust

Liquefiable soil
Liquefiable soil

(a) (b)

Non-liquefiable Non-liquefiable crust

crust

Liquefiable soil
Liquefiable soil

Non-liquefiable
crust

Liquefiable soil Geogrid

Non-liquefiable crust

Non-liquefiable
Cement-stabilised soil
crust
Liquefiable soil
Liquefiable soil Geogrid

(e) (f)

412

Seismic Design of Foundations

Liquefaction countermeasures

12.8.2 Structural modiﬁcation of shallow foundations

Another approach to mitigating liquefaction-induced damage to shallow foundations is to allow liquefaction
to happen, but strengthening the foundations against liquefaction-induced lateral spreading and differential
settlement. This can be done by increasing the size of footings or providing more reinforcement.

Figure 12.39(a) shows an unﬁnished house in Christchurch with the foundation exposed. Such concrete slab-
on-grade foundations have been popular in New Zealand for single- and two-storey timber-framed houses.
The perimeter beams, measuring 250 mm × 530 mm, have one D12 reinforcement each at the top and bottom,
with concrete cover of about 80 mm. The protruding bars consisted of D10 bars, which will be bent inward,
and concrete will be poured over the site to form the unreinforced concrete slab. It is clear that this house does
not have enough resistance against uneven settlement or lateral spreading due to liquefaction. Note that sand
boils were observed adjacent to the exposed foundation. On the other hand, Figure 12.39(b) illustrates a house
being constructed at a site in Tokyo, Japan. Note the amount of reinforcement for the tie beams, indicating a
stiffened foundation system.

Japanese practice has increasingly employed both grade beams and continuous reinforced foundations for
low-to-moderate height structures, and the performance of these types of systems in earthquakes has been
good. Note that although stiff, shallow foundations can undergo differential settlement resulting in rotational
tilting of the structure, these can be easily re-levelled after earthquake-induced settlement through several
techniques, including jacking, under-excavation and compaction grouting.

Another way of preventing damage to shallow foundations is to underpin with deep foundations, such as
piles. This was adopted in many houses in Christchurch following the CES. Typically, the decision was not
based solely on the performance of the foundation after CES but on the risks of damaging settlement from
future events, based on proper analysis of the ground conditions. While differential settlement as a result of
previous earthquake may be within tolerable limits for the structure, another earthquake could produce similar
or greater differential movement, cumulative to the ﬁrst, which could then lead to severe structural damage or
failure.

Figure 12.39 Examples of shallow foundation construction: (a) in Christchurch, New Zealand; (b) in Tokyo, Japan

(a) (b)

413

Seismic Design of Foundations

12.9. Methods to strengthen pile foundations

Foundation elements are considered to be deep when the depth D to breadth B ratio is greater than 5 (D/B > 5).
Generally, the deep foundation types in use are

■ driven piles (timber, concrete or steel)

■ bored cast-in-place piles
■ bulb (Franki) piles
■ screw piles
■ continuous ﬂight auger (CFA) piles.

Most piles in liqueﬁable ground are end bearing and installed to found onto dense sand or gravel stratum.
Although often designed as end bearing with some contribution from side resistance, in reality, for many of
them, the gravity loads will have been carried since construction by the side resistance mechanism.

Deep foundations, whose lengths extend below the occurrence of liquefaction (or significant cyclic softening
due to partial liquefaction), can provide reliable vertical support and so can reduce or eliminate the risk of
unacceptable liquefaction-induced settlements. However, pile foundations do not necessarily prevent damage
that may occur as a result of differential lateral structural displacements, so piles must be coupled with
sufficient lateral structural connectivity at the foundation so as to safely resist unacceptable differential lateral
displacements. Moreover, pile foundations are likely to be damaged by lateral forces caused by a decrease in
the subgrade reaction accompanying liquefaction and associated lateral spreading of liquefied soil.

There are several issues that need to be considered for deep foundations in a liqueﬁable ground

■ Loss of side resistance (skin friction) in piles. This may occur from pore water pressure increase
during shaking, even if full liquefaction does not trigger. Where full liquefaction is triggered at depth,
all side resistance above may be effectively lost or reversed because of settlement of the overlying
strata. In such cases, so-called ‘negative skin friction’ may contribute to pile settlement.
■ Pile settlement when the soil below the base liquefies. Some types of piles, such as bored cast-in-place
piles, may be susceptible to settlement when pore water pressure builds up and liquefaction occurs
above the base of the pile. If overloaded, the pile will settle until the end-bearing mechanism is
mobilised, resulting in differential settlement that may have severe consequence for a structure.
■ Effect of liquefaction at depth on the pile capacity. Pile settlement may also occur as a result of
liquefaction of sand layers below the founding layer. It is possible that the site would have dense
gravel or sand layers that may be several metres thick but underlain by much looser sands. Deeper
liquefaction should be considered in the design of pile foundations.
■ Possible loss of lateral capacity. If the piles are not adequately embedded into the dense strata, the
lateral capacity of the pile may be lost, making it susceptible to damage in a laterally spreading
ground. Limited base sliding may also be possible, although this type of failure is seldom observed.
■ Kinematic interactions between the ground and the piles. Such interaction needs to be carefully
considered during design. Ground deformations induced by liquefaction, either cyclic or permanent,
may be significant; these may impose significant strains within piles, resulting in pile damage and
permanent deformation well below the ground surface. Therefore, estimation of pile strain levels and
the likelihood of pile damage should be conducted using available analytical procedures.
■ Buckling of slender piles in liquefied soil may also occur.

414

Seismic Design of Foundations

Liquefaction countermeasures

Similar to shallow foundations, liquefaction-induced damage to deep foundations can be prevented through
two general approaches: (a) by improving the ground adjacent to the pile foundations and (b) by strength-
ening the pile foundation.

Note that signiﬁcant research effort since the turn of the century has led to the development of a range of
methods to address the effect of laterally spreading ground on pile foundations, ranging from simpliﬁed
pseudo-dynamic approaches to fully non-linear, time domain, fully coupled effective stress analysis of the
soil–pile–superstructure interaction. These types of methods, complemented by appropriate conservatism,
can provide a suitable basis for analysis of this issue, and for the design and detailing of piles and pile–cap
connections.

12.9.1 Ground improvement adjacent to pile foundations

As mentioned above, liquefaction of the soil adjacent to or surrounding piles can lead to pile damage. If that
zone is improved to prevent liquefaction and/or the associated ground deformation, then pile integrity can be
maintained. In this vein, the approaches can be classiﬁed into two groups, depending on the concept involved

■ Prevention of the occurrence of liquefaction

– by improving the ground surrounding the piles by compaction/densiﬁcation (however, sufﬁcient

consideration must be given to the influence of the ground improvement installation on existing piles,
which may include an increase in bending stresses by compaction, excess pore water pressure and
vibration resulting from compaction, and the influence of earth pressure increased by compaction)
– by improving the ground by solidification
– by surrounding the piles with a diaphragm or cement mixing wall (see Section 12.6.2)

■ Allow liquefaction to occur, but prevent the occurrence of lateral spreading

– by installing stiff/compacted sand walls in the direction of movement

– by improving a zone adjacent to the piles by densiﬁcation or solidiﬁcation.

Examples of conditions of the ground adjacent to buildings supported by a pile foundation following
earthquakes are shown in Figure 12.40. These were observed in Urayasu City following the 2011 East Japan
earthquake. Figure 12.40(a) shows a pile protruding out of the surface due to the subsidence of the liquefied
ground around the piles. If the ground were treated by ground improvement methods, no such differential
settlement would occur; instead, a little subsidence would take place at the boundary between the improved
and unimproved areas. Figures 12.40(b) and 12.40(c) show the boundary between unimproved areas and
improved sections by the solidification and compaction methods, respectively. Comparing these figures,
sections near compaction-improved ground seem to deform smoothly at the boundary, possibly due to a
densification effect.

12.9.2 Structural modiﬁcation of pile foundations

In terms of strengthening the piles through structural retroﬁtting, several methods are available. These include

■ increasing the number of piles

■ increasing the stiffness of piles (increasing the diameter or amount of reinforcement; use of steel piles)
■ adding batter piles or inclined piles to increase the lateral capacity.

415

Seismic Design of Foundations

416

Seismic Design of Foundations

Figure 12.40 Typical observed liquefaction-induced damage at sites of buildings supported by pile foundations following the 2011 To
̄ hoku earthquake
(Japan) (Harada, 2011)
Seismic Design of Foundations

Improved
area
Unimproved area
Unimproved
area
Sand boil
Improved area

Pile exposure
due to ground settlement

Ground improvement: lattice-type Ground improvement: non-vibratory

deep-mixing method SCP method

Structure Structure Structure

Supported pile Settlement Settlement Settlement

Solidification Compaction
improvement improvement

(a) (b) (c)

Liquefaction countermeasures

12.10. Summary
Various remediation techniques are available to mitigate the liquefaction risk to structures constructed on
loose saturated sandy deposits. These techniques can be roughly classiﬁed into two groups: the ﬁrst is to
improve the soil so that liquefaction may not occur, while the second is to provide measures for structures to
prevent or minimise damage even if liquefaction occurs. Remedial treatment of soils to prevent liquefaction is
based on either the principle of improving the properties of soils or the principle of changing stress or
deformation conditions. In some cases, liquefaction is allowed to occur but the damage is prevented or
minimised by strengthening the structure or foundations.

Ground improvement is being used increasingly for remediating liquefiable soils due to the wide variety of
methods that are available. Backed by several decades of research and experience, these methods have been
used in many soil conditions and have been developed by both the construction industry and research
institutions. These techniques can be adapted to site-specific conditions and, in some cases, one or more
methods combined to provide economical solutions for liquefaction remediation problems. Often, new
ground improvement techniques are developed when difficult circumstances are encountered, and these new
techniques are then refined and become part of the expanding set of available mitigation measures.

A few of the remediation methods that have been implemented at many sites all over the world have been
tested by few large-scale earthquakes, and most of them performed well, with negligible or minor ground
damage. Thus, these remediation techniques will continue to play an important role in the mitigation of
seismic risk to existing and future structures.

REFERENCES
Andrus RD and Chung RM (1995) Ground Improvement Techniques for Liquefaction Remediation Near Existing
Lifelines. National Institute of Standards and Technology, Gaithersburg, MD, USA, Report NISTIR 5714.
Baez JI (1995) A Design Model for the Reduction of Soil Liquefaction by Using Vibro-Stone Columns. PhD thesis,
University of Southern California, Los Angeles, CA, USA.
Baez JI. and Martin GR (1993) Advances in the design of vibro systems for the improvement of liquefaction
resistance. Ground Improvement 7th Annual Symposium, Vancover, BC, Canada.
Barksdale RD and Bachus RC (1983) Design and Construction of Stone Columns. Federal Highway Adminis-
tration, Ofﬁce of Engineering and Highway Operations, Research and Development, Washington, DC, USA,
vol. I, Report FHWA/RD-83/026.
Boulanger R, Idriss I, Steward D, Hashash Y and Schmidt B (1998) Drainage capacity of stone columns or gravel
drains for mitigating liquefaction. Proceedings of Geotechnical Earthquake Engineering and Soil Dynamics III.
American Society of Civil Engineers, Reston, VA, USA.
Brennan AJ and Madabhushi SPG (2002) Effectiveness of vertical drains in mitigation of liquefaction.
Soil Dynamics and Earthquake Engineering 22: 1059–1065.
CERA (Canterbury Earthquake Recovery Authority) (2013) Geotechnical Database for Canterbury Earthquake
Sequence. https://canterburygeotechincaldatabase.projectorbit.com (accessed 12/08/2012).
Cooke HG and Mitchell JK (1999) Guide to Remedial Measures for Liquefaction Mitigation at Existing Highway
Bridge Sites. Multidisciplinary Center for Earthquake Engineering Research, Buffalo, NY, USA, Technical
Report MCEER-99-015.
Cox BR and Grifﬁths SC (2010) Practical Recommendations for Evaluation and Mitigation of Soil Liquefaction in
Arkansas. Department of Transportation, Washington, DC, USA, Project Report MBTC 3017.
DBH (Department of Building and Housing) (2011) Repairing and Rebuilding Houses Affected by the Canterbury
Earthquake Sequence, revised issue. DBH, Wellington, New Zealand.

417

Seismic Design of Foundations

Dise K, Stevens MG and von Thun JL (1994) Dynamic Compaction to Remediate Liquefiable Embankment
Foundation Soils. American Society of Civil Engineers, Reston, VA, USA.
EQC (Earthquake Commission) (2015) Residential Ground Improvement: Findings from Trials to Manage
Liquefaction Vulnerability. EQC, Wellington, New Zealand.
Fudo Construction (1971) Design Manual of Compaction System. (In Japanese.)
Fudo Construction (1993) Investigation Report on the 1993 Hokkaido Nansei-oki Earthquake. Fudo Construction,
Tokyo, Japan. (In Japanese.)
Fudo Construction (2000) Investigation Report on the 2000 Tottori-ken Seibu Earthquake. Fudo Construction,
Tokyo, Japan. (In Japanese.)
Harada K (2011) Non-vibratory SCP method as countermeasures against liquefaction along coastal areas –
improvement effectiveness during the 2011 Off the Pacific Coast of Tō hoku earthquake. Doboku Sekou 12:
54–57. (In Japanese.)
Harada K, Nozu M, Shinkawa N and Matsushita K (2015) Countermeasures against liquefaction and verification of
their effectiveness during past large-scale earthquakes in Japan. Proceedings of the 40th Annual Conference on
Deep Foundations, Oakland, CA, USA.
Hausler EA and Sitar N (2001) Performance of soil improvement techniques in earthquakes. 4th International
Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics, San Diego, CA,
USA, paper 10.15.
Hausmann MR (1990) Engineering Principles of Ground Modification. McGraw-Hill, New York, NY, USA.
Hayward Baker (2015) Vibro Replacement. http://www.haywardbaker.com/solutions/techniques/vibro-replace-
ment (accessed 22/08/2018).
Iai S and Koizumi K (1986) Estimation of earthquake induced excess pore pressure for gravel drains. Proceedings
of the 7th Japan Earthquake Engineering, Symposium, Tokyo, Japan, pp. 679–684.
Iai S, Matsunaga Y, Morita T et al. (1993). Effects of remedial measures against liquefaction at 1993 Kushiro-oki
earthquake. Proceedings of the 5th US–Japan Workshop on Earthquake Resistant Design of Lifeline Facilities and
Countermeasures against Soil Liquefaction. National Center for Earthquake Engineering Research, Taiwan.
Idriss IM and Boulanger RW (2008) Soil Liquefaction During Earthquakes. Earthquake Engineering Research
Institute, Oakland, CA, USA.
Iida K, Kogai Y and Ohbayashi J (2005) Improvement effectiveness of sand compaction pile method as a
countermeasure against liquefaction at Nemuro Port. Proceedings of the 40th National Conference of the
Japanese Geotechnical Society, Hakodate, Japan, pp. 2191–2192. (In Japanese.)
Ishihara K, Kawase Y and Nakajima M (1980) Liquefaction characteristics of sand deposits at an oil tank site
during the 1978 Miyagiken-Oki earthquake. Soils and Foundations 20(2): 97–112.
JGS (Japanese Geotechnical Society) (1996) Investigation Report on the 1994 Sanriku Haruka-oki Earthquake.
JGS, Tokyo, Japan. (In Japanese.)
JGS (1998) Remedial Measures against Soil Liquefaction. Balkema, Rotterdam, the Netherlands.
JSCE (Japan Society of Civil Engineers) (1986) Investigation Report on the 1983 Nihonkai-Chubu Earthquake.
JSCE, Tokyo, Japan. (In Japanese.)
Kakurai M, Aoki M, Hirai Y and Matano H (1996) Investigation of spread foundations on liquefied man-made
islands. Tsuchi-to-Kiso 44(2): 64–66.
Kinoshita H, Nozu M, Ohbayashi J and Harada K (2012) Sand compaction pile technology and its performance in
both sandy and clayey grounds. ISSMGE – TC 211 International Symposium on Ground Improvement IS-GI
Brussels, Belgium.
Kitazume M (2005) The Sand Compaction Pile Method. Balkema, Rotterdam, the Netherlands.
Kitazume M and Terashi M (2013) The Deep Mixing Method. CRC Press/Balkema, Leiden, the Netherlands.

