Intermediate Offshore

Foundations

The late Steve Kay was an independent geotechnical consultant with thirty-
three years’ experience as a principal engineer with Fugro, and over forty-
five years as a geotechnical specialist, mainly in the oil and gas industry, both
with contractors and consultants. His expertise was in shallow and interme-
diate (caisson, bucket, can) foundation design, with extensive worldwide
experience in offshore, nearshore and land engineering. He gave suction
foundation courses and master classes and wrote the commercially available
software package CAISSON_VHM.

Susan Gourvenec is Royal Academy of Engineering Chair in Emerging

Technologies in Intelligent & Resilient Ocean Engineering, and Professor of
Offshore Geotechnical Engineering at the University of Southampton, UK.
Susan is currently Convenor of the International Standardisation
Organisation (ISO) committee responsible for developing industry standards
for marine soil investigation, marine geophysical investigation and offshore
geotechnical design. Susan co-authored Offshore Geotechnical Engineering
(CRC Press, 2011) and co-edited the proceedings of the inaugural and sec-
ond International Symposia on Frontiers in Offshore Geotechnics (ISFOG).

Elisabeth Palix has eighteen years’ experience in offshore geotechnics. She

spent twelve years working for Fugro Geoconsulting before joining EDF
Renouvelables, where she is working on design and installation aspects of
offshore projects. Elisabeth is also a member of the TC 209 (ISSMGE) and
has been involved in several geotechnical R&D projects (e.g. SOLCYP,
SOLCYP+, PISA, Unified CPT-based methods).

Etienne Alderlieste is a senior geotechnical researcher/consultant for

Deltares, where he is working on installation and in-place capacity of inter-
mediate and shallow offshore foundations. Before joining Deltares, Etienne
worked as Senior Geotechnical Engineer at SPT Offshore, where he designed
suction foundations for the oil, gas and offshore wind industry. He has also
installed and reinstalled numerous single-suction anchors and several jacket
structures with suction foundations worldwide.
Intermediate Offshore
Foundations

Steve Kay, Susan Gourvenec,

Elisabeth Palix and Etienne Alderlieste
First edition published 2021
by CRC Press
2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
and by CRC Press
6000 Broken Sound Parkway NW, Suite 300, Boca Raton, FL 33487-2742
© 2021 Steve Kay, Susan Gourvenec, Elisabeth Palix, Etienne Alderlieste
CRC Press is an imprint of Informa UK Limited
The right of Steve Kay, Susan Gourvenec, Elisabeth Palix, and Etienne Alderlieste to be
identified as authors of this work has been asserted by them in accordance with sections 77
and 78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or utilised in any
form or by any electronic, mechanical, or other means, now known or hereafter invented,
including photocopying and recording, or in any information storage or retrieval system,
without permission in writing from the publishers.
For permission to photocopy or use material electronically from this work, access www.
copyright.com or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive,
Danvers, MA 01923, 978-750-8400. For works that are not available on CCC please contact
[email protected]
Trademark notice: Product or corporate names may be trademarks or registered trademarks,
and are used only for identification and explanation without intent to infringe.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging‑in‑Publication Data
Names: Kay, Steve, author. | Gourvenec, Susan, author. | Palix, Elisabeth,
author. | Alderlieste, Etienne, author.
Title: Intermediate offshore foundations / Steve Kay, Susan Gourvenec,
Elisabeth Palix, Etienne Alderlieste.
Description: First edition. | Abingdon, Oxon ; Boca Raton, FL : CRC Press,
2021. | Includes bibliographical references and index.
Identifiers: LCCN 2021000720 (print) | LCCN 2021000721 (ebook) | ISBN
9781138353534 (hardback) | ISBN 9780367706708 (paperback) | ISBN
9780429423840 (ebook)
Subjects: LCSH: Offshore structures--Foundations.
Classification: LCC TC197 .K39 2021 (print) | LCC TC197 (ebook) | DDC
627/.98--dc23
LC record available at https://lccn.loc.gov/2021000720
LC ebook record available at https://lccn.loc.gov/2021000721

ISBN: 978-1-138-35353-4 (hbk)

ISBN: 978-0-367-70670-8 (pbk)
ISBN: 978-0-429-42384-0 (ebk)
Typeset in Sabon
by SPi Global, India
Contents

Preface xiii
Notes from the co-authors xv

1 Introduction 1
1.1 Intermediate foundations 1
1.2 Matching models and data quality for good design 4
1.3 Structure of the book 6

2 Offshore foundation types and mode of operation 7

2.1 Definitions – shallow, intermediate and deep foundations 7
2.2 Modes – shallow, intermediate and deep foundations 9
2.3 Intermediate foundation geometry 20
2.3.1 General 20
2.3.2 Internal stiffeners 20
2.3.3 External stiffeners 21
2.3.4 Protuberances 21
2.3.5 Embedment ratio 21
2.4 Summary of intermediate foundation terms 23

3 Loads 25
3.1 Introduction 25
3.2 Units, sign conventions and reference point 25
3.2.1 Example – units 25
3.3 Structural to geotechnical load conversion 28
3.3.1 Example – load conversion 29
3.4 Geotechnical stresses and strains 30
3.5 Commentary 30

v
vi Contents

4 Marine geology 31
4.1 Geology, sediment types and depositional environments 31
4.2 Lateral variability top layers 31
4.3 Seafloor conditions 33

5 Loading conditions and soil drainage 35

5.1 Introduction 35
5.2 Drained-undrained 36
5.2.1 Non-dimensional velocity – penetration rate 36
5.2.2 Dynamic drainage factor – dynamic loading of
solid piles 37
5.2.3 Laterally loaded pile 38
5.2.4 Generic 39
5.3 Closure 40

6 Hazards, uncertainties and risk minimisation 41

6.1 Introduction and case histories 41
6.1.1 Low penetration resistance during installation in
NC clay 41
6.1.2 Excessive misalignment during installation in
NC clay 44
6.1.3 Cylinder buckling during installation 46
6.1.4 Sand plug liquefaction during installation 46
6.1.5 Underpressures close to/above critical during
installation in competent sands 48
6.1.6 Excessive scour during operation 48
6.1.7 Anchor chain trenching during operation 48
6.2 Hazards 49
6.3 Uncertainties 51
6.3.1 Geotechnical data 51
6.3.2 Geotechnical design 52
6.4 Risk minimisation 52
6.4.1 Ground investigation 52
6.4.2 Geotechnical data 53
6.4.3 Laboratory testing 53
6.4.4 Geotechnical design 53
6.5 Closure 53
Contents vii

7 Investigation programs 55
7.1 Introduction 55
7.2 Desk study 55
7.3 Geophysical and geotechnical 56
7.4 Laboratory 58

8 Design basis 59
8.1 General principles 59
8.1.1 Introduction 59
8.1.2 Installation/retrieval/removal 61
8.1.3 In-place resistance 63
8.1.4 In-place resistance – non-co-planar MH loads 64
8.1.5 In-place response 68
8.2 Sign conventions, nomenclature and reference point 68
8.3 Foundation stiffness and fixity 69
8.3.1 Seafloor VHMT loads 71
8.3.2 Foundation lateral and rotational fixity 71
8.4 Load and material factors 72
8.5 Commentary 73
8.6 Closure 73