418

Seismic Design of Foundations

Liquefaction countermeasures

Lukas RG (1995) Geotechnical Engineering Circular No. 1: Dynamic Compaction. Federal Highway Adminis-
tration, Washington, DC, USA.
Mitchell JK and Katti RK (1981) Soil improvement - state of the art report. Proceedings of the 10th International
Conference on Soil Mechanics and Foundation Engineering, Stockholm, Sweden, vol. 4, pp. 509–565.
Namikawa T, Koseki J and Suzuki Y (2007) Finite element analysis of lattice-shaped ground improvement by
cement-mixing for liquefaction mitigation. Soils and Foundations 47(3): 559–576.
Nashed R (2005) Liquefaction Mitigation of Silty Soils Using Dynamic Compaction. PhD thesis, State University
of New York, Buffalo, NY, USA.
Nishikawa J, Kamata T and Kaji M (1995) Damage to highways, railways and river embankments from the 1994
Hokkaido Toho-oki earthquake. Tsuchi-to-Kiso 43(4): 7–10. (In Japanese.)
Nunokawa N, Harada K, Imai Y, Mishiro T and Taguchi Y (2012) Verification on liquefaction index for judgment
and improvement effectiveness during the 2011 off the Pacific Coast of Tō hoku earthquake. Proceedings of the
Japanese Conference on Geotechnical Engineering in Kanto Region, Tokyo, Japan. (In Japanese.)
Ohbayashi J, Harada K, Yamamoto M and Sasaki T (1998) Evaluation of liquefaction resistance in compacted
ground. Proceedings of the 10th JAEE Symposium, pp. 1411–1416. (In Japanese.)
Onoue A (1988) Diagrams considering well resistance for designing spacing ratio of gravel drains. Soils and
Foundations 28(3): 160–168.
Orense R, Yamada S and Otsubo M (2012) Soil liquefaction in Tokyo Bay area due to the 2011 Tō hoku (Japan)
earthquake. Bulletin of the New Zealand Society for Earthquake Engineering 45(1): 15–22.
Poh TY and Wong IH (2001) A field trial of jet grouting in marine clay. Canadian Geotechnical Journal 38:
338–348.
Port and Harbour Research Institute (1997) Handbook on Liquefaction Remediation of Reclaimed Land. Balkema:
Rotterdam, the Netherlands.
Priebe HJ and Grundbau K (1998) Vibro replacement to prevent earthquake induced liquefaction. Proceedings of
the Geotechnique-Colloquium, Darmstadt, Germany, technical paper 12-57E.
Sasaki Y, Tamura K, Yamamoto M and Ohbayashi J (1995) Soil improvement work for river embankment
damaged by 1993 Kushiro-oki earthquake. Proceedings of the 1st International Conference on Earthquake
Geotechnical Engineering, Tokyo, Japan, vol. 1, pp. 43–48.
Schaefer V (ed.) (1997) Ground Improvement, Ground Reinforcement, Ground Treatment: Developments 1987–1997.
American Society of Civil Engineers, Reston, Reston, VA, USA.
Seed HB and Booker JR (1977) Stabilization of potentially liquefiable sand deposits using gravel drains. Journal of
the Geotechnical Engineering Division, ASCE 103(GT7): 757–768.
Slocombe BC (2004) Dynamic compaction. In Ground Improvement (Moseley MP and Kirsch K (eds)), 2nd edn.
Spon Press, London, UK, ch. 3.
Suzuki Y, Saito S, Onimaru S, Kimura T, Uchida A and Okumura R (1996) Grid-shaped stabilized ground
improved by deep cement mixing method against liquefaction for a building foundation. Tsuchi-to-Kiso 44(3):
46–48. (In Japanese.)
Thevanayagam S, Martin G, Nashed R, Shenthan T, Kanagalingam T and Ecemis N (2006) Liquefaction
Remediation of Silty Soils Using Dynamic Compaction. Multidisciplinary Center for Earthquake Engineering
Research, Buffalo, NY, USA, Technical Report 06-0009.
Tokimatsu K and Seed HB (1987) Evaluation of settlements in sand due to earthquake shaking. Journal of
Geotechnical and Geoenvironmental Engineering, ASCE 113(8): 861–878.
Tokimatsu K, Mizuno H and Kakurai M (1996) Building damage associated with geotechnical problems. Soils and
Foundations, pp. 219–234 (special issue).
Tonkin & Taylor (2011) AMI Stadium Geotechnical Report. Vbase, Christchurch, New Zealand.

419

Seismic Design of Foundations

van Ballegooy S, Malan P, Lacrosse V et al. (2014) Assessment of liquefaction-induced land damage for residential
Christchurch. Earthquake Spectra 30(1): 31–55.
Watanabe T (1966) Damage to oil reﬁnery plants and a building on compacted ground by the Niigata earthquake
and their restoration. Soils and Foundations 6(2): 86–99.
Welsh JP (1992) Grouting techniques for excavation support. Proceedings of Excavation Support for the Urban
Infrastructure. American Society of Civil Engineers, Reston, VA, USA, pp. 240–261.
Wotherspoon LM, Orense RP, Jacka M, Green RA, Cox BR and Wood CM. (2014) Seismic Performance of
improved ground sites during the 2010–2011 Canterbury Earthquake Sequence. Earthquake Spectra 30(1):
111–129.
Xanthakos PP, Abramson LW and Bruce DA (1994) Ground Control and Improvement. Wiley, New York, NY,
USA.
Yamada H and Sato M (1990) Site tests related to work to implement countermeasures to liquefaction at Kushiro
Port. Presentation Outlines (4th set) of the 34th Hokkaido Development Agency Technical Research Meeting,
Hokkaido, Japan, pp. 259–264. (In Japanese.)
Yamamoto S and Suzuki Y (1984) Liquefaction of ground and its prevention – the Southern Noshiro Road seismic
damage and restoration works. Foundation Engineering and Equipment 12(7): 44–49 (special issue). (In
Japanese.)
Yasuda S, Ishihara K, Harada K and Shinkawa N (1996) Effect of soil improvement on ground subsidence due to
liquefaction. Soils and Foundations, pp. 99–107 (special issue).
Yasuda S, Harada K, Shinkawa N and Kanemaru Y (2012) Characteristics of liquefaction in Tokyo Bay area by the
2011 Great East Japan earthquake. Soils and Foundations 52(5): 793–810.
Youd TL, Idriss IM, Andrus RD et al. (2001) Liquefaction resistance of soils: summary report from the 1996
NCEER and 1998 NCEER/NSF workshops on evaluation of liquefaction resistance of soils. Journal of the
Geotechnical Engineering Division, ASCE 127(4): 297–313.
Zen K, Katoh H, Miyama S, Adachi M and Osawa K (2001) Liquefaction countermeasures of compaction grouting
– part 2. Proceedings of the 36th Japan National Conference on Geotechnical Engineering, Tokushima, Japan,
2223–2224. (In Japanese.)
Zen K, Nogami T, Matsushita N, Yamamoto R and Taki M (2000) Liquefaction countermeasures of compaction
grouting – part 1. Proceedings of the 35th Japan National Conference on Geotechnical Engineering, Gifu,
Japan, pp. 2411–2422. (In Japanese.)

420

Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

Appendix
Engineering correlations for the design
of foundations
This appendix is a summary of important formulas that are often helpful in analysis.

A.1. Notation in this appendix

a Net area ratio in the CPT

Ac Cone cross-sectional area in the CPT
Adiff Area behind the cone on which the pore pressure acts in the CPT
cu Undrained cohesion
Dr Relative density of the soil
e Void ratio
emax Maximum voids ratio
emin Minimum voids ratio
E Modulus of elasticity (Young’s modulus) (E = 2G(1 + ν) for an elastic material)
Er Rod energy of the SPT
fs Pile shaft resistance; sleeve friction in the CPT
G Shear modulus
Gmax Shear modulus at very low strain or the maximum shear modulus (strain = 1 × 10–4%)
k OCR exponent
K0 Coefﬁcient of earth pressure at rest (K0 = 1 – sin ϕ′ for normally consolidated soils)
K 1 , K2 Correlation factors
N SPT N value; cycle number during soil loading
N60 SPT N value based on 60% rod energy
N70 SPT N value based on 70% rod energy
(N1)60 Corrected SPT N60 value for 1 atm
(N1)70 Corrected SPT N70 value for 1 atm
Nb SPT N60 value close to the pile tip
Nk Bearing capacity factor or ‘cone factor’
N′k Cone factor accounting for the effective overburden pressure

421

Seismic Design of Foundations

Nr Average SPT N60 value of the pile footing radius

Ns SPT N60 value along the pile shaft
pa Atmospheric pressure
qc Cone resistance in the CPT
qu Unconﬁned compressive strength
qt Total cone resistance in the CPT
su Undrained shear strength
t Time
TG Natural period of a soil layer
u Pore water pressure; horizontal displacement
vp Compressional (P) wave velocity
vs Shear wave velocity
vs,30 Shear wave velocity in the ﬁrst 30 m of soil
z Depth
α Correlation factor
γ Shear strain
γc Cyclic shear strength
γwet Unit weight of saturated soil
ν Poisson’s ratio
ρ Density
σ′h Horizontal effective stress (σ′h = K0σ′v)
σ′m Mean effective stress (σ′m = (σ′v + 2 σ′h)/3)
σ′v Vertical effective stress or overburden pressure
τ Shear stress in a horizontal plane
ϕ Internal friction angle of the soil

A.2. Note on engineering correlations

Correlations are links between two observations (also parameters) that are purely empirical but can be used with
conﬁdence within the limits of the observations. In other words, one has to be careful to extrapolate the correlation.

A.3. Standard penetration test (SPT)

The SPT is a common, simple and inexpensive in situ testing method to obtain a ground proﬁle. The test
involves driving a standard thick-walled sample tube into the ground at the bottom of a borehole by blows
from a slide hammer with a standard weight (typically 63.5 kg) and a falling distance of 75 cm. The sample
tube is driven 150 mm into the ground, and then the number of blows needed for the tube to penetrate each
150 mm up to a depth of 450 mm is recorded. The sum of the number of blows required for the second and
third 150 mm of penetration is reported as the SPT blow count value N. The N value from the ﬁrst 150 mm of
the penetration is usually discarded. This is due to the likely damage to the soil layer caused by soil drilling.

422

Seismic Design of Foundations

Engineering correlations for the design of foundations

Figure A.1 (a) Schematic of the SPT. (b) SPT N-value enumeration

Number of blows per 150 mm

of each penetration

150 mm
5
150 mm
10
150 mm
15
Soil
SPT N value = 10 + 15 = 25

(a) (b)

Figure A.1 illustrates the SPT and also SPT N-value enumeration. The test procedure is described in most codes of
practice: for example, the International Standard ISO 22476-3 (ISO, 2004), the European Standard EN 1997-2
(CEN, 2007), the US Standards ASTM D1586 for clayey soil (ASTM, 2011a) and ASTM D6066 for sandy soil
(ASTM, 2011b), and the Australian Standard AS 1289.6.3.1. Depending on the country, the hammer type and
release mechanism used may differ and the amount of energy for a blow vary. There are three different types of
hammer that are used to carry out the SPT: pin, doughnut and safety hammers. Table A.1 shows recommended
correction factors to achieve 60% rod energy based on Seed et al. (1985).

An SPT is typically stopped if the following are not met: (a) 50 blows for any 150 mm increment; (b) 100 blows
to drive the required 300 mm; (c) and ten successive blows produce no penetration.

The SPT N value can be used to estimate the relative density of soils and approximate shear strength
parameters based on correlations and are given below.

Table A.1 Summary of energy ratios for SPT procedures

Country Hammer type Release mechanism Rod energy: % Correction factor

Japan Doughnut Free fall 78 1.3

Doughnut Rope and pulley 67 1.12
USA Safety Rope and pulley 60 1
Doughnut Rope and pulley 45 0.75
Argentina Doughnut Rope and pulley 45 0.75
China Doughnut Free fall 60 1
Doughnut Rope and pulley 50 0.83

Seed et al. (1985)

Pilcon hammers (used in the UK and Hong Kong) develop an energy ratio of about 60%

423

Seismic Design of Foundations

A.4. SPT correlations for sandy soil classiﬁcation and unit weight
The SPT N value can be used in correlations for different soil parameters such as the unit weight of saturated soil
gwet. Bowles (1988) suggested correlations between N value and soil parameters, and these are given in Table A.2.

N values can be inﬂuenced by many factors such as the quality of the equipment, the performance of the test,
the depth of test and the groundwater.

Note that the different values of energy can be converted by using

Er1 N1 = Er2 N2 (A:1)

where Er is the rod energy of the SPT. For example, if a test is carried out at 70% rod energy (N70) and the
measured value is 15, to convert this to N60 (60% rod energy) the equivalent value will be 70 × 15 = 60 × N60,
which gives us 17.5. It is customary to use whole numbers, so we will take the converted N value as 17.

Yoshida (2015) provided a range of soil unit weights based on the N value, using research from Railway
Technical Research (1997), and these are listed in Table A.3.

Table A.2 Empirical values for the unit weight of soil γwet and soil classiﬁcation based on the SPT

Description Very loose Loose Medium Dense Very dense

SPT N70: ﬁne sand 1–2 3–6 7–15 16–30

Medium sand 2–3 4–7 8–20 21–40 >40
Course sand 3–6 5–9 10–25 26–45 >45
γwet: kN/m 3
11–16 14–18 17–20 17–22 20–23

Bowles (1988)

Table A.3 Unit weight of soil

SPT N value Unit weight (kN/m3)

Saturated Buoyant

Sandy soil 50 20 10
30–50 19 9
10–30 18 8
10 17 7
Clayey soil 30 19 9
20–30 17 7
10–20 15–17 5–7
10 14–16 4–6

Yoshida (2015)

424

Seismic Design of Foundations

Engineering correlations for the design of foundations

A.5. SPT correlations for sandy soil for relative density and angle of
internal friction
Table A.4 lists typical values of the voids ratio e for predicting the relative density of soils.

Table A.5 shows the correlation between the N value and the relative density and angle of friction of sandy soils.
Other correlations have been proposed. It is advisable to obtain the parameters to be used in design calculations
based on all of these relationships and to select the most appropriate using engineering judgement.

Table A.4 Maximum (emax) and minimum (emin) voids ratios for predicting the relative density of soil

Source Soil type emax emin

Sowers and Sowers (1961) Uniform sub-angular sand 0.85 0.5

Well-graded sub-angular sand 0.70 0.35
Very well-graded silty sandy gravel 0.65 0.25
Micaceous sand and silt 1.25 0.80
McCarthy (1977) Well-graded ﬁne to coarse sand 0.70 0.35
Uniform ﬁne to medium sand 0.85 0.50
Silty sand and gravel 0.80 0.25
Micaceous sand and silt 1.25 0.80
Lambe and Whitman (1979) Standard Ottawa sand 0.80 0.50
Clean uniform sand 1.00 0.40
Uniform inorganic silt 1.10 0.40
Silty sand 0.90 0.30
Fine to coarse sand 0.95 0.20
Silty sand and gravel 0.85 0.14

Table A.5 Correlation between the SPT N value and the relative density Dr and angle of internal friction of soil ϕ

SPT N value Dr : % ϕ: °

<4 <20 (very loose) <30

04–10 20–40 (loose) 30–35
10–30 40–60 (medium) 35–40
30–50 60–80 (dense) 40–45
>50 80–100 (very dense) >45

Based on Terzaghi and Peck (1967)

425

Seismic Design of Foundations

Figure A.2 Correlation between the internal friction angle ϕ and the SPT N value (Kitazawa et al., 1959)

50
Internal friction angle: ° (ø)

10 Holocene sand
Terrace sand
Tokyo formation sand
0
1 2 3 45 10 20 30 40 50 60 70 80
SPT N value

A.5.1 Empirical relation 1

Kitazawa et al. (1959) suggested an equation to evaluate the internal friction angle of sand based on data from
Japan (Figure A.2)
pffiffiffiffiffiffiffiffiffi
f = 15 + 20N (A:2)

where f is the internal friction angle and N is the SPT N value.

A.5.2 Empirical relation 2

Bowles (1988) suggested an empirical equation to obtain the internal friction angle from the relative density

f = 28° + 15Dr ± 2° (A:3)

where Dr is the relative density of the soil (for 70% use 0.7 in the equation).

A.5.3 Empirical relation 3

Meyerhof (1957) proposed an empirical correlation between the N value, the effective vertical stress and
Dr (%)
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
N
Dr = 21 0 (A:4)
sv + 0:7

where s′v is the vertical effective stress (kgf/cm2).

426

Seismic Design of Foundations

Engineering correlations for the design of foundations

A.5.4 Empirical relation 4

Kulhawy and Mayne (1990) approximated the correlation between the effective overburden pressure s′v, the
N value and the peak drained friction angle f, in the form
0:34
N
f = sin−1 (A:5)
12:2 + 20:3(sv0 /pa )

where pa is the atmospheric pressure.

A.5.5 Empirical relation 5

Shioi and Fukui (1982) gave correlations between N and f (obtained from the Japanese Railway Standards),
which was slightly modified by Bowles (1988) as
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
f= 18(N1 )70 + 15 ðfor roads and bridgesÞ (A:6)

f = 0:36N70 + 27 ðfor buildingsÞ (A:7)

where N70 is the SPT value for the 70% energy ratio and (N1)70 is the SPT value for the 70% energy ratio and
1 atm pressure

A.5.6 Empirical relation 6

Hatanaka and Uchida (1996) expressed the relationship between the soil friction angle and the corrected
N value of a sandy soil in the form
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
f= 15:4(N1 )60 + 20 (A:8)

A.5.7 Empirical relation 7

Dunham (1954) proposed for uniform, angular well-graded sand that
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
f = 25° + 12(N1 )60 (A:9)

and for round well-graded or angular uniform graded soil particles that
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
f = 20° + 12(N1 )60 (A:10)

A.6. SPT correlations with the modulus of elasticity for sandy soil
Due to the uncertainty of estimating the modulus of elasticity E, Kulhawy and Mayne (1990) proposed the
following as an initial approximation

E
= 5N60 ðsand with finesÞ (A:11)
pa

427

Seismic Design of Foundations

E
= 10N60 ðclean normally consolidated sandÞ (A:12)
pa

E
= 15N60 ðclean over-consolidated sandÞ (A:13)
pa

Based on Bowles, E (kPa) can be expressed in relation to the soil type and the degree of consolidation as

E = 500½(N1 )60 + 15 ðnormally consolidated sandÞ (A:14)

E = 250½(N1 )60 + 15 ðsaturated sandÞ (A:15)

E = 300½(N1 )60 + 6 ðsilty sandÞ (A:16)

E = 300½(N1 )60 + 15 ðclayey sandÞ (A:17)

if (N1 )60 ≤ 15, E = 600½(N1 )60 + 6

ðgravelly sandÞ (A:18)
if (N1 )60 > 15, E = 600½(N1 )60 + 6 + 2000

E = 18 000 + 750(N1 )60 ðoverconsolidated sandÞ (A:19)

A.7. SPT correlations for sandy soil for the shear modulus
For cohesionless sands and gravels, the maximum shear modulus Gmax (MN/m2) can be estimated from the
N value according to the following relationships, following Imai and Tonouchi (1982)

Gmax = 14:4N 0:68 (A:20)

Gmax = 7N (A:21)

However, note that the scatter for these relationships is fairly high: see Figure A.3. It is advisable to take the
average of these two relationships in determining the value of Gmax. Wong and Pun (1997) reviewed the data
published by Imai and Tonouchi (1982) and derived a best ﬁt line (see Figure A.3), which is

Gmax = 8:8N 0:86 (A:22)

428

Seismic Design of Foundations

Engineering correlations for the design of foundations

Figure A.3 Correlation between the shear modulus Gmax and the SPT N value for cohensionless soils (Modiﬁed
from Wong and Pun, 1997.)