9 Installation, retrieval and removal 79

9.1 Introduction 79
9.2 General considerations – suction assistance 79
9.3 General considerations for monopile installation – impact
driving, vibratory and drilled and grouted 81
9.4 Best and high estimates – installation resistance 85
9.5 Under-penetration and over-penetration 85
9.5.1 General 85
9.5.2 Under-penetration 87
9.5.3 Over-penetration 87
9.6 Difficult soil profiles 88
9.7 Maximum pump underpressure 90
9.8 Penetration resistance assessment 90
9.8.1 Introduction 90
9.8.2 CPT qc Method 91
9.8.3 CPT method coefficients kp and kf 92
viii Contents

9.8.4 CPT method coefficient αu 95

9.8.4.1 Commentary – steady state 101
9.8.5 Classical bearing capacity method 103
9.9 Landing on seafloor, minimum self-weight penetration
and free-fall penetration 105
9.9.1 Landing on seafloor 105
9.9.2 Minimum self-weight penetration 106
9.9.3 Free-fall penetration 107
9.9.3.1 Example – pile free-fall 108
9.9.3.2 Example – pile free-fall case A
(suction foundation) 110
9.9.3.3 Example – pile free-fall case B
(OWT monopile) 110
9.9.3.4 Example – self-weight penetration cases
A and B 110
9.9.3.5 Example – commentary 111
9.10 Installation in clay 111
9.10.1 Base failure in clay 111
9.10.2 Plug heave in clay 113
9.10.3 Clay installation/retrieval example 114
9.10.3.1 Example – clay installation and retrieval 117
9.10.3.2 Commentary 118
9.10.4 Friction set-up in clay 120
9.10.5 Boulders in clay 121
9.10.5.1 Forces on boulder and pile tip 122
9.10.5.2 Refusal 126
9.10.5.3 Refusal example 1 – weak clay with
dropstone 126
9.10.5.4 Refusal example 2 – competent clay /
glacial till with boulders 127
9.10.5.5 Foundation tip integrity 128
9.10.5.6 Commentary – suction pile 129
9.10.5.7 Commentary – anchor chain 130
9.11 Installation in sand 131
9.11.1 Sand plug liquefaction, piping and heave in sand 131
9.11.2 Models for sand 134
9.11.3 Sand installation example 135
9.11.3.1 Example – sand installation 139
9.11.3.2 Commentary 142
9.11.4 Friction set-up in sand 142
9.11.5 Back analysis of installation data in sand 143
Contents ix

9.11.6 Observational method in Peck (1969) 144

9.11.7 Observational method in CEN (2004) 145
9.11.8 General 145
9.11.9 Water pocket model 146
9.11.9.1 Example – water pocket model 153
9.11.10 Reverse punch-through failure 158
9.11.10.1 Example – reverse punch-through 160
9.11.10.2 Commentary 162
9.12 Installation in (weak) rock 163
9.12.1 Impact driving 163
9.12.1.1 Pile driving refusal 165
9.12.1.2 Risk of buckling 165
9.12.2 Drilling 168
9.12.2.1 Drive drill drive 168
9.12.2.2 Drilled and grouted 169
9.12.3 Vibratory 171
9.13 Presentation of installation assessment 171
9.14 Retrieval and removal resistance assessments 172
9.14.1 Suction foundations 172
9.14.2 OWT monopiles 172
9.15 Closure 173

10 In-place resistance 175

10.1 Introduction 175
10.2 Loading conditions and soil response 175
10.3 In-place failure modes 175
10.4 Tension cracks and gapping 176
10.5 Maximum axial resistance 176
10.5.1 Failure modes for maximum axial
tensile resistance 176
10.5.1.1 Tensile V loads (anchor foundations) in
“undrained” soil 177
10.5.1.2 Compressive V loads (support foundations)
in “undrained” soil 180
10.5.1.3 Example – maximum axial resistance 181
10.5.2 Undrained (“clay”) soil response 182
10.5.3 Undrained (“sand”) soil response 183
10.5.4 Drained (“sand”) soil response 184
10.5.4.1 Compressive loads 184
x Contents

10.5.4.2 Tensile loads 184

10.5.5 Axial myths 184
10.5.5.1 Bearing capacity factor Nc 184
10.5.5.2 Skin friction and end-bearing 186
10.6 Maximum lateral resistance (support foundations) 187
10.7 Maximum lateral resistance and lug position (anchor piles) 190
10.8 Maximum torsional resistance 194
10.9 Tilt and twist (anchor piles) 195
10.9.1 Example – twist 196
10.10 Resistance under combined VHM(T) loads 197
10.11 In-place resistance analysis methods 198
10.11.1 General 198
10.11.2 Undrained soil response (clay) 198
10.11.3 Drained soil response (sand) 201
10.12 VHM(T) resistance envelope methods 201
10.12.1 General 201
10.12.2 Undrained soil response (“clay”) 201
10.12.2.1 Undrained soil response (“clay”): 201
10.12.2.2 Example seafloor VHM loads 204
10.12.3 MH ellipses 205
10.12.4 V-Hmax ellipsoids 205
10.12.5 VHM envelope – equations 208
10.12.6 VHM envelope – yield function 208
10.12.7 Modifying lateral and Vmax resistance 209
10.12.8 Resistance comparisons 212
10.12.8.1 Support foundations: 212
10.12.8.2 Anchor pile (and chain) foundations: 212
10.12.9 MH ellipse – design examples 214
10.12.10V-Hmax ellipsoid – design example 216
10.12.11 VHM envelope – support foundation design
example 217
10.12.12 VHM envelope – anchor foundation design
example 220
10.12.13Drained soil response (“sand”) 223
10.13 Resistance at shallow penetration 224
10.14 Resistance in (weak) rock 224
10.14.1 General 224
10.14.2 Axial resistance of driven piles in weak rock 224
10.14.3 Axial resistance of drilled and grouted piles in rock 226
10.14.4 Lateral resistance in rock 232
10.15 Group resistance 235
Contents xi

10.15.1 General 235

10.15.2 Considerations 235
10.15.3 Design procedures 235
10.15.4 Design tools 235
10.15.5 Braced support groups 235
10.15.5.1 Shallow foundation groups 235
10.15.5.2 Anchor pile groups 238
10.16 Closure 239
10.16.1 Clay 239
10.16.2 Sand 239

11 In-place response 241

11.1 Introduction 241
11.2 Displacement under static loads 241
11.2.1 General 241
11.2.2 Immediate displacement 243
11.2.3 Primary consolidation settlement 245
11.2.4 Secondary compression (creep) settlement 247
11.2.5 Settlement induced by cyclic loads 247
11.2.6 Regional settlement 248
11.3 Displacement under dynamic and cyclic loads 248
11.4 Response in (weak) rock 250
11.5 Closure 253

12 Miscellaneous design considerations 255

12.1 Introduction 255
12.2 Scour protection 256
12.3 Interaction with adjacent infrastructure 256
12.3.1 Introduction 256
12.3.2 Jack-up spudcans 257
12.3.3 Adjacent intermediate foundations 261
12.3.4 Adjacent shallow foundations and pipelines 261
12.3.5 Conductor wells 261
12.4 Shallow gas 261
12.5 Permanent passive suction 262
12.5.1 Example – permanent passive suction 262
12.6 Top plate vent design 263
12.6.1 Example – top plate vent design 264
12.7 Suction pump design 265
xii Contents

12.7.1 Example – suction pump flow rate 265

12.8 Foundation instrumentation 266
12.9 Steel design 267
12.10 Soil reactions 267
12.11 Underbase grouting 268
12.12 Anchor chain trenching 269
12.12.1 Commentary 274
12.13 Closure 274

Acronyms 275
Notation 277
References 281
Index 301
Preface

The origin of this book is as follows. International standard ISO 19901-

4:2016 deals with geotechnical and foundation design for the offshore
petroleum and natural gas industry. It is due for renewal in 2020. A sub-
group, chaired by Susan Gourvenec, was tasked with providing advice (i.e.
text, figures and references) on intermediate support foundations. This
foundation type is currently not included in ISO 19901-4. In early 2017,
since I am now in the knowledge transfer stage of my professional career,
and a member of her subgroup, I provided Susan with an 80-page document
about suggested advice topics, together with justification for their inclusion.
The document had basically four key sections dealing with installation, in-
place resistance, in-place response and miscellaneous design considerations.
Even if its suggestions were acceptable to the subgroup, it would be inevi-
table that, because of its length, not all the information would appear in the
revised, third edition of ISO 19901-4. Both Susan and I were concerned that
much of the material would disappear forever into either the Recycle Bin
(MS Windows) or Trash (Mac OS).
Hence this book is basically the above document augmented with mate-
rial acquired over the past 40-odd years. It is hoped that this book has
“plugged” the gap existing between shallow and pile foundations.
The book has two main authors. Due to my ill-health, I provided a good
250-page draft to Susan Gourvenec, who then took on the laborious task of
reviewing, editing, improving and getting it into publisher-ready format. In
addition, Elisabeth Palix (EDF-EN) kindly contributed sections on interme-
diate foundations for renewables and weak rock, and Etienne Alderlieste
(Deltares, formerly with SPT) was a reviewer and redrew many of the fig-
ures ready for publication.
Other acknowledgements are necessary. Firstly, to Ian Smith, my research
supervisor, who instilled in me a love of numerical methods in geomechan-
ics, starting with non-linear FEA using 3-noded constant strain triangles.
His book programs have been modified to solve the axisymmetric fluid flow
and laterally loaded caisson problems presented herein. Secondly, for the
transfer of knowledge and practical advice with numerous engineers in
many organisations with whom I worked as a civil/geotechnical engineer,

xiii
xiv Preface

later on as an independent geotechnical consultant. Notable organisations

include Fugro Engineers BV and SPT Offshore BV. Also, thank you to all
CAISSON_VHM users for being sceptical, and not afraid to ask me perti-
nent questions about the dark corners of in-place resistance and associated
intermediate support and anchor foundation topics. Finally, I thank my wife
Angela for putting up with my obsession while preparing this book.
Steve Kay
Leiderdorp
The Netherlands
May 2018
Notes from the co-authors

Susan Gourvenec
One afternoon in late 2017, I got a phone call in my office from Steve. He
said he would like to write a book on intermediate offshore foundations,
based on the materials he had provided to the ISO working group. He
wanted the information and the knowledge to be available for engineers to
use. He wanted to know if I would be keen to get involved. It seemed like a
good idea – so I said yes.
Steve called often to discuss things – always interesting, enlightening and
often amusing. Steve retained his humour during all our conversations even
in the late stages of his illness, and his desire for ‘knowledge transfer’ was
sincere.
Undoubtedly a talented geotechnical engineer, Steve also struck me as a
great enthusiast and citizen of our engineering community. Always generous
with his time and knowledge whether at conferences or committee meetings.
It is telling that he sent apologies for missing a meeting of the Technical
Panel of ISO WG10 just a few days before he passed away. A life cut too
short – but Steve certainly made an impression on offshore geotechnics and
on the people he worked with and interacted with in his professional life.
I know several geotechnical engineers of my generation, who worked
closely with and knew Steve much better than I did. They all speak highly of
his generosity in sharing knowledge and of how much they learnt from him.
Two such individuals were drafted (by Steve) into developing this book –
and I have thoroughly enjoyed the opportunity to work with both Elisabeth
and Etienne in creating ‘Steve’s book’, as it has been referred to over the last
couple of years as we worked on it.
I would like also to acknowledge the support from the publisher Taylor &
Francis, in particular Tony Moore, who enthusiastically took on the book
title and worked with Steve and me to put in a proposal. Thanks also to
Gabriella Williams, always so understanding about our various delays and
missing deadlines, and to Frazer Merritt, who took the book over the finish
line to final publication.

xv
xvi Notes from the co-authors

Elisabeth Palix
I got the chance to work with Steve at the start of my career on few projects
involving suction anchor design. We also got the opportunity to write a few
papers together. Steve was a passionate engineer who could give you a call a
few weeks after your last meeting to tell you that he had some thoughts
about your last discussion. I liked his ability to think out of the box and to
take into consideration the opinion of anyone, even a young inexperienced
female engineer. I remember that we had to put energy at first to convince
some engineers that despite the use of the word pile, anchor piles should not
be designed as a pile.
When Steve contacted me to provide him with some support for his book,
it was the first I heard about his illness. He was in a hurry to transfer the
knowledge he acquired during his career. He was afraid that everything
would be lost and disappear into the “Windows recycle bin” (his own
words).
This book has been the opportunity to strengthen the links between us at
the end of his life. I believe that working on this book gave him some energy
to fight his illness and some peace toward the end. I am very glad that I got
the chance to be part of this story with Susan and Etienne. We tried together
to finish this book as Steve would have done if he was still among us.

Etienne Alderlieste
Not too long after I started working at SPT Offshore, I met Steve. He was
contracted to provide engineering support and gave a course on suction
foundation design. For another project in 2014, Steve was asked to be
involved as reviewer for the detailed design of the suction anchors for an
FPSO in the Gulf of Guinea. Throughout the course of that project we dis-
cussed a lot of subjects, including engineering and New Zealand. He enjoyed
sharing knowledge and simultaneously challenging several design choices,
which he was good at, with the aim to end up with a sane design. The final
design accounted for potentially detrimental effects of the by-then novel
phenomenon known as trenching, without being overly cautious. Not con-
servative, that term belongs in politics.
During the years that followed we stayed in touch, discussing various
engineering matters, personal matters, but also matters concerning
CAISSON_VHM – checking functionality of a newer version, assistance
when a Windows update messed with things. With Steve’s health declining,
our contact intensified. He got me involved in this book, and asked me to
provide technical support for CAISSON_VHM to assist Richard, who
would be taking over.
Susan, Elisabeth and I happily accepted the relatively straightforward
task to “finish the book” Steve had drafted. It proved, however, significantly
more challenging for all of us than just those three words suggest. A big
thanks to Susan and Elisabeth; I really enjoyed our meetings and open
discussions!
Notes from the co-authors xvii

All co-authors
We would also like to acknowledge, Richard Kay, Steve’s son, for seeking
out and providing Steve's Mathcad files for recreating figures and analyses
presented in the book.