Upper limit
Gmax(u) = 20 × N0.86 for 2 ≤ N ≤ 200
Simud (1988)
G = 7N
1000

500
Shear modulus, Gmax: MPa

Imai & Tonouchi (1982)

G = 14.4 × N0.86
200

100

20
Lower limit
10 Gmax(l) = 3.8 × N0.86 for 2 ≤ N ≤ 200

5
Average line
Gmax(av) = 8.8 × N0.86 for 2 ≤ N ≤ 200
2

0.2 0.5 1 2 5 10 20 50 100 200 500

SPT N value

A.8. SPT correlations for the shear wave velocity

Ohta and Goto (1978) proposed the following empirical equation to ﬁnd vs, based on the SPT N value, the
depth H of the deposit and the soil type
0 1 0 1
1:00 clay
B C B C
B 1:091 C B fine sand C
B C B C
B C B C
B 1:029 C B medium sand C
0:219 0:230 B C B C
vs = 62:14N H B C for B C (A:23)
B 1:073 C B coarse sand C
B C B C
B C B C
B 1:151 C B sand and gravel C
@ A @ A
1:485 F gravel F

429

Seismic Design of Foundations

where ‘F’ indicates the soil type. This has been derived for data collected mostly from alluvial plains in Japan.

The Japanese highway code relations for vs (m/s) and the N value are (JRA, 2002)

vs = 80N 1/ 3 ðsandsÞ (A:24)

vs = 100N 1/ 3 ðclaysÞ (A:25)

Table A.6 lists some correlations between vs and the N value.

Table A.6 Correlation between the shear wave velocity vs (m/s) and the SPT N value

Source Soil type Correlation

Imai (1977) All soils vs = 91N0.337

Sand vs = 80.6N0.331
Clay vs = 80.2N0.292
Ohta and Goto (1978) All soils vs = 85.35N0.348
Seed and Idriss (1981) All soils vs = 61.4N0.5
Sykora and Stokoe (1983) Sand vs = 100.5N0.29
Okamoto et al. (1989) Sand vs = 125N0.3
Pitilakis et al. (1999) Sand vs = 145N0.178
Clay vs = 132N0.271
Kiku et al. (2001) All soils vs = 68.3N0.292
Jafari et al. (2002) Sand vs = 22N0.77
Clay vs = 27N0.73
Hasancebi and Ulusay (2007) All soils vs = 99N0.309
Sand vs = 90.82N0.319
Clay vs = 97.89N0.269
Dikmen (2009) All soils vs = 58N0.39
Sand vs = 73N0.33
Silt vs = 60N0.36
Clay vs = 44N0.48

430

Seismic Design of Foundations

Engineering correlations for the design of foundations

A.9. SPT correlations for the shear strength of clay

An unconﬁned compression test is usually used to obtain the shear strength of clay (as the undrained
strength su – also known as cohesion), from which the unconﬁned compressive strength qu is obtained
using

qu
su = (A:26)
2

Terzaghi and Peck (1948) related qu (kPa) and the SPT N value through

qu = 12:5N (A:27)

Kitazawa et al. (1959) proposed the following based on the Japanese data (qu in kPa)

qu = 40 + 5N (A:28)

Kulhawy and Mayne (1990) suggested a rough estimate of the undrained shear strength as

cu 0:72
= 0:29N60 (A:29)
pa

where cu is the undrained cohesion (qu = 2cu). However, sometimes Equation A.29 gives unrealistically high
estimates for cu.

Jamiolkowki et al. (1985) suggested that for overconsolidated clays of low to moderate Plasticity Index (PI)

cu
= (0:23 ± 0:04)OCR0:8 (A:30)
sv0 oc

where s′v is the effective overburden pressure and OCR is the overconsolidation ratio. Equation A.30 can be
approximated for these overconsolidated clays as

cu
= 0:23 ± 0:04 (A:31)
sp0 oc

where s p′ is the maximum past vertical effective stress. Figure A.4 shows the correlation between the N value
and qu, and Table A.7 lists that for the N value and cu.

431

Seismic Design of Foundations

Figure A.4 Correlation between the unconﬁned compression strength qu and the SPT N value (Yoshida, 2015)

5000
Holocene clay (Port region)
Pleistocene clay (Port region)
2000
Honmaki mudstone
1000 Pleistocene clay (Inland)
Takenaka and Nishigaki
500

ck
q0, σl – σ3: kPa

Pe
d
an
hi
200

ag
rz
Te
100

Ohsaki

20
1 2 5 10 20 50 100
SPT N value

Table A.7 Correlation between the SPT and undrained cohesion cu

Soil consistency (N1)60 cu: kPa

Very soft 0–2 <20

Soft 3–5 20–50
Normal consistency 6–9 50–100
Stiff 10–16 100–150
Very stiff 17–30 >150

Weltman and Head (1983)

A.10. Correlations for soil properties

This section provides standard correlations for determining various soil parameters. While dealing with case
studies, it may be useful to obtain the average shear wave velocity and then estimate the shear modulus G of
the soil. Based on the stratigraphic proﬁle description, Eurocode 8 (CEN, 2004) gives an indication of the
average value of the shear wave velocity in the ﬁrst 30 m of soil (denoted vs,30). An extract of this information
is given in Table A.8.

432

Seismic Design of Foundations

Engineering correlations for the design of foundations

Table A.8 Extract from Eurocode 8, Part 5

Ground type Description of the stratigraphic proﬁle vs,30: m/s

A Rock or other rock-like geological formation including at most 5 m of weaker >800

material at the surface
B Deposit of very dense sand gravel, or very stiff clay, at least several tens of metres in 360–800
thickness, characterised by a gradual increase of mechanical properties with depth
C Deposit of dense or medium-dense sand, gravel or stiff clay with a thickness from 180–360
several tens to many hundreds of metres
D Deposit of loose-to-medium cohesionless soil (with or without some soft cohesive <180
layers) or predominantly soft-to-ﬁrm cohesive soil

CEN (2004)

Table A.9 lists typical values of soil densities. Table A.10 gives typical values of Poisson’s ratio for different
types of soils. Table A.11 lists typical values of coefﬁcient of earth pressure at rest for different types of soils.

In undrained conditions (e.g. saturated clayey soils) Poisson’s ratio is normally taken as 0.499 (in numerical
analysis).

Table A.9 Typical values of soil densities for different types of soils

Soil type Saturated density: kg/m3 Bulk density: kg/m3 Dry density: kg/m3

Sandy gravel (loose) 2050 2010 1630

Sandy gravel (dense) 2230 2180 1900
Soft clay 1530 1500 840
Stiff clay 1940 1920 1500

Muir Wood (2009)

Table A.10 Typical values of Poisson’s ratio for different types of soils

Soil type Poisson’s ratio

Loose sand and stiff clay 0.35–0.45

Dense sand 0.25–0.30
Saturated clays ≈0.50

433

Seismic Design of Foundations

Table A.11 Typical values of the coefﬁcient of earth pressure at rest K0 for different types of soils

Soil type K0

Dense sand 0.60

Loose sand 0.35
Normally consolidated clay 0.5–0.6

Craig (2004)

Often, we need the modulus of elasticity of the soil E, which can be expressed in terms of the shear wave
velocity vs, the shear modulus G, the soil density r and Poisson’s ratio n. The relationships between these
properties are

E = 2G(1 + n) (A:32)

G = rv2s (A:33)

A.11. SPT correlations for foundation bearing capacity

Poulos (2014) reports the work of Decourt (1995) given in Table A.12 for shallow and deep foundations to
determine bearing capacities

qu = K1 Nr kPa ðshallow foundations: ultimate bearing capacityÞ (A:34)

fb = K2 Nb kPa ðpiles: ultimate base resistanceÞ (A:35)

fs = a(2:8Ns + 10) kPa ðpiles: ultimate shaft resistanceÞ (A:36)

Table A.12 Correlation factors K1 and K2 (after Decourt (1995))

Soil type K1 (shallow footings) K2 (displacement piles) K2 (non-displacement piles)

Sand 90 325 165

Sandy silt 80 205 115
Clayey silt 80 165 100
Clay 65 100 80

434

Seismic Design of Foundations

Engineering correlations for the design of foundations

where Nr is the average SPT (N60) value of the pile footing radius; Ns is the SPT value along the pile shaft;
Nb is the SPT value close to the pile tip; K1 and K2 are the factors in Table A.12; and the factor a = 1 for
displacement piles in all soils and non-displacement piles in clay, and a = 0.5–0.6 for non-displacement
piles in granular soil.

Table A.13 gives the correlation between the pile shaft resistance fs and the SPT N value.

Table A.13 Correlation between the pile shaft resistance fs and the SPT N value with fs = α + βN kN/m2

Pile type Soil type α β Remarks References

Driven Cohesionless 0 2.0 fs is the average value over the shaft Meyerhof (1956), Shioi
displacement and Fukui (1982)
N is the average value along the shaft
Halve for a small displacement pile
Cohesionless 10 3.3 Pile type not speciﬁed Decourt (1982)
and cohesive
50 ≥ N ≥ 3

fs ≤ 170 kN/m2
Cohesive 0 10 Shioi and Fukui (1982)
Cast in place Cohesionless 30 2.0 fs ≤ 200 kN/m2 Yamashita et al. (1987)
0 5.0 Shioi and Fukui (1982)
Cohesive 0 5.0 fs ≤ 150 kN/m2 Yamashita et al. (1987)
0 10.0 Shioi and Fukui (1982)
Bored Cohesionless 0 1.0 Findlay (1984), Shioi and
Fukui (1982)
0 3.3 Wright and Reese (1979)
Cohesive 0 5.0 Shioi and Fukui (1982)
Cohesive 10 3.3 Piles cast under bentonite Decourt (1982)
50 ≥ N ≥ 3
fs ≤ 170 kN/m2
Chalk –125 12.5 30 > N > 15 Fletcher and Mizon (1984)
fs ≤ 250 kN/m 2

435

Seismic Design of Foundations

A.12. Cone penetration test (CPT)

The CPT – initially developed in the 1950s in the Netherlands – is also known as the Dutch cone test, and is
one of the most used offshore site investigation tools. The test consists of pushing an instrumented cone, with
the tip facing down, into the ground at a controlled rate (typically 1.5–2.5 cm/s). The resistance to penetration
is continuously measured. The cone penetrometer is used to measure the cone resistance qc. The CPT rig can
also be instrumented to measure the pore water pressure and shear wave velocity.

Figure A.5 illustrates the CPT for the case of pre-existing pore water pressure. For pre-existing pore water
pressure, the total cone resistance can be calculated using

qt = qc + (1 − a)u (A:37)

where qt is the total cone resistance, qc is the measured cone resistance, u is the pore water pressure and a is
the net area ratio, which can be calculated from

Ac − Adiff
a= (A:38)
Ac
where Ac is the cross-sectional area of the cone and Adiff is the area behind the cone on which the pore pressure
acts. In Figure A.5, fs is the sleeve friction.

There are at least ﬁve different types of cone: mechanical, electric friction, electric piezo, electric piezo-
friction and seismic. There are also three types of CPT that can be used to carry out for on-shore investi-
gations: truck, crawler and trailer mounted.

Table A.14 provides a guide for the correlation between cone resistance and the internal friction angle in ﬁne
sand. It should be noted that this guide is applicable to unaged, uncemented sands up to about 10–15 m depth.

Readers are referred to the excellent text books on CPT testing Lunne et al. (1997).

Figure A.5 Schematic of the CPT

Soil
qt

436

Seismic Design of Foundations

Engineering correlations for the design of foundations

Table A.14 Correlation between the cone resistance qc and the internal friction angle ϕ

Cone resistance, Compaction of SPT N value Relative density, Angle of internal

qc: MPa ﬁne sand Dr: % friction, ϕ: °

<2 Very loose <4 <20 <30

2–4 Loose 4–10 20–40 30–35
4–12 Medium dense 10–30 40–60 35–40
12–20 Dense 30–50 60–80 40–45
>20 Very dense >50 80–100 >45

A.13. CPT correlation for soil properties

Meyerhof (1976) proposed for normally consolidated sand
pffiffiffiffiffi
f = 29° +2:5 qc (A:39)

where f is the internal friction angle of the soil.

The cone resistance can be used to determine the shear modulus of different sands using either of two
relationships. The ﬁrst (Baldi et al., 1989) is

vs = 277(qc )0:13 (sv0 )0:27 (A:40)

where vs is in m/s and qc and s′v are in MN/m2. This relationship was derived from in situ test data for
two natural predominantly silica sands from the same site in Italy. The second relationship (Rix and
Stokoe, 1991) is

Gmax = 1634(qc )0:25 (sv0 )0:375 (A:41)

where Gmax, qc and s ′v are in kN/m . This equation is derived from ﬁeld test results from two Italian
2

sites, three US sites and one calibration chamber test on quartz sand.

The cone resistance can be used to determine the shear modulus of clay using (Mayne and Rix, 1993)

Gmax = 406(qc )0:695 e−1:13 (A:42)

where Gmax and qc are in kN/m2 and e is the voids ratio. This relationship is derived from ﬁeld tests at sites
worldwide.

A.14. Conversion between SPT and CPT data

Many comparative studies have been carried out by various researchers, and it is known that the particle
size distribution has an important inﬂuence on the correlation between SPT and CPT data. The following

437

Seismic Design of Foundations

Table A.15 Correlations between the CPT cone resistance qc and the SPT N value

Soil type qc/N

Silts, sandy silts and slightly cohesive silt–sand mixtures 0.2–0.3

Fine to medium sands and slightly silty sands 0.4–0.5
Coarse sands and sands with some gravel 0.6–1.0
Sandy gravels and gravel 1.1–1.8

correlations (Table A.15) may be used, where the cone resistance qc is in MPa. Bearing capacity theory
indicates that, for cohesive soils, qc should be related to the effective overburden pressure s′v and the
undrained shear strength su in the following way (Sanglerat, 1972)

qc = su Nk + sv0 (A:43)

where Nk is the bearing capacity factor or ‘cone factor’

In practice (e.g. the Fugro CPT system), a modiﬁed expression in which the effect of the overburden pressure
is included in the cone factor is used

qc = su Nk0 (A:44)

For Equations A.43 and A.44, the cone factor must be determined empirically, or be known from correlations
based on previous investigations in the same clay. The value of the cone factor depends on the stress–strain
properties of the clay, and is frequently found to lie in the range 15–20, although it should be noted that values
outside this range have been observed.

A.15. Equations for wave propagation

The equation of motion for a soil element at a semi-inﬁnite level and in an elastic medium is

∂t ∂2u
=r 2 (A:45)
∂z ∂t

where t is the shear stress in a horizontal plane, r is the soil density, u is the horizontal displacement, t is time
and z is the depth.

The shear stress and shear strain can be correlated through the shear modulus G as

∂u
t = G (shear strain) = G (A:46)
∂z

Combining Equation A.45 and A.46, the shear wave velocity vs can be linked with G through density, which gives

∂2 u 2∂ u
2
= vs (A:47)
∂t 2 ∂z2

438

Seismic Design of Foundations

Engineering correlations for the design of foundations

The general solution for Equation A.47 is of the form

z z
u(t, z) = E t + +F t− (A:48)
vs vs

where E and F are arbitrary functions.

A.16. Engineering base layer

The definition of the engineering base layer in the design specifications for Japan required for site response
analysis is defined by the shear wave velocity vs. Representative definitions in Japan are shown in Table A.16.
Yoshida (2015) mentions that the engineering seismic base layer defined for port facilities and for road
bridges seems to correspond to the layers with an SPT N value of 50. However, for buildings, it seems to
indicate layers deeper than N = 50.

A.17. P-wave velocity

The velocity of P waves is given by
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
E(1 − n)
vp = (A:49)
r(1 − 2n)(1 + n)

The typical velocity of P waves is often 1400–1500 m/s.

A.18. Dynamic soil properties

All soils under moderate cyclic loads yield progressively, try to compact, generate positive pore water
pressure, and soften. Collapsible soils such as loose sands can liquefy in the absence of drainage. The fre-
quency of the cyclic loading condition is therefore important, since the pore pressure generated by one full
load cycle may only partially dissipate before the next load cycle.

For clays, not only strength but also stiffness degrades. The amount of stiffness degradation is a function of
the strain level and the number of cycles of loading. Extensive laboratory and ﬁeld studies have clariﬁed many
aspects of the behaviour of clays under such loading (Vucetic and Dobry, 1991). Stiffness degradation has
been correlated with the plasticity index (PI) of the soil. For a particular cyclic strain in the soil, more-plastic
clays exhibit less degradation.

Table A.16 Deﬁnition of the engineering seismic base layer in Japanese design speciﬁcations

Design speciﬁcation vs: m/s

Port facility 300

Road bridge 300–350
Buildings and houses 400
Nuclear power plant 700

439

Seismic Design of Foundations

Figure A.6 shows a graph based on 16 sets of published experimental data from strain-controlled tests on
clays having different PIs and OCRs. In the graph, a PI of zero indicates sand and a PI of 200 represents highly
plastic Mexico City clays, which show little non-linearity up to 0.1% cyclic shear strain. Gmax is the secant
shear modulus at small strains. Figure A.7 shows the effect of degradation on G/Gmax for different numbers of
cycles (N = 1, 10, 100 and 1000) for normally consolidated clays (OCR of 1).

Figure A.6 Relationship between G/Gmax and the cyclic shear strain γc for non-consolidated and overconsolidated clays

1.0

0.8

0.6
G/Gmax

OCR = 1–15
0.4
PI =
200
100
0.2 50
30
15
0
0
0.0001 0.001 0.01 0.1 1 10
Cyclic shear strain, γc: %

Figure A.7 Effect of cyclic stiffness degradation on G/Gmax versus the cyclic shear strain γc for soils having different
PIs and loaded to different numbers of cycles

1.0

0.8 N=1
N = 10
N = 100
N = 1000
0.6
G/Gmax

OCR = 1
0.4
PI =
N=1 200
N = 10 100
0.2 N = 100 50
N = 1000 30
15
0
0.0
0.0001 0.001 0.01 0.1 1 10
Cyclic shear strain, γc: %

440

Seismic Design of Foundations

Engineering correlations for the design of foundations

Element tests are often used to study cyclic soil behaviour under various conditions of one-way and two-way
cyclic loading. For the design of the Auger tension leg platform (TLP) in the Gulf of Mexico, stress-controlled
cyclic undrained direct simple shear tests were carried out (Dutt et al, 1992). Figure A.8 shows the plot of
the cyclic stress ratio versus the logarithm of the cycle number observed in a two-way cyclic test with an average
shear stress of zero. The graph shows that the number of cycles to failure increases rapidly with decreasing
cyclic shear stress. Also, the cyclic shear strain contours asymptotically approach a critical shear stress level
below which the cyclic shear strains are essentially independent of the number of cycles.