It has been a significant task, but also a great privilege to work on this
book. We hope you find it useful.
Susan Gourvenec,
Elisabeth Palix and Etienne Alderlieste
June 2020

If you are interested in using CAISSON_VHM, visit

https://www.casksoftware.com.
Steve Kay
28 November 1946–4 December 2018
Chapter 1

Introduction

1.1 INTERMEDIATE FOUNDATIONS

Intermediate foundation describes a foundation geometry and response that

falls between “shallow” and “deep” foundations. Shallow foundations are
generally taken as having a length to diameter (or breadth) ratio L/D < 0.5,
while deep foundations (or ‘piles’) are taken to have L/D > 10. Arguably
more significant, though, is how the foundation responds under applied
loading. Load interaction (between different load components) is significant
for intermediate foundations, and they behave as essentially rigid when
laterally loaded – distinguishing the response from deep (piled) foundations;
while failure mechanisms involve mobilisation of soil strength both deep
and shallow – distinguishing the response from shallow foundations.
Intermediate offshore foundations may be broadly divided into the
following two types:

• support foundations (point of load transfer at seafloor);

• anchor foundations (point of load transfer usually at depth below
seafloor).

The support foundation type includes use for hydrocarbons – platforms,

subsea templates, manifolds, wellhead and pipeline supports, riser tower
bases; and renewables – wind turbines, transformers and met masts.
The anchor foundation type is mainly used to anchor floating systems or
vessels, as a starter pile for pipeline pulls, or to salvage capsized or sunken
vessels (including detaching the bow of the Kursk nuclear submarine).
Figures 1.1 and 1.2 give commonly encountered support and anchor
foundation examples.
Intermediate foundations have various synonyms, including the terms (in
alphabetical order) “anchor pile”, “braced caisson”, “bucket”, “caisson”,
“can”, “free-standing caisson”, “monopile”, “pile”, “skirted foundation”,
“suction anchor pile”, “suction installed skirted foundation”, “suction
foundation”, and “suction pile anchor”. A comparison with offshore
shallow and deep (pile) foundations is given in Chapter 3.

1
2 Introduction

(a) (b)

Figure 1.1 Intermediate support foundation type – load applied at seafloor:

(a) jacket with suction buckets, Aberdeen Bay offshore wind farm and
(b) monopile for offshore wind turbine.

(a)

(b)

Figure 1.2 Intermediate anchor foundation type – load applied beneath seafloor
level (at lug): (a) suction anchor (Dijkhuizen et al., 2003) and (b) tor-
pedo pile (De Araujo et al., 2004).
Introduction 3

Hogervorst (1980) described the first field trials of full-size suction piles,
nearshore Holland, and Senpere and Auvergne (1982) described the first
application of suction pile anchors for the Gorm Field, offshore Denmark.
Due to their simplicity, reliability and cost savings, suction foundations are
also applied to fixed platforms. Bye et al. (1995) described design and instal-
lation to support a jacket structure in very dense North Sea sands. Kolk
et al. (2001) presented studies for the first lightweight platform (Offshore
Cabinda, West Coast of Africa) founded on soft clays and subjected to ten-
sile leg loading. Andersen et al. (2005) presented a keynote paper on the
application of deepwater suction anchors, drawn from an industry spon-
sored study between the American Petroleum Institute (API) and the
Deepstar Joint Industry Project VI. The paper identified at least 485 suction
caissons had been installed at more than 50 locations in water depths up to
2000 m. These were for oil and gas applications. At the time of writing this
book over 40 complete jackets with suction buckets, and more than 500
suction foundation piles, for both oil and gas and renewable platform proj-
ects, have been installed worldwide (Tjelta, 2015). In Europe, up to the end
of 2019, 4258 monopile foundations have been installed as offshore wind
turbine (OWT) substructures, amounting to 80% of the market share by
facility (Wind Europe, 2020). Table 1.1 shows that monopiles represent
over 80% of all European OWT foundation substructures.
Intermediate foundations, which may be driven, suction installed, vibro-
or dynamically installed, have become established for a variety of offshore
applications. In many cases they are an economical alternative to piling,
with savings on materials, installation equipment and installation time.
Perhaps uniquely among foundation designs for fixed structures, intermediate
suction foundations offer a realistic option for complete removal. Apart
from drilled rock sockets for wind turbines and pylons, there are no
intermediate foundation precedents in onshore practice.

Table 1.1 S ubstructures for Offshore Wind Turbines in Europe (Wind Europe,
2020).

Cumulative

Substructure Type Number Market Share [%]

Monopile 4258 81.0
Jacket 468 8.9
Gravity Base 301 5.7
Tripod 126 2.4
Tripile 80 1.5
Floating Spar Buoy 6 0.1
Floating Barge 1 0.02
Semi-Submersible 2 0.04
Others 16 0.3

Note(s): 5258 total to 31 December 2019

4 Introduction

At the time of writing no (text)book exists that deals solely with interme-
diate foundations. This book attempts to fill the gap. Aspects of intermedi-
ate foundations are covered in the international standard ISO 19901-4:2016
(ISO, 2016a), but the scope is brief and limited to the petroleum and natural
gas industries. Design guidance based on existing practice for suction
installed intermediate foundations for fixed offshore wind are addressed in
the recent Offshore Wind Accelerator guidelines (Carbon Trust, 2019),
while new design methods for monopiles for fixed wind is the subject of the
PISA project (Byrne, 2017). This book aims to comprehensively address
intermediate foundations for both hydrocarbons and renewables drawing
from the historical context to current state of the art. The target audience of
this book is mainly offshore geotechnical engineers involved in installation
and design. In addition, those from other offshore disciplines may be inter-
ested in reading selected topics.

1.2 MATCHING MODELS AND DATA QUALITY FOR

GOOD DESIGN

A theme throughout this book is that models used should be only as accu-
rate as the data used. In particular for offshore geotechnics, this generally
precludes using more advanced models since the required data are simply not
obtained during routine offshore investigations, and engineers have to resort
to empirical correlations. Examples of missing (or inaccurate) parameter val-
ues include (mass) coefficients of consolidation cv and permeability k, pre-
consolidation pressure p'c, and compression moduli Cc/(1 + e0), Cs/(1 + e0).
Geotechnical engineers should not be lulled into a false state of security when
(for example) a 3D finite element analysis (complete with imagery and ani-
mations) has been let loose on poor quality data.
This point – that accuracy of prediction is a function of the quality of the
method as well as the data used – was first made by Lambe in his 1973
Rankine lecture (Lambe, 1973). His original figures are reproduced in
Figure 1.3. Note that, to state his case, the accuracy contour lines in these
figures are either straight or elliptical. Other (intermediate) contour lines are
possible. For example, assuming parabolic-shaped accuracy contours
(assuming equal weights and consistent with the simple equation “accuracy
= model × data”), Figure 1.4 has been prepared. Again, this figure shows
that little is to be gained by using more accurate models with low accuracy
data. For example, using 50% accurate data and model, 25% (= 50% ×
50%) accuracy is obtained. This increases slightly to 40% accuracy (= 50%
× 80%) if the model accuracy is increased to 80%. Note that the same 40%
accuracy can be achieved if data and model accuracy are both modestly
increased to 64%.
A prime example of high-quality offshore predictions being made was for
the North Rankin “A” platform, North West Shelf, Australia. In the 1980s,
Introduction 5

Figure 1.3 Interdependency of accuracy of methods and data on predictions

(Lambe, 1973).

Figure 1.4 A
ccuracy of predictions – equal weights and “accuracy = model ×
data.”

concerns were raised about the belled pile tip resistance in layered weak
calcarenite. High-quality data (triple tube PQ size rock coring providing
virtually 100% recovery and detailed core logging, plus stress-path-specific,
high-pressure consolidated-drained triaxial testing and in-situ plate load
tests) were matched using a double hardening constitutive soil model to
perform finite element analyses of single bells and groups of bells (Smith
et al., 1988).
6 Introduction

1.3 STRUCTURE OF THE BOOK

Chapter 2 sets out definitions of shallow, intermediate and deep founda-

tions, compares characteristics of modes of operation and outlines geometry
and terminology for intermediate foundations; Chapter 3 goes on to address
determination of appropriate loads for foundation design; Chapter 4 intro-
duces some basics of marine geology; and Chapter 5 considers the inter-
relationship between loading conditions and soil properties in terms of
drainage response. Chapter 6 presents hazards, uncertainties and risk mini-
misation illustrated with case histories. Chapter 7 presents the process and
components of geotechnical design for intermediate foundations from desk
study to site investigation. Chapter 8 outlines the design basis for intermedi-
ate foundations in terms of installation, in-place resistance and in-place
response. Chapters 9–11 present detailed discussion about these three key
topics. At the time of writing, there is no general consensus as to which
design methods/models are appropriate for intermediate offshore founda-
tion installation (Chapter 9), in-place resistance (Chapter 10) and in-place
response (Chapter 11). Hence, there is comprehensive discussion and advice
on models to be found in these three key chapters. Finally, in Chapter 12,
some ‘miscellaneous’ design considerations are presented.
Chapter 2

Offshore foundation types and mode

of operation

2.1 DEFINITIONS – SHALLOW, INTERMEDIATE AND

DEEP FOUNDATIONS

Table 2.1 gives definitions for “shallow”, “intermediate” and “deep”

foundations. “Shallow” foundations are mostly for lightly loaded seafloor
structures, while heavily loaded fixed platforms are founded mainly on
“deep” (pile) foundations – with the exception of platform gravity-based
structures (GBS). A third foundation type is “intermediate”, situated between
shallow and deep. Examples include small stubby caissons to support
seafloor structures, monopiles for offshore wind turbines, and caisson
anchors for semi-submersible or other types of floating structures.
Intermediate foundation installation and in-place response (resistance
and stiffness) both need to be considered, usually concurrently. Intermediate
anchor resistance in normally consolidated clay is a function of time after
installation, while an appropriate Factor of Safety (FOS) on in-place resis-
tance has to be assigned a priori in order to determine the foundation geom-
etry. Design requires simultaneous input from (and interaction between)
geotechnical and structural engineers. These aspects of intermediate founda-
tion design differ from those normally used to design shallow or deep pile
offshore foundations.
Intermediate foundations must be installed to their design penetration
below seafloor. For intermediate support foundations, this ensures sufficient
in-place resistance to withstand combined vertical, horizontal, moment and
twist (VHMT) loads. Unlike piles, proof of sufficient V resistance during
installation is not necessarily proof of sufficient resistance under VHMT
loads. This is due to interaction effects: the available axial V resistance
decreases as H and/or M and/or T load increase. Similarly, anchors need lug
levels at the correct depth below seafloor to optimise pullout resistance and
ensure a near horizontal translation mode.

7
8 Offshore foundation types and mode of operation

Table 2.1 Offshore Foundation Definitions.

Foundation Type Approximate Resistance a /Load Transfer/ Foundation

Embedment Response b Characteristics Examples
Length: due to Seafloor VHMT
Diameter Loads c
Ratio, L/D [–]
“shallow” L/D < 0.5 Resistance is a function Pipeline and cable
of VHMT loads, i.e. infrastructure,
coupled. gravity base
Majority of VHMT load is structures,
transferred to jacket mudmats,
embedment depth. both skirted and
Similar responses at unskirted.
seafloor and
embedment depth.
“intermediate” 0.5 ≤ L/D Resistance is a function OWT monopiles,
≤ 10 of VHMT loads, i.e. caissons and
coupled. anchor piles
Part of VHMT load is (installed by
transferred to deadweight,
embedment depth. followed by
Lateral response is either suction
essentially rigid under assistance,
HM load, with ≈ impact driving
constant rotation θ xz or vibratory
with depth. Responses driving).
differ at seafloor and
embedment depth.
“deep” L/D > 10 VT resistance is a Pipe piles for
function of VT loads, fixed platform
HM resistance d is a or anchor piles
function of H load only, for floating
i.e. uncoupled axial/twist platforms.
and lateral resistance.
Some of V load e,f , but
little or no HMT load,
is transferred to
embedment depth.
Some V response, but
negligible HMT
response, at
embedment depth.

a Resistance: ultimate under combined axial, lateral, moment and twist (VHMT) loads at seafloor
b Response: forces (including moment M and torsion T) and displacements (including rotation θr and
twist θt)
c Seafloor VHMT loads: see Figure 2.2a
d Second order effects ignored (e.g. p-delta for laterally loaded piles)
e Ratio of seafloor V load: tip V resistance can be almost zero for a pile in clay, but up to 50% for a pile

end-bearing in sand
f Flexible foundation: soil system assumed (Murff, 1975)
Offshore foundation types and mode of operation 9

Table 2.2 Foundation In-place Resistance Assessment Models/Software.

Foundation Type Modes References/Software

Shallow Foundations Effective area ISO 19901-4 (2016a)
(embedment ratio concept API RP 2GEO (2011)
L/D < 0.5) Resistance ISO 19901-4 (2016a)
Figure 2.1 envelope API RP 2GEO (2011)
Feng et al. (2014)
Gourvenec et al. (2017)
Intermediate Foundations Rotational CANCAP2 (Fugro, 2009)
(0.5 ≤ L/D ≤ 10) failure HVMCAP (Norwegian
Figure 2.2 (clay) Geotechnical Institute, 2004)
Figure 2.3 (sand) CAISSON (Kennedy et al., 2013)
Zero rotation DNV RP-E303 (2017)
failure API RP 2SK (2005)
AGSPANC (AG, 2003)
FALL16 (OTRC, 2008)
Resistance Vulpe et al. (2014)
Envelope CAISSON_VHM (Kay and Palix,
2011; Kay, 2015)
Deep (Pile) Foundations Axial ISO 19901-4 (2016a)
(L/D > 10) Lateral API RP 2GEO (2011)
Figure 2.4 OPILE (Cathie Associates, 2014)

2.2 MODES – SHALLOW, INTERMEDIATE AND DEEP

FOUNDATIONS

Mode of failure is one method of comparing different responses of shallow,

intermediate or deep foundations. In this section, different modes of in-place
resistance are illustrated for each foundation type to indicate the variety of
response within each foundation type and key differences between founda-
tion type. Detailed considerations for in-place resistance, along with instal-
lation and in-place response (i.e. stiffness), are covered in Chapters 9–11.
Table 2.2 summarises modes and software (both proprietary and com-
mercial) for assessing in-place resistance of offshore shallow, intermediate
and deep foundations for undrained soil response.
The following paragraphs summarise and compare modes for the three
foundation types and show that, like shallow foundations, VHM load inter-
action and soil layering should be considered for intermediate foundations.