When analysing a soil system, three important features of soil behaviour must be considered

■ Soil is a non-linear material, in which the stiffness progressively decreases with increasing shear
stress, until at a sufﬁciently high shear stress level it deforms plastically.
■ Soil is an inelastic material. Therefore, when subject to cyclic loading, it exhibits hysteretic damping
(energy loss), which increases with increasing shear strain.
■ The soil properties, including strength, stiffness and pore water pressure, may be affected by repeated
cycles of load. This is particularly relevant to saturated sands and silts that may suffer a build-up in
pore water pressure with continued load application, possibly resulting in liquefaction.

Five main properties determine the response of a soil–structure system under dynamic load: the bulk density,
stiffness, material damping, strength and degradation due to cyclic loading. The bulk density affects the inertia
of the soil under dynamic loading. The stiffness, which is largely deﬁned by the shear modulus, determines the
natural frequency of the system and the response. Damping is a measure of energy dissipation – the higher the
damping, the greater the energy loss and, therefore, the smaller the response.

Figure A.8 Cyclic stress ratio versus the number of cycles to failure (Auger TLP platform clay) Dutt et al. (1992)

1.0

0.9 5.0%
2.0% 3.0%
Maximum obliquity
0.8
γc = 1.5%
Cyclic shear stress ratio

0.7

0.6

0.5

0.4

0.3
1 10 100 1000
Number of cycles, N

441

Seismic Design of Foundations

Table A.17 Typical soil properties

Soil type Description ϕ′: ° su: kPa Gmax: MPa

Gravel Dense 40 — 230

Sand Loose 30 — 10
Medium dense 32.5 — 20
35 — 55
Silt — 25 — 10
Clay Soft — 25 15
Firm — 50 30
Hard — 100 60
Rock Mudstone — — 2150
Sandstone — — 3100
Granite — — 15 700

A.19. Typical soil properties for preliminary design

When carrying out dynamic analyses, particularly in the early stages of projects, all soil information will not
be available. Table A.17 provides a range of standard parameters that can be used for preliminary design
(in the absence of any other information).

A.20. Stiffness of soil from laboratory tests

Laboratory tests suggest that the maximum shear modulus Gmax can be expressed by

OCRk pffiffiffiffiffiffiffiffiffiffi0
Gmax = 625 pa sm (A:50)
0:3 + 0:7e2

where e is the voids ratio, pa is atmospheric pressure, s m ′ is the mean effective stress, OCR is the
overconsolidation ratio and k is an OCR exponent that varies with the PI as shown in Table A.18. Equation
A.50 can be expressed in another form as

0 0:5
Gmax = 625F(e)OCRk p0:5
a (sm ) (A:51)
where F(e) is a function of voids ratio.

Hardin (1978) proposed that F(e) = 1/(0.3 + 0.7e2), while Jamiolkowski and Lo Presti (2000) suggested
that F(e) = 1/e1.3. Note that G0, pa and s m
′ must all be expressed in the same units.

442

Seismic Design of Foundations

Engineering correlations for the design of foundations

Table A.18 Variation of the OCR exponent k

PI k

0 0
20 0.18
40 0.3
60 0.41
80 0.48
100 0.5

Hardin and Drnevich (1972)

Ishibashi and Zhang (1993) suggested a relationship between shear modulus G and its maximum value Gmax
for ﬁne grained soil

G
= Ksm0 m (A:52)
Gmax

where

0:000102 + n 0:492
K = 0:5 1 + tanh ln (A:53)
g

0:000556 0:4
m = 0:272 1 + tanh ln exp½−0:0145 PI1:3 (A:54)
g

8 9
>
> 0 for PI = 0 >
>
>
> >
>
>
< 3:37 10−6 PI4:404 >
=
for 0 < PI ≤ 0
n= (A:55)
>
> 7:0 10−7 PI1:976 for 15 < PI ≤ 70 >
>
>
> >
>
>
: >
;
2:7 10−5 PI1:115 for PI > 70

g is the shear strain and PI is the plasticity index of the soil.

For cohesive soils, the relationship between G0/su, the PI and the OCR as summarised by Weiler (1988) can
be used (Table A.19).

443

Seismic Design of Foundations

Table A.19 G0/su values

PI: % OCR
1 2 5

15–20 1100 900 600

20–25 700 600 500
34–45 450 380 300

Weiler (1988)

The OCR can be estimated by relating the measured undrained shear strength to the normally consolidated
undrained shear strength, where

(su )oc
= OCR0:8 (A:56)
(su )NC

and the normally consolidated shear strength can be approximated by

(su )NC ≈ 0:25sv0 (A:57)

where s′v is the effective overburden pressure.

A.21. Natural period of the ground

The natural period of the ground TG can be obtained using

X thickness
TG = 4 (A:58)
layers
vs

where vs is the shear wave velocity (m/s).

A.22. Shear strength of liqueﬁed soil

In 1987, Seed proposed a relationship between the liqueﬁed shear strength and the SPT (N1)60 value, which
was later updated by Seed and Harder (1990) (Figure A.9).

Pillai and Salgado (1994) and Olson and Stark (2002) subsequently published updated relations. Olson and
Stark (2002) suggested

su (liq)
= 0:03 + 0:0075½(N1 )60 ± 0:03 ð ðN1 Þ60 in blows / 0:3 mÞ (A:59)
sv0

444

Seismic Design of Foundations

Engineering correlations for the design of foundations

Figure A.9 Undrained strength of liqueﬁed soil

100

Measured SPT and residual strength data

Undrained residual strength: kPa

Estimated SPT and residual strength data
Construction-induced liquefaction -
estimated data
60

0
0 4 8 12 16 20 24
Equivalent clean sand SPT blow count

Pillai and Salgado (1994) proposed

su (liq)
= 0:21 (A:60)
sv0

A.23. Properties of rock

The value of Gmax and the method of calculation depends on the strength of the rock. It is therefore
preferable to directly measure the shear wave velocity of the rock and hence derive Gmax. However, this is
not always feasible. For weak rocks (unconﬁned strength < 2 MPa) it is possible to determine Gmax as if it
were a soil. For strong rocks (unconﬁned strength > 40 MPa) the value of Gmax is very similar to the static
value, therefore the following formula can be used

jMr qc
Gmax = (A:61)
2 ð1 + n Þ

where j is the mass factor derived from RQD data, Mr is the ratio of the deformation modulus to the
unconﬁned compressive strength qu of the intact rock and qc is the CPT cone resistance (Tomlinson, 1995).

For intermediate rocks, Thompson and Leach (1985) proposed the relationships

Gmax = (4 to 9ÞN ðfor sandstoneÞ (A:62)

Gmax = (3 to 13ÞN ðfor mudstoneÞ (A:63)

Mayne et al. (2002) presented a range of dry unit weights for different types of rock (Table A.20).

445

Seismic Design of Foundations

Table A.20 Dry unit weights for different types of rock

Rock type Unit weight: kN/m3

Shale 20–25
Sandstone 18–26
Limestone 19–27
Schist 23–28
Gneiss 23–29
Granite 25–29
Basalt 20–30

Mayne et al. (2002)

A.24. Calculation of ﬁxity of piles at the non-liqueﬁable dense layer

below the liquefiable layer
Ideally, perfect fixity of a pile in soil can be achieved by grouting into a rock-socket. Fixity at the junction
between liquefiable and non-liquefiable soils below the liquefiable soils is, however, impossible. The base of
a liquefied soil layer is at zero effective stress, and therefore acts temporarily as a free surface for any
underlying soil. Fixity therefore requires pile penetration of several diameters into the dense soil.

Davisson and Robinson (1965) developed an approximate procedure for analyzing partially embedded piles
and this method is suggested in many codes of practices. In the procedure, the partially embedded pile is
assumed to be fixed at some point in the ground, the depth of which depends on the relative stiffness between
the soil and the pile. This method, widely used in practice, involves the computation of stiffness factor T for a
particular combination of pile and soil defined by Equation A.64
sffiffiffiffiffiffi
5 EI
T= (A:64)
hh

where EI is the stiffness of the pile and hh is the modulus of subgrade reaction having units of Force/Length3.

The ratio of the depth to the point of fixity (Dfix) to the pile diameter (d) can be taken as 1.8T for granular soils
whose modulus increases linearly with depth. Figure A.10 summarizes the results from a simple calculation
assuming that the non-liquefiable dense soil beneath the liquefiable layer is sand having 80% relative density.
The modulus of subgrade reaction is taken as 40 MN/m3, following the API (2000) code of practice. The
figure shows that the fixity can be achieved by embedding the pile to a depth equivalent to 3 to 5 times the pile
diameter.

446

Seismic Design of Foundations

Engineering correlations for the design of foundations

Figure A.10 Depth required for ﬁxity

6
Concrete pile
Steel tubular pile
5

Ratio of required embedment for

pile fixit (Dfix) to file diameter (d)
4

0
0 0.5 1 1.5 2 2.5 3
Pile diameter (d): m

REFERENCES
API (2000) “Recommended Practice for planning designing and constructing Fixed Offshore Platforms - Working
Stress Design”. American Petroleum Institute, USA.
ASTM (American Society for Testing and Materials) (2011a) ASTM D1586. Standard test method for standard
penetration test (SPT) and split-barrel sampling of soils. ASTM, West Conshohocken, PA, USA.
ASTM (2011b) ASTM D6066. Standard practice for determining the normalized penetration resistance of sands for
evaluation of liquefaction potential. ASTM, West Conshohocken, PA, USA.
CEN (Comité Européen de Normalisation) (2004) EN 1998-5. Design of structures for earthquake resistance.
Part 5: Foundations, retaining structures and geotechnical aspects. CEN, Brussels, Belgium.
CEN (2007) EN 1997-2. Geotechnical design. Part 2: Ground investigation and testing. CEN, Brussels, Belgium.
Baldi G, Bellotti R, Ghionna V, Jamiolkowski M and Lo Presti DCF (1989) Modulus of Sands from CPT’s and
DMT’s. Proceedings of the 12th International Conference on Soil Mechanics and Foundation Engineering,
Rio de Janeiro, Brazil, vol. 1, pp. 165–170.
Bowles JE (1988) Foundation Analysis and Design, 4th edn. McGraw-Hill, New York, NY, USA.
Craig RF (2004) Craig’s Soil Mechanics, 7th edn. Spon Press, London, UK.
Davisson MT and Robinson KE (1965) Bending and buckling of partially embedded piles. Proceedings of the 6th
ICDMFE 2: 243–246.
Decourt L (1982) Prediction of the bearing capacity of piles based exclusively on N values of the SPT.
Proceedings of the Second European Symposium on Penetration Testing (ESOPT II), Amsterdam, the
Netherlands, vol. 1, pp 29–34.
Decourt L (1995) Prediction of load–settlement relationships for foundations on the basis of the SPT-T.
Proceedings: Ciclo de Conferencias Internationales. ‘Leonardo Zeevaert’, Mexico City, Mexico, pp. 85–104.
Dikmen U (2009) Statistical correlations of shear wave velocity and penetration resistance for soils. Journal of
Geophysics and Engineering 6(1): 61–72.
Dunham JW (1954) Pile foundation for buildings. Proceedings of the ASCE Soil Mechanics and Foundations
Division 80(1): 1–21.

447

Seismic Design of Foundations

Dutt RN, Doyle EH and Ladd RS (1992) Cyclic behavior of a deepwater normally consolidated clay. Proceedings
of Civil Engineering in the Oceans, College Station, TX, USA.
Findlay JD (1984) Discussion. In Piling and Ground Treatment. Thomas Telford, London, UK, pp. 189–190.
Fletcher MS and Mizon DH (1984) Piles in chalk for Orwell bridge. In Piling and Ground Treatment. Institution of
Civil Engineers. Thomas Telford, London, pp. 203–209.
Hardin BO (1978) The nature of stress stain behavior of soils. Earthquake Engineering and Soil Dynamics, ASCE
1: 3–90.
Hardin BO and Drnevich VP (1972) Shear modulus and damping in soils: measurement and parameter effects.
Journal of Soil Mechanics and Foundations Division, ASCE 98(6): 667–692.
Hasancebi N and Ulusay R (2007) Empirical correlations between shear wave velocity and penetration resistance
for ground shaking assessments. Bulletin of Engineering Geology and the Environment 66(2): 203–213.
Hatanaka M and Uchida A (1996) Empirical correlations between penetration resistance and internal friction angle
of sandy soils. Soils and Foundations 36(4): 1–10.
Imai T (1977) P- and S-wave velocities of the ground in Japan. Proceedings of the 9th Meeting of the International
Society for Soil Mechanics and Geotechnical Engineering, Tokyo, Japan, vol. 2, pp. 257–260.
Imai T and Tonouchi K (1982) Correlation of N-value with S-wave velocity and shear modulus. Proceedings of the
2nd European Symposium of Penetration Testing, Amsterdam, the Netherlands, pp. 57–72.
Ishibashi I and Zhang X (1993) Unified dynamic shear moduli and damping ratios of sand and clay. Soils and
Foundations, JSSMFE 33(1): 182–191.
ISO (International Organization for Standardization) (2004) ISO 22476-3:2005. Geotechnical investigation and
testing – field testing. Part 3: Standard penetration test. ISO, Geneva, Switzerland.
Jafari MK, Shafiee A and Razmkhah A (2002) Dynamic properties of the fine grained soils in south of Tehran.
Journal of Seismology and Earthquake Engineering 4(1): 25–35.
Jamiolkowski M and Lo Presti D (2000) Shear strength of coarse grained soils from in situ tests. A compendium.
Proceedings of the 4th International Geotechnical Engineering Conference, Cairo, Egypt.
Jamiolkowski M, Ladd CC, Germaine JT and Lancellotta R (1985) New developments in field and laboratory
testing of soils. Theme Lecture. 11th International Conference on Soil Mechanics and Foundation Engineering,
San Francisco, CA, USA.
JRA (Japan Road Association) (2002) Specifications for highway bridges. Part V: seismic design. JRA, Tokyo,
Japan. (In Japanese.)
Kiku H, Yoshida N, Yasuda S et al. (2001) In-situ penetration tests and soil profiling in Adapazari, Turkey.
Proceedings of the 15th International Conference on Soil Mechanics and Geotechnical Engineering, TC4
Satellite Conference on Lessons Learned from Recent Strong Earthquakes, Istanbul, Turkey, pp. 259–269.
Kitazawa G, Kitayama K. Suzuki K, Ohkawa H and Ohsaki Y (1959) Tokyo Group Map. Gihodo, Tokyo, Japan.
Kulhawy FH and Mayne PW (1990) Manual on Estimating Soil Properties for Foundation Design. Electric Power
Research Institute, Palo Alto, CA, USA.
Lambe TW and Whitman RV (1979) Soil Mechanics, SI version. Wiley, New York, NY, USA.
Lunne T, Powell JJM and Robertson PK (2014) Cone Penetration Testing in Geotechnical Practice CRC Press.
McCarthy DF (1977) Essentials of Soil Mechanics and Foundations. Prentice Hall, Reston, VA, USA.
Mayne PW and Rix GJ (1993) Gmax-qc relationships for clays. Geotechnical Testing Journal, ASTM 16(1): 54–60.
Mayne PW, Christopher BR and DeJong JT (2002) Manual on Subsurface Investigations. Federal Highway
Administration. Washington, DC, USA.
Meyerhof GG (1956) Penetration tests and bearing capacity of cohesionless soils. Journal of Soil Mechanics and
Foundations Division, ASCE 82(SM1): 1–19.
Meyerhof GG (1957) The ultimate bearing capacity of foundations on slopes. 4th International Conference on Soil
Mechanics and Foundation Engineering, London, UK, vol. 3, pp. 384–386.
Meyerhof GG (1976) Bearing capacity and settlement of pile foundations. Journal of the Geotechnical
Engineering Division, ASCE 102(GT3): 195–228.