Shallow Foundations (embedment ratio L/D < 0.5), Figure 2.1:

There is a wide variety of possible failure modes and associated models,
each with its own assumptions and limitations. For example, for an
embedded shallow foundation or jack-up spudcan under pure V load,
10 Offshore foundation types and mode of operation

(a)

(b)

Figure 2.1 In-place failure models for shallow foundations (0 ≤ L/D < 0.5):
(a) general shear, squeezing and punch-through failure modes under
pure V load (ISO 19905-01:2016b), (b) punch-through sand into clay
failure mode under VH load (Ballard et al., 2010). (Permission to
reproduce extracts from 19901-4 for a is granted by BSI.)
Offshore foundation types and mode of operation 11

(c)

(d)

Figure 2.1 Continued: In-place failure models for shallow foundations (0 ≤ L/D
< 0.5): (c) VH resistance envelopes for sand and clay (ISO 19901-
4:2016a), (d) rotational failure under VHM loading. (Permission to
reproduce extracts from 19901-4 for c is granted by BSI.)
12 Offshore foundation types and mode of operation

(e)

(f)

Figure 2.1 Continued: In-place failure models for shallow foundations (0 ≤ L/D <
0.5): (e) internal scoop mechanisms (Bransby and Yun, 2009), (f) VHM
yield envelopes for surface foundations with and without tension
(Taiebat and Carter, 2002).
there are six axisymmetric failure modes (Figure 2.1a). The modes, and
models to capture them, become more complicated for combined VH(M)
loading, especially for multi-layered soils (e.g. Figure 2.1b), and are rou-
tinely simplified (e.g. Figure 2.1c). A traditional simplification for shallow
foundations is to use the “effective area” method (ISO 19901-4: 2016a).
The effective foundation width and lengths are given by B' = B − 2 ex and
L' = L = 2 ey where ex = My/V and ey= Mx/V. The point of V load applica-
tion must lie in the middle third in order to avoid tensile stresses at foun-
dation embedment level. It is cautious, especially for embedment ratios of
around 0.5. In addition, plane strain vertical slices may be considered,
usually with shear contributions on the two out-of-plane faces to account
for 3D effects (Figure 2.1d). More possibilities occur if the foundation is
not solid but has ‘skirts’ (Figure 2.1e). A variety of VHM(T) resistance
envelopes are available for rectangular and circular foundations on und-
rained soil (Figure 2.1f). Unlike traditional bearing capacity methods,
which use linear load inclination factors, these use 3D FEA to derive
VHM(T) interaction effects more rigorously. Also, the “effective area” is
no longer needed, since the M resistance is automatically obtained for a
given VH load.
Intermediate Foundations (embedment ratio 0.5 ≤ L/D ≤ 10), Figure 2.2
(clay) and Figure 2.3 (sand):
All the modes shown in Figure 2.2 are for rigid foundations in undrained
soil and under combined VHM load. The first three modes use 2D plane
Offshore foundation types and mode of operation 13

(a)

Shallow Rotational Failure

Intermediate Rotational Failure

Deep Rotational Failure

Figure 2.2 In-place failure modes for intermediate foundations (0.5 ≤ L/D ≤ 10)
in undrained soil: (a) “shallow, intermediate and deep” caisson rota-
tional failure (Kolk et al., 2001).
14 Offshore foundation types and mode of operation

(b)

Figure 2.2 Continued: In-place failure models for intermediate foundations

(0.5 ≤ L/D ≤ 10) in undrained soil: (b) rotational mechanisms (Kennedy
et al., 2013).

strain limit equilibrium and shear contributions on the out-of-plane faces,

a variety of rotational failure modes, and options for pile head fixity and
gapping behind the pile. The fourth (Murff and Hamilton, 1993) uses
upper bound plasticity theory and models the soil as a conical soil wedge
with horizontal plane strain flow around the pile at depth. Figure 2.2e
shows a “tongue”-shaped VHM envelope developed from ≈ 5500 finite
element analyses (FEA). More resistance envelope models are given in
Section 10.12.2 (Undrained Soil Response) and Table 10.6. Note that,
like shallow foundations, these resistance envelope models permit foun-
dation optimisation without resorting to 3D FEA. Finally, Figure 2.2f
shows a failure model specifically for anchor chain trenching.
Offshore foundation types and mode of operation 15

(c)

Figure 2.2 Continued: In-place failure models for intermediate foundations

(0.5 ≤ L/D ≤ 10) in undrained soil: (c) “fixed-head” anchor pile under
chain T load (after Andersen and Jostad, 1999).

There are fewer models for sand than for clay, and these are shown in
Figure 2.3. The first (Broms, 1964a, 1964b) is for laterally loaded
caissons and the second shows the Butterfield and Gottardi (1994)
“rugby-ball”-shaped VHM envelope for shallow (not intermediate)
foundations.
Deep Foundations L/D > 10), Figure 2.4:
Unlike for shallow and intermediate foundations, the axial and lateral fail-
ure models for deep pile foundations are uncoupled. This is reasonable:
uncoupling accounts for the fact that the soil fails first during axial load-
ing whereas the pile fails first (bending stress) under lateral load. There
are no equivalent VHM(T) resistance envelopes for deep piles – only
axial resistance at pile head, Vmax versus depth profiles for various pile
diameters D.
16 Offshore foundation types and mode of operation

(d)

Figure 2.2 Continued: In-place failure models for intermediate foundations

(0.5 ≤ L/D ≤ 10) in undrained soil: (d) pile under VHM load (Murff and
Hamilton, 1993). (With permission from ASCE).

Axially loaded open-ended pipe pile compressive resistance is the sum of

skin friction and end-bearing, and there are two possibilities. As
shown in Figure 2.4a, essentially failure occurs either in a “coring” or
“plugging” mode, and the minimum resistance value is governing.
The remaining three figures deal with lateral resistance modes; the first
two show a single and double plastic hinge occurring for surface HM
load and anchor pile Tlug load respectively (Figure 2.4b and 2.4c).
Offshore foundation types and mode of operation 17

(e)

(f)

Figure 2.2 Continued: In-place failure models for intermediate foundations (0.5
≤ L/D ≤ 10) in undrained soil: (e) VHM envelope for clay (Kay, 2015),
and (f) anchor pile chain trenching failure model (Alderlieste et al.,
2016).

Figures 2.4b and c are for clay profiles. In sand, earth pressures are
slightly more difficult to compute and are usually assessed using pas-
sive wedge theory (Figure 2.4d).

Note that:

• Intermediate foundations behave like piles when axially loaded – i.e.

they usually either respond “plugged” or “coring”. However, unlike
piles, there is a third mode (“leaking”) which may occur during pull-
out, more details of which are given in Section 10.5 (Maximum Axial
Resistance).
• Modes become simpler and less numerous as one progresses from
“shallow” through “intermediate” to “deep” foundations. This is
because shallow foundation resistance is extremely sensitive to small
variations in soil layering and strength. On the other hand, deep foun-
dations obtain their resistance by averaging shear strength over larger
areas, and are therefore not so sensitive to weaker layers.
18 Offshore foundation types and mode of operation

(a)

(b)

(c)

Figure 2.3 In-place failure models for intermediate foundations in drained soil: (a) HM load
on laterally loaded rigid pile (Randolph and Gourvenec, 2011) and (b) VHM
envelope for shallow foundations (Butterfield and Gottardi, 1994).
Offshore foundation types and mode of operation 19

(a) (b)

(c)

(d)

Figure 2.4 I n-place failure models for deep (pile) foundations (L/D > 10): (a) axial
(soil) failure, (b) and (c) lateral (pile steel) failure, (d) lateral (soil
wedge) failure (Reese et al., 1974).
20 Offshore foundation types and mode of operation