448

Seismic Design of Foundations

Engineering correlations for the design of foundations

Muir Wood D (2009) Soil Mechanics: A One-Dimensional Introduction. Cambridge University Press, Cambridge, UK.
Ohta Y and Goto N (1978) Empirical shear wave velocity equations in terms of characteristic soil indexes.
Earthquake Engineering and Structural Dynamics 6: 167–187.
Okamoto T, Kokusho T, Yoshida Y and Kusuonoki K (1989) Comparison of surface versus subsurface wave
source for P–S logging in sand layer. Proceedings of the 44th Annual Conference of the Japan Society of Civil
Engineers, Osaka, Japan, vol. 3, pp. 996–997. (In Japanese.)
Olson SM and Stark TD (2002) Liquefied strength ratio from liquefaction flow case histories. Canadian
Geotechnical Journal 39: 629–647.
Pillai VS and Salgado FM (1994) Post liquefaction stability and deformation analysis of Duncan Dam. Canadian
Geotechnical Journal 3l(6): 951–966.
Pitilakis K, Raptakis D, Lontzetidis KT, Vassilikou T and Jongmans D (1999) Geotechnical and geophysical
description of EURO-SEISTESTS using field and laboratory tests, and moderate strong ground motions.
Journal of Earthquake Engineering 3(3): 381–409.
Poulos HG (2014) Tall building foundations – design methods and applications. Proceedings: Conference XXVII
RNIG, Puerto Vallarta, Mexico.
Railway Technical Research Institute (1997) Design Standard of Railway Facilities and Commentary. Maruzen,
Tokyo, Japan. (In Japanese.)
Rix GJ and Stokoe KH (1991) Correlation of initial tangent modulus and cone penetration resistance. Proceedings of
the 1st International Symposium on Calibration Chamber Testing/ISOCCT1, Postdam, NY, USA (Huang AB (ed.)),
pp. 351–362.
Sanglerat G (1972) The Penetrometer and Soil Exploration. Elsevier, New York, NY, USA.
Seed HB (1987) Design problems in soil liquefaction. Journal of Geotechnical Engineering Division, ASCE 113(8):
827–845.
Seed RB and Harder LF (1990) SPT-based analysis of cyclic pore pressure generation and undrained residual
strength. Proceedings of the H. B. seed Memorial Symposium. BiTech Publishing, Richmond, BC, Canada,
vol. 2, pp. 351–376.
Seed HB and Idriss IM (1981) Evaluation of liquefaction potential of sand deposits based on observations and
performance in previous earthquakes. In situ testing to evaluate liquefaction susceptibility. ASCE Annual
Conference, St Louis, MI, USA, preprint 81-544.
Seed HB, Tokimatsu K, Harder LF and Chung RM (1985) Influence of SPT procedures in soil liquefaction
resistance evaluations. Journal of Geotechnical Engineering, ASCE 111: 1425–1445.
Shioi Y and Fukui J (1982) Application of N-value to design of foundations in Japan. Proceedings of 2nd European
Symposium on Penetration Testing, Amsterdam, the Netherlands, vol. 1, pp. 159–164.
Sowers GB and Sowers GF (1961) Introductory Soil Mechanics and Foundations, 2nd edn. Macmillan, New York,
NY, USA.
Standards Australia (2004) AS 1289.6.3.1. Methods of testing soils for engineering purposes Soil strength and
consolidation tests. Determination of the penetration resistance of a soil. Standard penetration test (SPT)
Standards Australia, Sydney, Australia.
Sykora DE and Stokoe KH (1983) Correlations of in-situ measurements in sands of shear wave velocity.
Soil Dynamics and Earthquake Engineering 20: 125–136.
Terzaghi K and Peck RB (1948) Soil Mechanics in Engineering Practice. Wiley, New York, NY, USA.
Terzaghi K and Peck R (1967) Soil Mechanics in Engineering Practice, 2nd edn. Wiley, New York. NY, USA.
Thompson RP and Leach BA (1989) The application of the SPT in weak sandstone and mudstone rocks.
Proceedings of the Conference on Penetration Testing in the UK, Birmingham, UK. Thomas Telford, London,
UK, pp. 83–86.
Tomlinson MJ (1995) Foundation Design and Construction. Wesley Longman, Harlow, UK.
Vucetic M and Dobry R (1991) Effect of soil plasticity on cyclic response. Journal of Geotechnical Engineering,
ASCE 117(1): 89–117.

449

Seismic Design of Foundations

Weiler WA (1988) Small strain shear modulus of clay. Proceedings of the ASCE Conference on Earthquake
Engineering and Soil Dynamics II: Recent Advances in Ground Motion Evaluation, Park City, UT, USA,
pp. 331–335.
Weltman AJ and Head JM (1983) Site Investigation Manual. Construction Industry Research and Information
Association, London, UK.
Wong ACW and Pun WK (1997) Pilot Study of Effects of Soil Ampliﬁcation of Seismic Ground Motions in
Hong Kong. Geotechnical Consulting Ofﬁce, Hong Kong.
Wright SJ and Reese LC (1979) Design of large diameter bored piles. Ground Engineering 12(8): 47–51.
Yamashita K, Tomono M and Kakurai M (1987) A method for estimating immediate settlement of piles and pile
groups. Soils and Foundations 27(l): 61–76.
Yoshida N (2015) Seismic Ground Response Analysis. Springer, Berlin, Germany.

450

Seismic Design of Foundations

Bhattacharya, Orense and Lombardi

Index
Illustrations (ﬁgures and tables) are comprehensively referred to from the text. Therefore, signiﬁcant items in
illustrations have only been given a page reference in the absence of their concomitant mention in the text
referring to that illustration.

absolute modal combination, 111 base shear (design seismic) calculations for three-storey
acceleration (spectral), 109, 111 building, 115–118
case studies see case studies basements
elastic, 107, 121 shallow foundations, 409–411
input, pile-supported LNG tank in layered soils, 287 on slopes, 335
in performance-based design, 119 battered piles, 282–285
zero-period, 111 beaches and liquefaction, 175
see also peak ground acceleration beam
accelerograms and accelerometers, 66, 67 Euler–Bernoulli, 250
hypothetical nuclear power plant in UK, 123–124 on non-linear Winkler foundation, 125–136,
Port Island (Kobe area), 157–158 266, 276
synthetic, 70–74 bearing capacity (loads), 215–217
active earth pressure, 341–343, 346, 351 failures, 165–167, 215, 217, 233, 238
active length (piles), 255, 255–258 retaining wall, 349
air (pore) in partially saturated soil, settlement due to shallow foundations, 215–217
compressibility of, 233 strip footing on slope, 336
Amatrice earthquake (Japan, 2016), 17, 27 bedrock
AMI Stadium, 405–409 peak bedrock level amplitude (PBRA), 65
analysis (of design/foundations/structures), 103–140 target bedrock spectrum, 70
of ground response see response bending (piles), 266–274
anchored bulkheads, 347 buckling and, 266–274
Asia, South, peak ground acceleration, 47, 70 and dynamics, 257, 258
Assam failure (piles), 255, 256
earthquake (1950), 70 loading causing, 263–264
Saraighat Bridge, 258, 313–316, 318–320, 328–329 settlement and, interactions, 257
attenuation relationships, 39, 40, 41, 43 bending moment, 252, 270, 271, 291
Auger tension leg platform, Gulf of Mexico, 441 inertial, 252, 253, 254, 255, 287, 289
axial loads, pile foundations, 246, 248, 250, 252, kinematic see kinematic bending moment
253, 255 Bhuj earthquake, India (2001), 2, 4, 5, 16, 20–21, 22,
effects on bending response, 264–266 49, 252, 257, 279, 281, 282, 283
Kandla Port (incl. Tower), 21, 22, 242, 257, 281,
backbone curves, empirical, 89 282, 283, 290, 343
base layer (seismic), 82–83 Bihar earthquake (India, 1934), 70
engineering, 83, 91, 287, 439 blasting, 368

451

Seismic Design of Foundations

body waves, 81–82 caisson-supported bridges, 297, 306–307, 311–316,

P waves see P waves 318–320
S waves, 4, 11, 81–82 double D, 307, 311–316, 318
boiling sands, 144, 167, 168, 194, 233, 400, 413 liquefaction effects, 320
bore log data, evaluation of soil parameters from, performances comparisons with pile-supported
130 bridges, 258, 328–329
bored piles, 263, 435 Saraighat Bridge (India), 258, 318–320
boundary-element methods, kinematic bending Calcutta (Kolkata, India), 69–71
moment, 276 Campania earthquake (Italy, 63AD), 13
bridges, 297–330 Canterbury earthquake sequence (CES, 2010-2011),
abridged list of failures, 298 189, 199, 215, 219–221, 405–409
caisson-supported see caisson-supported bridges Christchurch earthquake (New Zealand, 2011), 17,
Hanshin Expressway Bridge, 20, 356, 357 165, 166, 219–221, 340–341, 405–409, 413
Indian Road Congress code (IRC:78) for, 262 ﬁnite element model see ﬁnite-element models
Landing Bridge (New Zealand), 282, 283 strengthening of shallow foundations following, 411
performance comparisons poor, 258, 328–329 cantilever retaining wall, 347, 348, 354, 357
Rio Banano Bridge (Costa Rica), 282, 283 capacity curve, 120
Rio Visacaya Bridge (Costa Rica), 298 case studies/histories and surveys
river see river bridges acceleration, 112–113
Saraighat Bridge (Assam), 258, 313–316, peak ground, 46–47, 70, 96, 98, 99, 325–326
318–320, 328–329 bridge foundation performance, 297–305
Showa Bridge, 14, 17, 18, 112, 113, 258, 281, geotechnical engineering, 17–29
298, 305, 320–326, 328 ground improvement see ground improvement
types of foundations, 305–307 ground movement/motion, 15, 17, 19, 47, 158
buckling (piles), 255, 272–274 ground response analysis, 79, 94–98, 101
bending and see bending Kakkonda hydropower station outlet tunnel damage, 44
buildings, 164–168 liquefaction see liquefaction
codes of practice see codes numerical simulations, 157–160
design seismic base shear calculations for pile foundations, 257, 258, 281
three-storey building, 115–118 raked pile performance, 283
ground improvement effectiveness with retaining wall performance, 354–359
Hyogo-ken Nambu earthquake (1993), 401 settlement see settlement
Tōhoku (East Japan, 2011), 404–405 spectrum-compatible ground motion, 69–74
overturning see overturning cast-in-situ (cast in place) piles, 263, 435
piles cement, 306
bending–buckling interaction, 269–270 deep cement mixing (DCM), 394, 395
buckling load estimation, 273–274 centrifuge tests, 226–230, 231, 235
residential see residential buildings see also VELACS
settlement, case studies, 217–223 Chamoli earthquake, 96, 98
on slopes, 334–341 chemical methods (in ground improvement), 306, 367, 373
uprooting of buildings, 4, 21 Chi Chi earthquake (Taiwan, 1999), 11, 15, 66, 91, 359
built environment, 164–168 China
geotechnical damage, 4 Haicheng earthquake (1975), 298, 326–328
liquefaction effects, 4, 164–168 hazard map, 53
buoyant rise of buried structures, 167 Tangshan earthquake (1976), 298, 302, 326
buried and underground structures Wenchuan earthquake (2008), 4, 298, 299–300
buoyant rise, 167 Christchurch earthquake (New Zealand, 2011), 17, 165,
damage, 4 166, 219–221, 340–341, 405–409, 413

452

Index

circular caissons, 307, 311 cone penetration test (CPP; Dutch cone test), 179,
twin, 307 182, 199–200, 390, 408, 436–438
clays (clayey soils) ground improvement, 378, 390
liquefaction susceptibility, 177–178 standard penetration test and, data conversion
shear modulus (G), 437 between, 437–438
shear strength, standard penetration test consolidation, settlement due to, 232
correlations, 431 constitutive models, 153–155, 160, 170, 199, 236
soft, liquefied soil behaving as, 129 Kawagishi-cho, 159
unit weight, 424 Manzari–Dafalias, 92
coastal zones and liquefaction, 175 containment, reinforcement and, 375
codes and standards (design codes of practice/ continental deposits and liquefaction, 175
building codes) contractive behaviour of soil, 145, 148, 162
EN 1998-1 (Eurocode 8; EC8), 65, 103, 109, COSMOS (Consortium of Organizations for Strong-
115, 118, 190–191, 238, 245, 261, 288, Motion Observation Systems), 69, 96
310, 432, 433 cost (monetary loss), 1, 2
Japanese, 15, 191–193, 221–222, 261–262, 322 Costa Rica earthquake (1991), 282, 283, 298
pile foundations, 261–262 Coulomb theory, 341, 344, 351
response spectra in, 107–110, 120 coupled stiffness, rigid caisson foundations, 315, 316
spectrum-compatible motion, 65–67, 190–192 critical load, Euler, 250, 261, 274
cohesion, undrained, in standard penetration test, critical-state shear modulus, 131, 136
431, 432 critical-state soil mechanics, 137–138
compaction crust, unliquefiable, damage in presence of, 188
dynamic, 366, 371–372, 382–383 cut and fill, 335, 337, 339
grouting, 367, 372–373 cyclic behaviour/response (soil), 148–149, 441
case study, 389–393 localised volumetric strains, 231
piles, 370, 376, 378, 396 prediction, 12
sand see sand compaction piles sand
by vibration, 370, 396 dense, 148–149, 162–163
see also densification loose, 148–149
complete quadratic combination (CQC) method, saturated, 151, 231
111 cyclic displacements, 194, 205
complex frequency response, 138 cyclic loads/loading, 12, 87, 141, 149, 170, 187, 439,
compositional characteristics and liquefaction, 441
174 partially drained, 231
compressibility of pore air in partially saturated soil–structure interactions (SSIs) and, 233
soil, settlement due to, 233 cyclic resistance curve, 152
compression test cyclic resistance ratio (CRR), 178, 179–186, 204
triaxial, 146, 147 cyclic shear strain, 87, 88, 440, 441
unconfined, 431 cyclic shear stress ratio (cyclic stress ratio; CSR), 148,
computation 149, 155, 178, 179, 181, 186, 191, 195, 441
lateral displacements, 198–199 Auger tension leg platform (Gulf of Mexico), 441
seismic hazard, 47
concrete dam(s), 4
reinforced see reinforced concrete Niteko Dam, 164
reinforced damage (geotechnical)
bridges, 299, 302, 304 built environment see built environment
retaining walls, 347, 354, 355, 357 examples/types, 4
unreinforced, retaining walls, 354 MMI scale and descriptions of, 36

453

Seismic Design of Foundations

piles, observed in detailed ﬁeld investigations, dilatancy/dilative behaviour, 142

241–245 sand, 142, 148
with unliquefiable surface layer or crust, dip angle (of fault), 34
188 dip-slip fault, 34
Deans Stand, 405–408 direct integration, implicit and explicit, 138
deep cement mixing (DCM), 394, 395 directionality in Hyogoken Nambu earthquake, 158
deep foundations discrete wavelet transform (DSW), 68
bridges, 305–306 displacement and deformation, 4, 38, 173, 194–200
fundamental frequency estimations, cyclic, 194, 205
316–320 large, 4
stiffness, 308–316 lateral see lateral displacements
standard penetration test, 434 liquefaction-induced, 194–200
strengthening, 414–415 peak, bridges, 325, 326, 328
deep soil mixing, 373 in performance-based design, 119
DEEPSOIL, 63, 95 permanent, 13, 56–58, 194
deformation see displacement and deformation shallow foundations, 217, 227, 230, 231, 235, 236
degree of freedom (DOF) systems, 116 deviatoric, 217, 233
multi (MDOF), 105, 116–117 ultimate lateral displacement, 134
single (SDOF), 106, 109, 110, 111, 120 displacement piles (driven piles), 434, 435
Delaware earthquake (2017), 137 Dobry and O’Rourke (1983) approach (bending
dense layer (non-liquefiable), pile fixity at, below momement in pile), 276–277
liquefiable layer, 446 LNG tanks, 289–290
densification (ground improvement), 143, 366, double D caissons, 307, 311–316, 318
367, 369–373, 377–382 drains (and drainage), 367–368, 375–376
by vibration and compaction, 396 gravel, 375–376, 388–389
worked example, 386–388 vertical, 384–386
densification (natural), 143 drivability studies (piles), 286
density driven (displacement) piles, 434, 435
relative, 425–427 ductility in performance-based design, 119
estimation/prediction, 130, 423, 425–427 duration (of ground motion)
ground improvement methods and, 380, 423 prediction at particular site, 12
typical values for different soil types, 433 quantification, 60
depth (of liquefaction) Dutch cone test see cone penetration test
settlement and, relationship between, 222–223 dynamic(s) (and dynamic soil properties), 439–441
strengthening of pile foundations and effects of, in ground response analysis, 87–89
414 pile-supported structures
weighting in evaluation of liquefaction effects, bridges, 320–328
189 case study of interactions with bending and
depth of fixity, 269, 272 buckling, 257, 258
deterministic hazard analysis, 39–41, 41 dynamic failure, 255
developed countries, impact, 1 dynamic instability, 250–252
developing (emerging) countries, 1, 64 dynamic centrifuge tests, 227
deviatoric soil deformations, 217, 233 dynamic compaction, 366, 371–372, 382–383
dewatering, permanent, 377
Dharahara Tower, 26 earth pressure (seismic), 13, 341–346, 351
differential settlements of buildings different soil types, 434
Canterbury earthquake sequence, 219 ground improvement and, 391
estimation, 233–236 theory, 341–346, 359

454

Index

earth structures, damage, 4 Europe, hazard map, 52

East Japan (Tōhoku) earthquake (2011), 1, 4, 5, European Stable Continental Region, 123
16–17, 21–26, 137, 167, 199, 221–222, 243, explicit direct integration, 138
278, 297, 298, 336–340, 393, 396, 404–405,
415 factor of safety see safety factor
EC8 see Eurocode 8 Fatehgarh dam, 4, 5, 21, 94
Edgecumbe earthquake (New Zealand, 1987), 282 fault(s), 34
effective stress see stress dip-slip, 34
ejecta causing ground loss, 233 reverse, 11, 12, 34, 35, 56, 72
elastic acceleration spectra, 107, 121 rupture parameters in moment magnitude (MW)
elastic critical load, 271 scale, 35, 37
elastic modulus strike-slip, 11, 12, 34, 56, 57, 58
standard penetration test in sandy soil, 427–428 thrust, 34, 71, 72
Young’s, 234, 250, 272, 289 fault movement/motion, 4, 11–12, 38, 39
elastic–plastic constitutive model, 154 consequent damage/destruction, 4, 11–12
elastic rebound, 33 case study, 44
elastic strain energy, 33, 35 ground displacement with, 38
electrokinetic injection, 368 quantification, 56
element tests (laboratory), 170, 441 field investigations, observed pile damage,
cyclic resistance ratio evaluation, 184 241–245
simulations, 154–155 financial (monetary) loss, 1, 2
embedment of piles, 253, 271 fine-grained soils and liquefaction susceptibility,
emerging (developing) countries, 1, 64 177–178
empirical backbone curves, 89 see also sands
empirical charts, 126, 178, 185, 227, 380, 382 finite-difference methods, 83, 187, 199
empirical methods of assessing liquefaction, 181, finite-element (FE) models and methods, 152–153, 199,
185, 186 231, 236
permanent lateral displacements, 196 advanced, 58
settlement, 235, 238, 272 buried pipeline
EN 1998-1 see Eurocode 8 crossing a landslide, 58
energy ratios and standard penetration test, 181, crossing a strike-slip fault, 57
190, 423, 427 kinematic bending moment, 276
engineering (seismology), 33–61, 421–450 liquefaction-induced ground displacements, 199
basic concepts, 33–61 non-linear time domain, 152–153
correlations, 421–450 simple, 58
geotechnical see geotechnical engineering five-storey building on piles, 269–271
structural, 103–140 fixed-base natural frequency, caisson-supported bridge,
engineering base layer, 83, 91, 287, 439 318–319
epicentre, 33 fixity (pile), 446
equivalent linear ground response analysis, 83, depth of, 269, 272
94 at non-liquefiable dense layer below liquefiable
equivalent linear static analysis of multi degree layer, 446
of freedom system, 116 flexibility coefficients, bridges, 316, 317–318, 318,
Euler–Bernoulli beam, 250 319, 320
Euler critical load, 250, 261, 274 flexible pile foundations, 255
Eurocode 8 (EN 1998-1; EC8), 65, 103, 109, 115, bridges, 310
118, 190–191, 238, 245, 261–262, 288, 310 flow failures, 13
standard penetration test and, 190, 432 liquefaction-induced, 164