2.3 INTERMEDIATE FOUNDATION GEOMETRY

2.3.1 General
Intermediate suction foundations are closed at the top, open at the base, and
generally cylindrical. In clay, shapes other than cylindrical may be feasible
(for example, square or triangular).
Suction foundation top plates, which contain vent(s), are usually flat. Top
plates may be domed in order to reduce steel bending stresses should
extremely high underpressures be anticipated. Domed top plates have a sin-
gle central vent. A geotechnical side benefit is that domes can accommodate
more soil plug heave. If inverted T-beam stiffeners are placed below the flat
top plate, “rat-holes” are generally added to facilitate water egress. Apart
from depth markers on the outer surface (used for installation monitoring),
all steel cylinder surfaces are left unprotected in order to maximise available
friction resistance. If the surfaces are not rusty, or are painted or treated in
other ways that reduce skin friction, reduced interface friction factors should
be considered. Suction foundation embedment ratios are a function of foun-
dation type and soil conditions, and typically vary between 0.5 and 6.
Cylinder wall thicknesses are much less than driven pipe piles, typically
D/250 to D/100. Top plate thicknesses are higher than cylinder wall thick-
nesses, usually between D/100 and D/50.
With the increase of turbine size, monopiles for offshore wind farms are
experiencing an increase in diameter, but similar lengths, leading to a
decrease of their slenderness ratio (L/D). In the last decade or so, turbine
capacity increased from 4 to 12 MW and turbine suppliers have already
announced that 14 MW turbine will be on the market soon. Current proj-
ects consider monopile diameters in the range of 7 to 10 m. Due to the
conjunction of increasing turbine size and the growing maturity of this
industry (i.e. better optimisation to be linked with a significant R and D
effort in the past few years), the associated embedment ratios are decreasing
from a L/D of 5 to L/D of 3 (or even less). Limits on the size of monopiles
are likely to be driven mainly by the manufacturer’s limits in terms of
fabrication and transport (in Europe some manufacturers can fabricate a
monopile 12 m diameter and up to 120 m long and 2000t) or the offshore
installation vessels available on the market.

2.3.2 Internal stiffeners
Internal stiffeners may be added to intermediate foundations. Examples
include:

• Full-height cross plates for support foundations – to either correct

tilt during installation by using unequal compartment underpressures,
or to reduce soil plug primary consolidation settlement in normally
consolidated clay profiles.
Offshore foundation types and mode of operation 21

• Full- or part-height cross plates for support foundations – to prevent

against an internal scoop mechanisms in caissons with embedment
ratios < 2 in non-uniform soil.
• Ring stiffener at lug level for anchor foundations – to distribute chain
loads and reduce structural steel stresses.
• Ring stiffeners along the length of the caisson – to improve pile bend-
ing stiffness and reduce internal friction during penetration.

2.3.3 External stiffeners
External stiffeners (usually two vertical beams) may also be added to the
outside of the cylinder to reduce structural steel stresses for foundations
installed by skidding (instead of lifting by crane). External stiffeners
added above the top plate are not usually of interest to geotechnical
engineers.
The cylinder is usually of constant outside diameter and mainly of uni-
form wall thickness. However, in sand profiles, the wall (and stiffener) thick-
ness may increase at or near tip level; these friction breakers reduce internal
friction. Similarly, for anchor foundations, the wall thickness may increase
at or near lug level. In clays, the use of vertical stiffeners is preferred over
ring stiffeners to reduce the possibility of gapping (vertical water-filled
annuli or slots) and twist (misalignment in plan).

2.3.4 Protuberances
Protuberances include mooring and lifting padeyes/lugs, longitudinal or ring
stiffeners, changes in wall thickness, mooring chain, launching skids, water
injection tubing and others. Occasionally, fins may be added to increase
lateral resistance of intermediate anchor foundations. Fins are two steel
plates (with stiffeners to reduce bending stresses) sticking out from the cais-
son wall orthogonal to the H load direction.

2.3.5 Embedment ratio
Generally the objective is to get the foundation as deep as possible in order
to maximise foundation resistance. This is because soil generally becomes
stronger (and hence unit values of in-place resistance increase) with depth.
However, suction installation (not in-place resistance) considerations are
usually governing. If an infinitely large underpressure can be applied, and
the steel cannot yield, then installation stop criteria are usually soil plug
liquefaction (sand) and soil base failure (clay). However, intermediate
foundations installed by impact driving have the same stop criteria as deep
(pipe) piles – steel fatigue and maximum blow count.
22 Offshore foundation types and mode of operation

For monopile foundations for offshore wind farms, the diameter is gener-
ally governed by frequency and fatigue requirements. The embedment is
adjusted to limit the accumulated displacement at the top and to ensure that
a small variation of the length would not induce a significant variation of the
mudline displacement. There is no specific installation constraint that would
limit the embedment ratio apart from site specific stratigraphy (e.g. presence
of relatively shallow bed-rock). As discussed in Section 2.3.1, L/D ratio tends
to decrease with increase of turbine size, leading to more rigid foundations.
Figure 2.5 shows that (unlike suction support and anchor foundations),
offshore wind turbine monopile foundations are likely to increase in diam-
eter D (and decrease embedment ratio L/D) in the future.

(a)

(b)

Figure 2.5 E uropean monopile geometry trends. Geometry data from Negro
et al. (2017); wind farm inauguration dates from www.4coffshore.com.
Offshore foundation types and mode of operation 23

2.4 SUMMARY OF INTERMEDIATE FOUNDATION TERMS

Intermediate foundation: foundations with an embedment ratio between

0.5 and 10. Usually circular in plan. Foundation behaves essentially rigid
when laterally loaded. See Section 3.1.
Intermediate suction foundation: steel cylinders which are (temporarily or
permanently) closed at the top by a steel top plate and installed using
underpressure (“active suction”) and/or water pressure less than ambi-
ent pressure is generated inside the cylinder under the top plate during
tensile loading (“passive suction”) and possibly removed/extracted using
overpressure.
Underpressure: water pressure inside the cylinder under the top plate which
is less than ambient pressure.
Overpressure: water pressure under the top plate which exceeds ambient
pressure inside the cylinder.
Both pressures are relative to hydrostatic at seafloor (not under the founda-
tion top plate).
Installation: launching or lowering the foundation into place on the sea-
floor, followed by further lowering using self-weight and suction assis-
tance to its design penetration depth below seafloor.
Retrieval: foundation recovery as a contingency measure during installation,
and subsequent re-installation.
Removal: permanent recovery of foundation after completion of the opera-
tional phase. Unlike retrieval, this may include raising it back to the
installation vessel.
Base failure: reverse end-bearing failure at foundation tip level (when using
underpressure, clay).
Liquefaction: zero effective stress in soil plug (underpressure, sand).
Piping: hydraulic leak between water plug under the top plate and seafloor
around the foundation tip (under pressure, sand).
Plug heave: soil plug surface heave (both clay and sand).
Set-up: inner and outer friction increase with time after installation (nor-
mally consolidated and lightly over consolidated clay, and loose to very
dense sands).
Tilt: foundation non-verticality or out of plumb.
Twist: foundation misalignment or misorientation in plan.
Stick-up: foundation height above seafloor to allow for general seafloor
slope, local seabed variations and steel displaced in soil, and other effects.
Embedment ratio: foundation embedment depth length divided by the
shortest plan dimension, equal to L/D for circular foundations.
Chapter 3

Loads

3.1 INTRODUCTION

Loads are the starting point for any foundation design and installation
assessment. This is because a lightly loaded structure (for example, a 100 kN
compressor) can be designed with a different foundation type than a heavily
loaded structure (for example, an offshore wind turbine monopile subjected
to significant seafloor lateral loads, H = 2 MN, M = 150 MNm).
In addition, offshore structural and geotechnical engineers have different
perspectives about loads. Table 3.1 shows that for support foundations: (a)
only the most critical (lowest factor of safety, FOS) load cases need
geotechnical analysis; (b) structural and geotechnical load formats are
different; (c) geotechnical engineers often assume MH loads are generally
co-planar, although less so for offshore wind, and (d) structural engineers
usually consider only in-place conditions. For anchor foundations, the
interface is simpler and perspectives are broadly the same – there are fewer
load cases, and structural loads have now only 2 degrees of freedom (DOF)
instead of 6.