455

Seismic Design of Foundations

flyovers, Indian Road Congress code (IRC:78) for, Global Seismic Hazard Assessment Program
262 (GSHAP), India/South Asia, 46, 49, 51, 70
footings (and their settlement), 216, 223–230, 238, Gorkha earthquake (Nepal, 2015), 17, 26
409–411 grain size (and liquefaction)
in physical model tests, 223–230 characteristics, 177, 191, 192
shallow footings, 434 distribution and mitigation of liquefaction, 377
strip footings on slope, 336 see also fine-grained soils
force-based design, 111–112, 119, 120 granular materials, 142, 143
piles, 264 loose, 142, 143, 184
Fourier transforms, 68, 287 penetration, 370
free face, permanent lateral displacements, see also grain size; sands
estimation, 196 gravel drains, 375–376, 388–389
frequency gravity retaining wall, 347, 348, 350, 354
complex frequency response, 138 Great East Japan (Tōhoku) earthquake (2011), 1, 4,
fundamental see fundamental frequency 5, 16–17, 21–26, 137, 167, 199, 221–222, 243,
natural see natural frequency 278, 297, 298, 336–340, 393, 396, 404–405, 415
friction, internal, angle of, 425–427, 436, 437 Great Hanshin (Hyogoken Nambu; Kobe) earthquake
Friuli earthquake (Italy, 1976), 34, 275 (1995), 1, 15, 20, 31, 79, 144, 157–159, 162,
Fujii River sand, 155 164, 177, 192, 194, 197, 199, 281, 283, 298,
Fukushima Daiichi nuclear power plant, 16, 17, 301, 354–359, 393–395, 401
21, 21–22 ground
fundamental frequency (of structures), 105, displacement/deformation see displacement and
316–320 deformation
estimating, 105, 316–320 ejecta causing loss of, 233
reduction with pile-supported structures, 252 liquefaction see liquefaction
fundamental period (of structures), 103–104, movement/motion, 38, 63–78
115, 118, 119, 269–270 case studies, 15, 17, 19, 47, 158
design and use of artificial strong motion,
Gangtok (Sikkim), 29, 31 63–78
Gaoyuan Bridge, 298, 300 duration see duration
gel push sampler, 184 Japanese code and, 193, 194
geological and geomorphological conditions and pile foundations and, as soil liquefies, 249
liquefaction, 173, 174, 176 prediction of characteristics at site, 12
geometry of pile foundations, 131, 324, 327 records see records
geosynthetic materials spectrum-compatible see spectrum-compatible
drainage, 376 ground motion
reinforced soil wall, 354–358 strong see strong motion
geotechnical damage, examples/types, 4 type I and II, 193
geotechnical engineering and investigations, 1–32, natural period of see natural period
208–209 oscillation, 165
case studies, 17–29 penetration tests see cone penetration test; standard
complexity, 29–32 penetration test
major project planning and design in seismic profile, p–y curves for liquefied soil obtained from,
areas, 207–210 134–136
major topics, 4–13 response see response
what it covers, 1–2 shaking see shaking
global capacity curve, 120 types, Eurocode 8 on, 433
global instability of retaining walls, 349 see also peak ground acceleration; soil

456

Index

ground improvement/remedial measures, horizontal seismic coefficient, 193, 333, 342, 346, 350,
365–409, 411 354
analysis and design, 377–386 horizontal spectra, 107
case studies, 389–409 houses see residential buildings
method application, 389–395 Hyogoken Nambu earthquake (Kobe; Great Hanshin,
seismic performance following improvement, 1995), 1, 15, 20, 31, 79, 144, 157–159, 162,
395–409 164, 177, 192, 194, 197, 199, 281, 283, 298,
summary of common methods, 366–368 301, 354–359, 393–395, 401
worked examples, 386–389 Port Island, 79, 157–159, 162, 169, 228, 229, 401
groundwater table, lowering, 377 hyperbolic model in ground response analysis, 89
grouting, 366, 367 hypocentre, 33
chemical, 367
compaction see compaction impact grid pattern in dynamic compaction, 383
jet, 367, 373–374 impedance functions
particulate, 367 flexible pile foundations, 310
GSHAP (Global Seismic Hazard Assessment rigid deep foundations, 310
Program), India/South Asia, 46, 49, 51, 70 implicit direct integration, 138
Gujarat, Bhuj earthquake see Bhuj earthquake importance factor of structure, 122
Gulf of Mexico, Auger tension leg platform, 441 in situ soil liquefaction, actual earthquake records as
Gutenberg–Richter model/law, 41, 42, 43 manifestations of, 181–183
in situ stress, increasing, 366, 367, 376–377
Haicheng earthquake (China, 1975), 298, 326–328 inclined ground see slopes
Haneda Airport (Tokyo), ground improvement, inclined piles (raked piles), 282–285
389–393 inclusions, restraining effect through, 396
Hanshin Awaji (Hyogoken Nambu; Kobe) India
earthquake (1995), 1, 15, 20, 31, 79, 144, Assam see Assam
157–159, 162, 164, 177, 192, 194, 197, 199, Bhuj earthquake see Bhuj earthquake
281, 283, 298, 301, 354–359, 393–395, 401 Bihar earthquake (1934), 70
Hanshin Expressway Bridge, 20, 356, 357 Harayana state, ground response analysis, 94–98
Harayana state, ground response analysis, 94–98 Kolkata, 69–71
Hasaki wind farm, 24, 26 peak ground acceleration for locations in, 46–47,
Hawaii (Pahal) earthquake (2017), 137 70
hazards, 38–60 pile foundations, 262
analysis, 39–44 bridges and flyovers, 262
LNG tanks, 286 Shillong earthquake (1897), 70
probabilistic (PSHA), 39, 41, 119, 123, 286 Sikkim earthquake (2011), 29, 31
magnitude related to trends in, 137 Tripura earthquake (2017), 137
primary vs secondary, 136 Uttarakhand state (incl. Chamoli district), 96, 98
uniform hazard spectra, 123–124 West Bengal see West Bengal
hillsides see mountain roads; slopes Indian Road Congress code (IRC:78), 262
historical perspectives, 13–29 Indian Standard (IS 1893), 70, 109, 261, 262, 270
Hōie earthquake (Japan, 1707), 13 Indonesia, Sumatra earthquake and tsunami (2004),
Hokkaido Nansei-oki earthquake (Japan, 1993), 16, 137
397, 400 inelasticity of soil, 87
Hokkaido Toho-oki Earthquake (Japan, 1994), inertial bending moment, 252, 253, 254, 255, 287,
397, 400 289
Holyhead, hypothetical nuclear power plant, inertial loads, 217, 255, 261, 263–264
123–124 infrastructure damage with landslide damage, 4

457

Seismic Design of Foundations

instability Nemuro-Hando-oki (2004), 397

piles see piles Nihonkai-Chubu (1983), 196
retaining wall, global, 349 Niigata see Niigata earthquake
intelligent scaling of strong motion record, 65–67 Niigataken Chuetsu-oki (2007), 166
intensity, 35 Sanriku Haruka-oki (1994), 397
in MMI scale, qualitative description, 36 Tokachi-oki (2003), 397
performance-based design relating to, 119 Tottori-ken Seibu (2000), 169, 397
see also magnitude; size Haneda Airport (Tokyo), ground improvement,
internal friction angle, 425–427, 436, 437 389–393
Ishihara-inspired liquefaction potential index, 189 hazard map, 53
Ishihara–Yoshimine method of settlement peak ground velocity map, 80
estimation, 199–200, 204 shallow foundations, modiﬁcations (Tokyo),
Italian building code (OPCM 3274), 262 413
Italy Japanese Road Association code on lateral
Campania earthquake (63AD), 13 spreading loads, 261
Friuli earthquake (1976), 34, 275) jet grouting, 367, 373–374
L’Aquila earthquake (2009), 16, 34
Northern, earthquake in 2012, 21, 120 k-Net, 69, 92
Reggio Messina earthquake (1908), 14 Kakkonda hydropower station, 44
zonation map for 475-year return period, 120 Kamisu wind farm, 24, 26
Iwate prefecture Kandla Port (incl. Tower), 21, 22, 242, 257, 281,
Nairiku-Hokubu earthquake (1998), 44 282, 283, 290, 343
Tōhoku earthquake (2011), 221 Kanto region (Japan)
Kanto earthquake (1923), 14, 221
Japan Tōhoku earthquake (2011), 221
codes and standards, 15, 191–193, 221–222, Kathmandu and Kathmandu Valley, 17, 26, 28
261–262, 322 Kawagishi-cho apartment complex, 17, 18, 112–113,
earthquakes 159, 159–160, 161, 167, 217, 218
Amatrice (2016), 17, 27 kinematic bending moment, 245, 252, 254, 255,
East Japan/Tōhoku (2011), 1, 4, 16–17, 261–262, 276–281
21–26, 137, 167, 199, 221–222, 243, LNG tank, 287, 289
278, 297, 298, 336–340, 393, 396, Kinematic interactions between ground and piles
397, 404–405, 415 relating to strengthening, 414
Hanshin–Awaji (Hyogoken Nambu; Kobe, kinematic loading, 217, 253, 255, 261, 263–264
1995), 1, 15, 20, 31, 79, 144, KMMH16 station near Mashiki town, 92, 93
157–159, 162, 164, 177, 192, 194, Kobe
197, 199, 281, 283, 298, 301, earthquake (Hyogoken Nambu earthquake, 1995), 1,
354–359, 393–395, 401 15, 20, 31, 79, 144, 157–159, 162, 164, 177,
Hōie (1707), 13 192, 194, 197, 199, 281, 283, 298, 301,
Hokkaido Nansei-oki (1993), 397, 400 354–359, 393–395, 401
Hokkaido Toho-oki (1994), 397, 400 ground improvement, 393–395
Iwate Nairiku-Hokubu (1998), 44 Port Island, 79, 157–159, 162, 169, 228, 229, 401
Kanto, 14, 221 Koceli earthquake (Turkey, 1999), 11, 15
Kumamoto (2016), 5, 17, 29, 79, 92, 93, Kolkata (India), 69–71
137 Kumamoto earthquake (Japan, 2016), 5, 17, 29, 79, 92,
Kushiro-oki (1993), 162, 168, 397, 93, 137
398–400 Kushiro-oki earthquake (Japan, 1993), 162, 168, 397,
Miyagi-ken-oki (1978), 336, 397 398–400

458

Index

laboratory tests assumptions, 84–85

element tests see element tests equivalent, 83, 94
soil stiffness, 442–444 example, 90–91
undrained tests, 145–151 procedure, 84–91
Landing Bridge (New Zealand), 282, 283 liquefaction (soil/ground), 38, 125–136, 141–214
landslides (and subsequent damage), 38 bridges and, 297–330
infrastructure, 4 profiles in collapse examples, 322, 336
pipelines, 57 case histories, 157–160, 161–162, 217–223
Tōhoku earthquake (Japan, 2011), 21, 24, Niigata earthquake (Japan, 1964), 17
336–339 countermeasures, 365–420
L’Aquila earthquake (Italy, 2009), 16, 34 classification, 365
large-diameter sampler, 159, 160 definition/meaning, 241
large-scale shaking table tests, 223, 264 effects of (incl. damage), 38, 187–189, 194–200,
lateral displacements, 196–199 208
cyclic, 194 built environment, 4, 164–168
permanent, 194, 196–199 caisson-supported bridges, 320
ultimate, 134 evaluation, 187–189
lateral loads (and capacity), pile foundations, 253, ground displacements, 194–200
255, 261, 264, 291 ground failure, 38, 164
strengthening and, 414 on modal parameters, 111–112
lateral pressure on retaining walls, increased, 168 soil strength, 13
lateral resistance, 246, 261, 262 factor of safety against see safety factor
initial, 133 investigation of mechanisms of, 145
ultimate, 134 pile foundations and, 341–395
lateral spreading, 13, 20, 165, 208, 227 pile fixity calculation at non-liquefiable dense
calculation, 206 layer below liquefiable layer, 446
evaluation, 209–210 potential
loads on pile due to, 261 assessment/calculation, 173, 182–183, 183, 186,
prevention, 415 187, 190, 191, 192, 202, 209, 210, 392
lateral stiffness geomorphological conditions and, 176
piles, 282, 284 index (LPI), 188–189
rigid caisson foundations, 311, 315, 316 practical aspects (evaluation), 173–215
lateral stress, increasing, 376 planning and design of a major project in
layered soil/media seismic areas, 209–210
natural period of ground, 101 worked example, 202–207
piles and, 245, 254–255 preventing, 415
LNG tanks, 287–290 p–y curves and, 125–136
stiffness of deep foundations and, 308 resistance to, 192–193, 369–377
wave propagation, 81–82 calculation, 192–193
LEAP (Liquefaction Experiments and Analysis cyclic see cyclic resistance curve; cyclic
Projects), 154 resistance ratio
length (of piles), 250, 270 lateral see lateral resistance
active, 255, 255–258 methods of increasing, 369–377
effective, 271 sand see sand
unsupported (bridges), 323, 327 severity number, 189
lime, pressure injected, 368 shallow foundations and, 215–240
limit state design, 357 in situ, actual earthquake records as manifestations
linear ground response analysis, 83, 84–93 of, 181–183

459

Seismic Design of Foundations

susceptibility, 173–178 MATLAB, 69, 85

theory, 141–171 Mercalli intensity scale, modified (MMI) scale, 35
triggering, 176–187 Meriken Park Hotel, ground improvement, 393–395
liquefied natural gas (LNG) tanks on piles, 246, Mexico (city)
285–290 earthquake (1985), 12, 15, 19–20, 81, 101
Lisbon earthquake (Portugal, 1755), 13 soil properties, 87
LNG (liquefied natural gas) tanks on piles, 246, Miaoziping Bridge, 298, 299
285–290 micro-zonation, 20
loads, 216 midspan collapse, 297
bearing capacity see bearing capacity Ministry of Business, Innovation and Employment
cyclic see cyclic loads (MBIE), Canterbury earthquake sequence, 220
estimation, 110 mixing methods
river bridge pile buckling load, 272–275 deep cement mixing (DCM), 394, 395
inertial, 217, 255, 261, 263–264 deep soil mixing, 373
kinematic, 217, 253, 255, 261, 263–264 mix-in-place technique, 367, 373
lateral spreading, Japanese Road Association Miyagi-ken-oki Earthquake (Japan, 1978), 336, 397
code on, 261 modal analysis, 104–105, 110, 138
monotonic, undrained behaviour under, modal combination methods, 111
145–151 modal mass, 110, 116
pile-supported structures, 246, 248, 249, 250, modal parameters, liquefaction effects on, 111–112
253, 255, 261, 262 modal superposition method, 138
modelling, 263–266, 271–274 modelling of pile loading, 263–281
localised volumetric strains, 231 see also constitutive models; finite-element
Loma Prieta earthquake (1989), 31, 72, 73, 75, 76, models; numerical modelling; physical
282, 283 model tests
Long Beach earthquake (USA, 1933), 14 moment magnitude (MW) scale, 35–37
lumped-mass models of soil deposits, 83, 84 monetary loss, 1, 2
lumped spring approach, 308 Mononobe–Okabe (M-O) earth pressure theory,
Luzon earthquake (Philippines, 1990), 218–219, 341–346, 359
222, 298, 302 monotonic loading, undrained behaviour under,
145–151
magnitude Monterey sand, 227
in deterministic hazard analysis, 39 Moro Gulf earthquake (Philippines, 1976), 298, 303
hazard trends related, 137 motion see fault; ground
liquefaction triggering and correction factors mountain roads, 359
for, 185 movement see fault; ground
scales, 35–37 multi degree of freedom (MDOF) systems, 105,
scaling factors, 185 116–117
selected records, 72 multi-linear regression method of estimating lateral
see also intensity; size spreading, 196, 206
Magsaysay Bridge, 298, 302 Mylonakis (2001) strain-transmissibility approach to
Manzari–Dafalias constitutive model, 92 kinematic bending moment analysis, 277–278
Mashiki town, KMMH16 station near, 92, 93
mat foundations, 238, 409 National Center for Earthquake Engineering Research
rigid, 227 (NCEER), 202, 386
mathematical forms/treatment natural frequency
ground response analysis, 89 caisson-supported bridge, 318–319
pile instability, 250–252 pile-supported structures, 291

460

Index

natural period constitutive see constitutive models

bridge pier, 321, 323, 323–324, 326, 327 densification methods, 383
ground, 444 finite-element see finite-element models
layered soil, 101 shallow foundations and settlement, 231, 236
simplified method to obtain, 101
Nemuro-Hando-oki earthquake (Japan, 2004), 397 Onagawa city, 4, 5, 23
Nepal, Gorkha earthquake (2015), 17, 26 Onoue approach, 386, 389
Nevada sand, 227 Open System for Earthquake Engineering Simulation
New Zealand (OpenSees), 92, 187
Canterbury earthquake sequence see Canterbury Oritate 5-chome, 339
earthquake sequence oscillation, ground, 165
Edgecumbe earthquake (1987), 282 Osterberg (piston) samplers, 159, 160, 209
Nihonkai-Chubu earthquake (Japan, 1983), 196, overburden pressure, 190
196, 396, 398 correction, 130, 181, 185–186
Niigata earthquake (Japan, 1964), 14, 17, overconsolidation ratios (OCRs), 87, 431, 440, 442,
112–113, 159–160, 165, 196, 217–218, 443, 444
222, 227, 231, 235, 241, 243, 259, 281, overturning, 4
297, 298, 305, 320–326, 395, 396 pile foundations, moment of, 269
Showa Bridge, 14, 17, 18, 112, 113, 258, retaining walls, 349
281, 298, 305, 320–326, 328 factor of safety against, 353, 354
Niigataken Chuetsu-oki (Japan, 2007), 166
Nishinomiya Bridge, 298, 301 P-delta effect, 271
Niteko Dam, 164 P waves (primary body waves), 4, 81–82
non-dimensional foundation stiffness parameters, velocity, 439
317, 319 Pahal (Hawaii) earthquake (2017), 137
non-linear ground response analysis, 84, 90–91 Panshan Highway Bridge, 298, 304, 326–328
example, 92–93 particulate grouting, 367
non-linear properties of soil, 87 passive (earth) pressure, 343–344, 352, 353
non-linear response history analysis, Paul Kelly Stand, 405–408
accelerograms, 123–124 PBRA (peak bedrock level amplitude), 65
non-linear time domain finite-element analysis, peak bedrock level amplitude (PBRA), 65
152–153 peak displacements, bridges, 325, 326, 328
non-linear Winkler springs and foundations, peak ground acceleration (PGA; peak acceleration),
125–136, 266, 291 12, 60
Northern Italy earthquake (2012), 21, 120 bridges, 325–326
Northridge earthquake (USA, 1994), 15, 72, 73, case studies, 46–47, 70, 96, 98, 99, 325–326
74, 75 peak ground velocity map, Japan, 80
Noshiro City, 196 PEER, 69, 71
nuclear power plant penetration, compaction by, 370
Fukushima Daiichi (Japan), 16, 17, 21, penetration tests see cone penetration test; standard
21–22 penetration test
UK, earthquake hazard for a hypothetical performance, 119–124
plant, 123–124 bridges (in seismic areas), 297
numerical modelling and methods caisson-supported vs pile-supported, 258,
(incl. simulations), 152–160 328–329
bending (piles), 264 improved-ground sites, 395–409
and buckling, 266–274 observed from past earthquakes with pile-supported
case histories, 157–160 structures, 278–281

461

Seismic Design of Foundations

period Poisson’s ratio

natural see natural period bridges and, 311, 313
of structures, fundamental, 103–104, 115, 118, different soil types, 433
119, 269–270 pore air in partially saturated soil, settlement due to
perturbation approach to spectral matching, 67 compressibility of, 233
Philippines pore water pressure, 141, 143–144, 161
Luzon earthquake (1990), 218–219, 222, 298, cone penetration test in measurement of, 436
302 dissipation/relief (of excess), 12, 144, 153, 159,
Moro Gulf earthquake (1976), 298, 303 167, 371, 375, 384, 386, 389, 396, 405, 414
physical model tests, settlement of footings in, shallow foundations, 223, 227, 232, 233, 238
223–230 effective stress and, 141
piles (pile-supported structures), 125–136, generation/development of excess, 12, 87, 147,
241–295, 414–415 148, 149, 153–154, 158–159, 160, 163,
base shear, 118 194, 208, 371, 375, 386, 414
bridges, 320–328 mechanism, 143–144
dynamic behaviour, 320–328 pile foundations, 248, 249, 275
performances comparisons with caisson- shallow foundations, 215, 223, 227, 228,
supported bridges, 258, 328–329 229, 232
countermeasures using port facilities, Kushiro-oki earthquake (Japan, 1993),
compaction piles see sand compaction piles 398
mix-in-place piles, 367 Port Island, 79, 157–159, 162, 169, 228, 229, 401
root piles, 368 Port of Oakland (1991), 282, 283
drivability studies, 286 Portugal, Lisbon earthquake (1755), 13
failures/collapse, 241, 242, 243, 257, 263 preliminary design, 208
dynamic, 255 screening in seismic areas, 208
mechanisms, 241, 245–252, 255, 256, 257, typical soil properties, 442
282 preloading, 376–377
fixity see fixity pressure
instability earth see earth pressure
buckling see buckling lateral (on retaining walls), increased, 168
mathematical treatment, 250–252 overburden see overburden pressure
methods of analysis, 125–136, 263–281 pore water see pore water pressure
modelling of problems, 263–281 pressure injected lime, 368
plastic hinge formation, 241, 242, 243, 253, probabilistic seismic hazard analysis (PSHA), 39,
255, 279, 297 41, 119, 123, 286
raked or battered, 282–285 pseudostatic analysis (slopes and retaining walls),
in seismic areas, 241–245 267, 331–333
standard penetration test and, 434–435 earth pressures, 341–346
strengthening, 414–415 simplified, 267, 331–333, 362
uses, 241 punching settlement, 233
pipeline hazards due to permanent grounds p–y curves, 125–136, 266, 267
deformations, 56–58
piston (Osterberg) samplers, 159, 160, 209
Quirino Bridge, 298, 304
plastic hinge formation piles, 241, 242, 243, 253,
255, 279, 297
plasticity index of soil (PI), 87, 178, 431, 439, raft foundations, 223, 238, 409, 411
443 raked piles, 282–285
plate tectonics see tectonics rapid settlement, 227

462

Index

reality vs code-based response spectra, 120 Richter local magnitude scale, 35

records (of motion) rigid foundations (bridges), 310–311
actual, as manifestations of in situ soil double-D caissons, 311–316
liquefaction, 181–183 rigid mat foundations, 227
of strong motion, 63–67 rigidity, modulus of see shear modulus
quantiﬁcation of PGA for locations in India Rio Banano Bridge (Costa Rica, 1991), 282, 283
with unavailable records, 46–47 Rio Visacaya Bridge (Costa Rica, 1991), 298
selection criteria, 65–67 river bridges (piles), 245, 263, 272–273, 298, 300,
Reggio Messina earthquake (Italy, 1908), 14 302, 305, 306, 311, 320–326
regression analysis of lateral spreading, 196, 206 buckling load estimation, 272–273
reinforced concrete see concrete Showa Bridge, 14, 17, 18, 112, 113, 258, 281,
reinforcement, 366, 367 298, 305, 320–326, 328
and containment, 375 river embankment and Kushiro-oki earthquake
of soil retaining walls, 347, 354–355 (Japan, 1993), 400
Rensselaer Polytechnic Institute (RPI) centrifuge, road damage (and reconstruction), 4
226 Christchurch earthquake, 165
replacement method, 377 Kumamoto earthquake (Japan, 2016), 30
residential buildings/houses mountain road, 359
Bhuj earthquake (India, 2001), 2 rock properties, 445
Christchurch earthquake (New Zealand, 2001), see also bedrock
219, 220, 340–341 rocking stiffness, rigid caisson foundations, 315, 316
Niigata earthquake (Japan, 1964) and the rod energy and standard penetration test, 423, 424
Kawagishi-cho apartment complex, 17, Rokko Bridge, 297, 298
18, 112–113, 159, 159–160, 161, 167, root piles, 368
217, 218 RspMatch and RspMatch2005, 67
on slopes, 336–341 rupture parameters in moment magnitude (MW) scale,
Tōhoku earthquake (Japan, 2011), 221–222, 35, 37
278, 336–340
resonant-column test, 89–90 S waves (secondary body waves), 4, 11, 81–82
response (ground), 79–102 safety factor (against liquefaction; FL), 186–187, 188,
analysis, 79–102, 179 203–204, 393
case studies, 79, 94–98, 101 against overturning, 353, 354
effective stress, 187, 199 against sliding, 352, 354
methods, 83–84 worked example of calculation, 203–204
site see site-response analyses samplers, 159, 160, 181, 184, 209
spectra see spectrum San Andreas fault (California), 33, 34
two types, 31–32 San Fernando earthquake (1971), 14, 34, 72, 74, 76, 196
restraining effect through inclusions, 396 San Francisco earthquake (1906), 14
retaining walls/structures, 331–364 sand (sandy soils/deposits), 142–151
case studies of performance, 354–359 boiling, 144, 167, 168, 194, 233, 400, 413
foundations, 347–349 contractive behaviour, 145, 148, 162
lateral pressure increases, 168 dense, 91, 147, 155, 162–163
pseudostatic analysis see pseudostatic analysis cyclic behaviour, 148–149, 162–163
seismic coefﬁcient, 331–334, 350, 354 dilative behaviour, 142, 148
worked examples, 350–354 liquefaction, 142–151, 154, 161–163, 167, 168, 170
return period, 123 resistance, 149–151
importance factor of structure vs, 122 susceptibility, 177, 178
reverse fault, 11, 12, 34, 35, 56, 72 vertical settlement estimation, 199–200

463

Seismic Design of Foundations

loose, 91, 143, 161–162, 173 seismic earth pressure see earth pressure
cyclic behaviour, 148–149 seismic performance see performance
monotonic behaviour in triaxial compression, seismic zone factor see zone factor
147 semi-empirical methods of liquefaction assessment,
Monterey sand, 227 178
Nevada sand, 227 permanent lateral displacements, 197–198
saturated, 143, 151, 168, 218, 223, 231, 268, Sendai City, 336, 339
375, 441 settlement, 165–167
shear modulus (G), 97, 428, 437 ground improvement and, and the Hyogo-ken
standard penetration test (SPT), 130, 424–428 Nambu earthquake (1993), 401
Toyoura sand, 146, 154 liquefaction countermeasures and, 401, 406, 407,
see also fine-grained soils 413, 414
sand compaction piles (SCP), 370, 376, 378, 380 of pile foundations, 254
performance, 396–405 bending interactions with, 257
worked example, 388 check, 255
Sanriku Haruka-oki (Japan, 1994), 397 strengthening methods with, 414
Saraighat Bridge (Assam), 258, 313–316, 318–320, of shallow foundations, 216–238, 405
328–329 building case studies, 217–223, 405
saturated soil factors affecting, 231–233
partially, settlement due to compressibility of footings see footings
pore air in, 233 mechanisms, 227
sandy soils/deposits, 143, 151, 168, 218, 223, vertical, 21, 194, 199–200, 220
231, 268, 375, 441 worked example of calculation, 204–205
scaling SH wave, 82, 84
centrifuge tests, 226 Shahe Bridge, 298, 302
shaking table tests, 223 SHAKE, 63, 94, 179
strong motion, 65–67 shaking (ground), 33, 47, 119, 144
scenario-based assessment in performance-based in MMI scale, 36
design, 119 vertical component, 120
screening (preliminary) in major project planning see also intensity; magnitude
and design in seismic areas, 207–208 shaking table tests, 223, 230
sedimentary deposits, liquefaction susceptibility, 175 large-scale, 223, 264
Seed–Booker model/approach, 382, 384, 386, 389 shallow foundations, 215–240
seismic areas footings, 434
bridge foundations see bridges standard penetration test, 434
major project planning and design, 207–210 strengthening, 409–413
pile foundations, 241–245 shear failure of piles, 255, 256
design criteria, 253–255 shear modulus (G; modulus of rigidity), 35, 85, 89,
typical foundation design process, 63 90, 94, 130, 131, 136, 276–277
seismic base layer and base shear see base layer; clays, 437
base shear critical state, 131, 136
seismic coefficient, 193, 331–334, 346 engineering correlations, 428, 438, 440, 441,
charts, 346 443
horizontal, 193, 333, 342, 346, 350, 354 initial, 131, 136
retaining walls/structures, 331–334, 350, 354 liquefied soils, 131, 132, 136
slopes, 331–334 sands, 97, 428, 437
standard, 193 standard penetration test, 428
vertical, 342, 350 see also base shear

464

Index

shear strain, 87, 89, 131, 148–149, 159, 162, soil(s), 31–32
197, 235, 438 bore log data evaluated from soil parameters, 130
cyclic, 87, 88, 440, 441 critical-state mechanics, 137–138
time history for real earthquake motion, 94 cyclic behaviour see cyclic behaviour
shear strength, 444 deep, mixing, 373
liquified soil, 444 density see density
standard penetration test, 423 expansion due to reduction in effective stress, 232
clay, 431 improvement see ground improvement
ultimate, 132 inelasticity, 87
shear stress, 87, 132, 145, 146, 149, 151, 190, Kumamoto City soil profile in ground response
193, 438, 441 analysis, 92
shear stress ratio, 151, 193 layered see layered soil
cyclic see cyclic shear stress ratio liquefaction see liquefaction
shear-wave velocity (Vs), 82, 83, 84, 86, 90, 91, loose vs dense, 147
101 pile and, interactions, 263, 266, 276, 290
cyclic resistance ratio and, 183 time history during earthquake, 248
standard penetration test correlations for, plasticity index (PI), 87, 178, 431, 439, 443
429–430 properties (general aspects)
wave propagation and, 438 cone penetration test correlations, 437
Shillong earthquake (India, 1897), 70 dynamic see dynamics
Shinozaki fault surface rupture, 45 standard penetration test correlations,
Showa Bridge, 14, 17, 18, 112, 113, 258, 281, 432–434
298, 305, 320–326, 328 typical properties for preliminary design,
Sichuan (Wenchuan) earthquake (China), 4, 442
298, 299 reinforced, retaining walls, 347, 354–355
Sikkim earthquake (India, 2011), 29, 31 saturated see saturated soil
simulation, numerical see numerical modelling shear modulus see shear modulus
single degree of freedom (SDOF) systems, 106, stiffness see stiffness
109, 110, 111, 120 strength see strength
site-response analyses, 179 structure and, interactions (SSIs), 110
algorithm, 86 beam on non-linear Winkler foundation, 125–136
geotechnical software, 94 settlements due to SSI-induced cyclic
LNG tanks, 286 loading, 233
site-specific seismic hazard studies, LNG tanks, system, soil behaviour in analysis of, 441
286 transitional, 178
size, evaluation, 35–37 undrained see undrained soil
see also intensity; magnitude unit weight see unit weight
slice method in pseudostatic analysis, 332 see also ground
sliding (retaining walls), 348 soil nailing, 368
factor of safety against, 352, 354 solidification, 367, 373–374
slopes/sloping ground (incl. hillsides), 331–364 South Asia, peak ground acceleration, 47, 70
foundation considerations for buildings on, spectrum (spectra)
334–341 acceleration see acceleration
permanent lateral displacements, estimation, horizontal, 107
196 matching, 67–68, 275
pseudostatic analysis see pseudostatic analysis wavelet-based see wavelet-based methods
seismic coefficient, 331–334 response (response spectrum/spectral response),
small-scale shaking table tests, 223 106–113, 138

465

Seismic Design of Foundations

applications of analysis, 110–113 volumetric see volumetric strains

in codes of practice, 107–110, 120 see also shear strain; stress–strain curve;
limitations, 111 stress–strain parameters
LNG tank, 287, 288 strain-transmissibility approach to kinematic bending
uniform hazard, 123–124 moment analysis, 277–278
vertical, 107 stratigraphic profile, Eurocode 8, 432
spectrum-compatible ground motion strength
case study, 69–74 increasing (strengthening), 365, 409–415
code-specified, 65–67 pile foundations, 414–415
spread footings, 216 shallow foundations, 409–413
square root of the sum of the squares (SRSS) partial bearing capacity failure due to loss of,
modal combination, 111 233
stability analysis (various structures), in performance-based design, 119
post-liquefaction strength of soil for, 13 post-liquefaction, 13
see also instability shear see shear strength
standard(s) see codes and standards stress
standard penetration test (SPT), 134, 181–182, effective, 141–142, 147, 149, 153, 155, 157, 161,
422–432 187
cone penetration test and, data conversion ground response analyses and, 187, 199
between, 437–438 sand, 145, 146
cyclic ground displacements and, 197 soil expansion due to reduction in, 232
cyclic resistance ratio (CRR) evaluation, shear see shear stress; shear stress ratio
181–182 in situ, increasing, 366, 367, 376–377
Eurocode 8 and, 190, 432 stress–density (S–D) model, 154, 155, 158
generalised soil profile, 91 stress reduction factor, 179
ground improvement, 390–391, 392 stress–strain curve (liquefied soil), 131–132
sandy soil, 130, 424–428 construction, 133–134
standard seismic coefficient, 193 in cyclic torsion shear test of sand, 150
static analysis, equivalent linear, of multi degree stress–strain parameters for liquefied soil, 136
of freedom system, 116 strike-slip fault, 11, 12, 34, 56, 57, 58
stationary wavelet transforms, 68 strip footing on slope, 336
stiffness strong motion (for design), 63–78
chemical or cement addition causing records of see records
stiffening, 396 selection, 63–78
deep foundations (bridges), 308–316 structures
laboratory tests, 442–444 bridges see bridges
in performance-based design, 119 buried and underground see buried and
stilt support, 335 underground structures
stone columns, 370, 380–382 engineering, 103–140
worked example, 388 enhancement, 369
storage tanks pile foundations see piles
Hokkaido Nansei-oki earthquake (Japan, 1993), settlement see settlement
397, 400 shallow foundations see shallow foundations
LNG (liquefied natural gas) on piles, 246, strength see strength
285–290 tilting, 165–166
storey force, 116 Sumatra earthquake (and tsunami), Indonesia (2004),
strain 16, 137
take-off, 131, 133–134, 136 Superstition Hills earthquake (1987), 163

466

Index

surface layer, unliqueﬁable, damage in presence transitional soils, 178

of, 188 trial wedge method, 345, 346
surface waves, 82 triaxial compression test, 146, 147
synthetic accelerograms, 70–74 Tripura earthquake (India, 2017), 137
tsunamis, 1
Taiwan Chi Chi earthquake (1999), 11, 15, 66, Moro Gulf, 303
91, 359 Sumatra (2004), 16, 137
Taiyo New Town (Japan), 23, 337 Tōhoku earthquake (2001, Japan), 21, 23, 221
take-off strain, 131, 133–134, 136 tunnel, Kakkonda hydropower station, 44
tamping (dynamic compaction), 366, 371–372, Turkey, Koceli earthquake (1999), 11, 15
382–383 twin circular caissons, 307
Tangshan earthquake (China, 1976), 298, 302,
326 UK see United Kingdom
tectonics (plate tectonics), 33 ultimate lateral resistance, 134
India, 71 underground structures see buried and underground
tension leg platform, Auger (Gulf of Mexico), structures
441 undrained cohesion in standard penetration test, 431, 432
three-storey building, design seismic base shear undrained soil
calculations, 115–118 behaviour under monotonic loading, 145–147
thrust fault, 34, 71, 72 laboratory tests, 145–151
tieback walls, 347 strength of undrained liquefied soil, 445
tilting of structures, 165–167, 165–166 uniform hazard spectra, 123–124
time-based assessment in performance-based unit weight, 91, 107
design, 119 for different types of rock, 446
time-domain finite-element (FE) analysis, standard penetration test, 424
non-linear, 152–153 United Kingdom (UK)
time histories, 63, 65, 67, 70–74, 77 hazard map, 55
Kawagishi-cho apartment complex during the nuclear power plant (hypothetical), earthquake
1964 Niigata earthquake, 112, 113, hazard, 123–124
114, 115 United States (USA)
pile–soil interaction during earthquake, 248 Delaware earthquake (2017), 137
pore water pressure development, 158–159, Loma Prieta earthquake (1989), 31, 72, 73, 75,
275 76, 282, 283
Tista Bridge (West Bengal), 311 Long Beach earthquake (1933), 14
Tōhoku earthquake (East Japan, 2011), 1, 4, 5, Northridge earthquake (1994), 15, 72, 73, 74, 75
16–17, 21–26, 137, 167, 199, 221–222, Pahal (Hawaii) earthquake, 137
243, 278, 297, 298, 336–340, 393, 396, San Andreas fault, 33, 34
397, 404–405, 415 San Fernando earthquake (1971), 14, 34, 72,
Tokimatsu–Asaka method of displacement 74, 76, 196
estimation, 194–195, 205 San Francisco earthquake (1906), 14
Tokimatsu–Seed method of settlement estimation, Superstition Hills earthquake (1987), 163
200, 205 Whittier Narrows (1987), 72, 73, 74, 76
Tokyo Wildlife Liquefaction Array (California), 163
Haneda Airport ground improvement, 389–393 uprooting of buildings, 4, 21
shallow foundations, modification, 413 see also overturning
Tottori-ken Seibu earthquake (Japan, 2000), 169, Urayasu City, 415
397 USA see United States
Toyoura sand, 146, 154 Uttarakhand earthquake, 96, 98

467

Seismic Design of Foundations

valley ﬁlls, 336, 337, 338 propagation, 4

VELACS (Verification of Liquefaction equations, 438–439
Analysis by Centrifuge Studies) project, layered media, 81–83
153–154 surface, 82
vertical component of ground shaking, 120 see also shear-wave velocity
vertical drains, 384–386 wavelet-based methods in spectral matching, 67
vertical pile, lateral stiffness, 282 wavelet transforms
vertical seismic coefficient, 342, 350 discrete, 68
vertical settlement, 21, 194, 199–200, 220 stationary, 68
vertical spectra, 107 WAVGEN, 67, 70
vibration, 82 weight see unit weight
ground improvement methods using, 366, Wenchuan (Sichuan) earthquake (China), 4, 298, 299
369, 370, 396, 397, 400 West Bengal, 70
vibro-replacement, 366, 370 Tista Bridge, 311
modes of, 104, 106, 110, 111 West Port, Kushiro-oki earthquake (Japan, 1993), 398
void ratio, 147, 148 Whittier Narrows (USA, 1987), 72, 73, 74, 76
and densification method, 377 width ratio, 235
Volterra series approach to spectral matching, Wildlife Liquefaction Array, 163
67–69 windfarms and Tōhoku earthquake, 24, 26
volumetric strains, 199, 200, 201 Winkler springs and foundations, 57, 125–136, 276, 291
settlement and, 231, 232, 234 non-linear, 125–136, 266, 291

water pressure see pore water pressure Young’s modulus of elasticity, 234, 250, 272, 289
water table, lowering, 377
wave(s), 81–83 zero-period acceleration, 111
body see body waves Zhuacun Bridge, 298, 302
in development of synthetic earthquakes, zone factor, 193
70 Indian cities, 49

468

Seismic Isolation, Energy Dissipation and Active Vibration Control of Structures - 17th World Conference On Seismic Isolation (17WCSI) - S
No ratings yet
Seismic Isolation, Energy Dissipation and Active Vibration Control of Structures - 17th World Conference On Seismic Isolation (17WCSI) - S
1,072 pages
Geotechnical Earthquake Engineering (Kramer 1996)
100% (4)
Geotechnical Earthquake Engineering (Kramer 1996)
673 pages
CI160 Civil Engineer S Handbook of Professional Practice
No ratings yet
CI160 Civil Engineer S Handbook of Professional Practice
374 pages
CivilFEM Theory Manual
No ratings yet
CivilFEM Theory Manual
648 pages
Seismic Isolation For Architects-Routledge (2017)
No ratings yet
Seismic Isolation For Architects-Routledge (2017)
205 pages
Earthquake Engineering Report
No ratings yet
Earthquake Engineering Report
20 pages
5th International Conference On New Developments in Soil Mechanics and Geotechnical Engineering Proceedings of ZM 2022
100% (1)
5th International Conference On New Developments in Soil Mechanics and Geotechnical Engineering Proceedings of ZM 2022
483 pages
Fema - Nehrp - Design Examples Training Materials - Volume 1 PDF
0% (2)
Fema - Nehrp - Design Examples Training Materials - Volume 1 PDF
457 pages
Load Out
100% (2)
Load Out
239 pages
Conflict Management and Negotiation - Team 5
No ratings yet
Conflict Management and Negotiation - Team 5
34 pages
Geotechnical Earthquake Engineering (Kramer 1996)
100% (2)
Geotechnical Earthquake Engineering (Kramer 1996)
673 pages
Geotechnical Earthquake Engineering Kramer 1996
No ratings yet
Geotechnical Earthquake Engineering Kramer 1996
673 pages
DR B K Maheshwari Professor Dept of EQ Engg
No ratings yet
DR B K Maheshwari Professor Dept of EQ Engg
77 pages
Wood 1973 - Earthquake Induced Soil Pressures On Structures
No ratings yet
Wood 1973 - Earthquake Induced Soil Pressures On Structures
327 pages
Plaxis Lectures
100% (2)
Plaxis Lectures
147 pages
Seismic Design Using Structural Dynamics SK Ghosh PDF
100% (2)
Seismic Design Using Structural Dynamics SK Ghosh PDF
196 pages
Topics in Modal Analysis & Parameter Identification,: Brandon J. Dilworth Timothy Marinone Michael Mains
100% (1)
Topics in Modal Analysis & Parameter Identification,: Brandon J. Dilworth Timothy Marinone Michael Mains
181 pages
SP 152
100% (1)
SP 152
260 pages
Sea Oc
100% (2)
Sea Oc
296 pages
ENP Energy Efficient Free Cooling For Data Centers
No ratings yet
ENP Energy Efficient Free Cooling For Data Centers
16 pages
Geotechnical Earthquake Engineering
No ratings yet
Geotechnical Earthquake Engineering
5 pages
Seismic Evaluation and Rehabilitation of Structures
No ratings yet
Seismic Evaluation and Rehabilitation of Structures
505 pages
Seismic Design of Foundations
No ratings yet
Seismic Design of Foundations
12 pages
Time-Saving Design Aids For Reinforced Concrete
100% (2)
Time-Saving Design Aids For Reinforced Concrete
19 pages
Dupré Judith Bridges A History of The World S Most Spectacular Spans Hachette Books Black Dog Le
No ratings yet
Dupré Judith Bridges A History of The World S Most Spectacular Spans Hachette Books Black Dog Le
491 pages
Soil - Structure Interaction
100% (2)
Soil - Structure Interaction
448 pages
FEMA 451 - NEHRP Design Examples 2006
100% (5)
FEMA 451 - NEHRP Design Examples 2006
860 pages
Nhi18031 Lateral Piles
100% (2)
Nhi18031 Lateral Piles
298 pages
RAM Foundation User Manual
100% (1)
RAM Foundation User Manual
122 pages
Seminar Presentation.1
No ratings yet
Seminar Presentation.1
28 pages
Gazetas 1991
No ratings yet
Gazetas 1991
41 pages
Soil-Structure Interaction For Deep Foundations, Excavations and Tunnels
100% (7)
Soil-Structure Interaction For Deep Foundations, Excavations and Tunnels
424 pages
ERD
No ratings yet
ERD
7 pages
DYNA102 Stanadyne Pump
100% (3)
DYNA102 Stanadyne Pump
4 pages
Earthquake Resistant Design of Foundation
100% (1)
Earthquake Resistant Design of Foundation
11 pages
Zenva - Javascript 101 Slides
No ratings yet
Zenva - Javascript 101 Slides
49 pages
AbdelRaheem2015 Article EvaluationOfSoilFoundationStru
No ratings yet
AbdelRaheem2015 Article EvaluationOfSoilFoundationStru
20 pages
Book 3
No ratings yet
Book 3
8 pages
Birla Institute of Technology and Science, Pilani: Pilani Campus AUGS/ AGSR Division
No ratings yet
Birla Institute of Technology and Science, Pilani: Pilani Campus AUGS/ AGSR Division
3 pages
Reinforced Concrete Analysis and Design Ray
No ratings yet
Reinforced Concrete Analysis and Design Ray
556 pages
4 Lecture NDMA ERDofShallowFounds V113 Part-A-Paper-1 - EarthquakeResistantDesignOfFoundations
No ratings yet
4 Lecture NDMA ERDofShallowFounds V113 Part-A-Paper-1 - EarthquakeResistantDesignOfFoundations
11 pages
Mathematical Optimization of Solar Thermal Collectors Efficiency Function Using MATLAB
No ratings yet
Mathematical Optimization of Solar Thermal Collectors Efficiency Function Using MATLAB
5 pages
FHWA Soils and Foundations Excerpt PDF
No ratings yet
FHWA Soils and Foundations Excerpt PDF
110 pages
Suggested Analysis and Design Procedures For Combined Footings and Mats
No ratings yet
Suggested Analysis and Design Procedures For Combined Footings and Mats
21 pages
Florin Pavel Viorel Popa Radu Vacareanu Impact of Long Period Ground Motions On Structural Design
No ratings yet
Florin Pavel Viorel Popa Radu Vacareanu Impact of Long Period Ground Motions On Structural Design
114 pages
Special Considerations and Challenges in Seismic Design of Tall Buildings
No ratings yet
Special Considerations and Challenges in Seismic Design of Tall Buildings
55 pages
Questionnaire Employee Name: Designation: Academic Qualification: Experience
No ratings yet
Questionnaire Employee Name: Designation: Academic Qualification: Experience
4 pages
Earthquake Design Practice For Buildings 2nd Ed
0% (1)
Earthquake Design Practice For Buildings 2nd Ed
32 pages
Contact Us Track Your Order: - 44-330-175-5511 - Europe - Language
No ratings yet
Contact Us Track Your Order: - 44-330-175-5511 - Europe - Language
18 pages
Final Thesis Print
No ratings yet
Final Thesis Print
71 pages
Plaxis Dynamic Loading Example
No ratings yet
Plaxis Dynamic Loading Example
11 pages
Continous Power UPS Selectivity DTC
No ratings yet
Continous Power UPS Selectivity DTC
28 pages
4 Lecture NDMA ERDofShallowFounds V113 Part-A-Paper-1 - Earthquake Resistant Design of Foundations
No ratings yet
4 Lecture NDMA ERDofShallowFounds V113 Part-A-Paper-1 - Earthquake Resistant Design of Foundations
11 pages
Review On Analysis of Multi-Storey Framed Structure Using ETABS in Various Zones Considering Soil Structure Interaction
No ratings yet
Review On Analysis of Multi-Storey Framed Structure Using ETABS in Various Zones Considering Soil Structure Interaction
7 pages
Nist GCR 11-917-15 GM Selection
No ratings yet
Nist GCR 11-917-15 GM Selection
256 pages
Pokétwitch Eng
No ratings yet
Pokétwitch Eng
5 pages
Handlebars
No ratings yet
Handlebars
5 pages
Poulos - Tall Building Foundation. Design Methods & Applications
No ratings yet
Poulos - Tall Building Foundation. Design Methods & Applications
51 pages
Study of Soil Structure Interaction On RCC Building
No ratings yet
Study of Soil Structure Interaction On RCC Building
5 pages
Design Concepts of Substructures
No ratings yet
Design Concepts of Substructures
8 pages
1965 Newmark - Earthquakes On Dams and Embankments
100% (1)
1965 Newmark - Earthquakes On Dams and Embankments
21 pages
Net Zero Tool
No ratings yet
Net Zero Tool
19 pages
Dynamic Soil Properties
No ratings yet
Dynamic Soil Properties
16 pages
Application of The Soil Spring Constant
No ratings yet
Application of The Soil Spring Constant
4 pages
Dynamic Analysis of Machine Foundation-INFO
No ratings yet
Dynamic Analysis of Machine Foundation-INFO
8 pages
OU Diary-2020 Informatica PDF
No ratings yet
OU Diary-2020 Informatica PDF
75 pages
Kramer, S. - Geothechnical Earthquake Engineering
No ratings yet
Kramer, S. - Geothechnical Earthquake Engineering
673 pages
Soil Mechanics
No ratings yet
Soil Mechanics
19 pages
Earthquake Tip: What Is Important in Foundations of Earthquake-Resistant Buildings?
No ratings yet
Earthquake Tip: What Is Important in Foundations of Earthquake-Resistant Buildings?
2 pages
Introduction To Earthquakes: Preface Acknowledgments
No ratings yet
Introduction To Earthquakes: Preface Acknowledgments
6 pages
Special Request/Authorization: For Official Use Only Privacy Sensitive
No ratings yet
Special Request/Authorization: For Official Use Only Privacy Sensitive
1 page
The Seismic Damage Mechanism of Daliang Tunnel by Fault Disl - 2023 - Engineerin
No ratings yet
The Seismic Damage Mechanism of Daliang Tunnel by Fault Disl - 2023 - Engineerin
20 pages
Case Study of A Piled Raft Foundation Constructed Using A Reverse Construction Method and Its Post-Analysis
No ratings yet
Case Study of A Piled Raft Foundation Constructed Using A Reverse Construction Method and Its Post-Analysis
18 pages
Direct Displacement-Based Design of Frame-Wall Structures-Sullivan, Priestley, Calvi-2006
No ratings yet
Direct Displacement-Based Design of Frame-Wall Structures-Sullivan, Priestley, Calvi-2006
34 pages
Chapter 7 Concrete Foundations and Earth-Retaining Structures
No ratings yet
Chapter 7 Concrete Foundations and Earth-Retaining Structures
50 pages
Spesifikasi Rig 450 HP (BMA#06)
No ratings yet
Spesifikasi Rig 450 HP (BMA#06)
21 pages
Soil Structure Interaction Analysis of Multi Storey Building Frame For Seismic Load Using SAP2000
No ratings yet
Soil Structure Interaction Analysis of Multi Storey Building Frame For Seismic Load Using SAP2000
6 pages
S-Frame Quick Start Guide
0% (1)
S-Frame Quick Start Guide
40 pages
BEP 328 - Project Management 8: Negotiating Solutions
No ratings yet
BEP 328 - Project Management 8: Negotiating Solutions
13 pages
Adf Scheme - List of The Colleges and Departments Approved by Aicte SL. NO. Name of The College Name of The Departments
No ratings yet
Adf Scheme - List of The Colleges and Departments Approved by Aicte SL. NO. Name of The College Name of The Departments
1 page
Housekeeping Operations - Chapter 4 Guestroom Cleaning & Maintenance
No ratings yet
Housekeeping Operations - Chapter 4 Guestroom Cleaning & Maintenance
34 pages
Effect of Soil - Structure Interaction Constitutive Models On Dynamic Response of Multi - Story Buildings PDF
No ratings yet
Effect of Soil - Structure Interaction Constitutive Models On Dynamic Response of Multi - Story Buildings PDF
5 pages
Machine Foundatioon
No ratings yet
Machine Foundatioon
12 pages
New Microsoft Word Document (3) BBBB
No ratings yet
New Microsoft Word Document (3) BBBB
85 pages
Sun and Eames in ST of Energy 1995
No ratings yet
Sun and Eames in ST of Energy 1995
16 pages
Test CAE
No ratings yet
Test CAE
10 pages
Earthquake Damage and Seismic Code For Buildings
No ratings yet
Earthquake Damage and Seismic Code For Buildings
10 pages
112 Experiment 3
No ratings yet
112 Experiment 3
3 pages
Vibrations of Soils and Foundations
No ratings yet
Vibrations of Soils and Foundations
3 pages
Foundation Analysis and Design: Michael Valley, P.E
No ratings yet
Foundation Analysis and Design: Michael Valley, P.E
0 pages
Horizon (Ceiling Hung) : Key Features
No ratings yet
Horizon (Ceiling Hung) : Key Features
2 pages
CU-2022 B.sc. (Honours) Zoology Semester-6 Paper-DSE-B (2) - 1 QP
No ratings yet
CU-2022 B.sc. (Honours) Zoology Semester-6 Paper-DSE-B (2) - 1 QP
2 pages
Estimation of A Population Mean
No ratings yet
Estimation of A Population Mean
1 page
Itinerary of Travel
No ratings yet
Itinerary of Travel
4 pages
SYNTHESIS
No ratings yet
SYNTHESIS
2 pages
Boolean Xor Based (K, N) Threshold Visual Cryptography For Grayscale Images
No ratings yet
Boolean Xor Based (K, N) Threshold Visual Cryptography For Grayscale Images
4 pages
Advantage of Using PLC in Industrial Automation
No ratings yet
Advantage of Using PLC in Industrial Automation
2 pages
Hard Work, Determination, and Persistence: 3 Keywords in Life
No ratings yet
Hard Work, Determination, and Persistence: 3 Keywords in Life
2 pages