3.2 UNITS, SIGN CONVENTIONS AND REFERENCE POINT

Offshore engineering practice is to use the following SI system:

  kN, m, tonne mass, seconds and degrees

3.2.1 Example – units
Note that using tonne mass eliminates the need for including gravity
(g = 9.81 m/s2) in dynamics problems. Three examples neatly illustrate this.

25
26 Loads

Table 3.1 Structural and Geotechnical Perspectives of Loads.

Perspective

Consideration Structural Engineer Geotechnical Engineer

Support Foundations (point of load transfer at seafloor)
Software SACS, Nastran, SESAM, ANSYS, Analytical, PLAXIS 3D,
etc. ABAQUS, etc.
Number of Load Many (FLS, ULS, ALS, dynamic, Fewer (ULS, SLS, ALS,
Cases seismic). seismic).
Analysis Type In place response only. Usually Routine projects – usually
linear (due to high # load ULS (installation and
cases), non-linear only for a in-place resistance) and
limited number of load linear (in-place
cases. response) and stiffness
response (both non-
linear and linear)
Foundation F x , F y , F z , M x , M y , M z (6 DOF) V, H, M, T (4 DOF) as well
Loads as
Fx, Fy, Fz, Mx, My, Mz (6
DOF)
Foundation δ x , δ y , δ z , φ x , θ y , Ψ z (6 DOF) δ z , δ x , θ xz, Ψ xy (4 DOF)) as
Displacements well as δx, δy, δz, φx, θy,
Ψz (6 DOF)
Miscellaneous Foundation Stiffness Foundation Settlements

Anchor Foundations (point of load transfer usually at depth below seafloor)

Software None for foundation steel. Analytical, PLAXIS 3D,
Mooring systems: Orcaflex, ABAQUS etc.
SESAM, etc., for chain loads
at seafloor.
Number of Load Few, usually either two (Intact Same number as
Cases condition and redundancy Structural
check – ISO 19901-7) or
three (All lines intact, One
line damaged, Transient
– O&G Organisations),
possibly assess MBL.
Analysis Type None – chain loads generally Anchor chain inverse
provided at seafloor. catenary in soil
ULS (installation and
in-place resistance)
Foundation Seafloor: T seafloor , θ seafloor V lug , H lug (= VHM at
Loads Lug level: T lug , θ lug (= V lug and seafloor)
H lug , 2 DOF)
Foundation Not considered. Not considered.
Displacements
Miscellaneous FEA lug stiffener stresses. In-place tilt and twist.

Note(s): Displacement symbols φ, θ and Ψ are consistent with those commonly used for ship motions:
x = surge, y = sway, z = heave
φ (phi) = roll, θ (theta) = pitch, Ψ (psi) = yaw
Loads 27

Example 1: Vp, the compression wave speed in an elastic solid, is given by

Vp = √(E/ρ). For steel, Young’s modulus E = 210e6 kPa and mass density
ρ = 7.7 tonnes/m3, giving Vp = 5200 m/s, which is the correct speed. No g
value has been used.
Example 2: The first natural frequency of a lumped spring-mass SDOF
(single degree of freedom) system is given by the equation ωn1 = √(k/m),
where ωn1 is frequency (radians/second), k is spring stiffness (kN/m) and
m is mass (tonnes). Again g = 9.81 m/s2 is absent in the equation.
Example 3: One newton (1 N) is defined as the force required to accelerate
an object with a mass of 1 kilogram (1 kg) by 1 m/s2. Using Newton’s
Second Law (F = ma), we have acceleration a = F/m, Hence unit accelera-
tion a1 = 1 N/1 kg. Multiplying the top and bottom by 1000, we now
have the same a1 value, i.e. a1 = 1000 N/1000 kg. Hence, using kN force
and tonne mass automatically gives correct acceleration units (m/s2), mak-
ing conversion factors unnecessary.

Geotechnical engineers should request/obtain/report centric loads at sea-

floor. This is the point of load transfer; see Figure 3.1a. This is important for
both shallow and intermediate support foundations under combined (VHMT)
loading. Centric means at the foundation geometric centre line. This is not
necessarily the point where the platform leg meets the foundation – they
could be offset. A typical reason for the offset is that platform installation
barges have a fixed width. Offsetting platform legs gives a few more metres
of clear spacing between the caissons. Another reason for offsetting legs
could be for optimisation of load transfer to the soil. For anchor founda-
tions, chain loads (and inclination angles) are usually given at seafloor, not
at anchor lug (padeye) level; see Figure 3.1b.

(a) (b)

Figure 3.1 L oad and displacement sign conventions – positive shown: (a) shallow
foundations and (b) anchor mooring.
28 Loads

Figure 3.2 C
omparison of sign conventions: (a) structural analysis (b) geotechni-
cal analysis – laterally loaded foundations – HM sign convention.

Geotechnical engineers almost universally adopt compression positive

with a right-handed axis and clockwise positive sign convention for loads
and moments, as shown in Figure 3.1a.
Figure 3.1a shows that an in-plane horizontal load acting left to right and
clockwise moment are taken as positive, and this is logical since the over-
turning moment generally results from an overturning horizontal load act-
ing above foundation level – i.e. both act in the same direction. Some
softwares take overturning H and restoring M as positive, so care is needed
to change sign convention if this is the case.
Structural and geotechnical engineers can adopt different conventions,
and it is important to transfer/convert loads properly. In particular the geo-
technical compression positive convention may differ from the structural
engineering interpretation that often considers tension positive. This is
straightforward in itself, but it also affects the direction of the y axis, which
needs to be accommodated in conversion (Figure 3.2). Multiple variations
of sign conventions can be adopted by different engineers across the design
process, and it is essential that the adopted convention is clearly communi-
cated between groups.

3.3 STRUCTURAL TO GEOTECHNICAL LOAD CONVERSION

Structural engineers generally supply all six structural load components (i.e.
Fx, Fy, Fz, Mx, My, Mz). Equivalent VHMT geotechnical loads need to be
derived from these. As expected, V and T loads are straightforward. However,
resultant H and M are slightly more complicated. Using Excel/MathCAD
function atan2(x,y), the equations to derive VHMT loads are given by:

H atan2 Fx ,Fy (3.1)

M atan2 M y , M x note the minus sign (3.2)

Loads 29

V Fz (3.3)

H sqrt Fx 2 Fy 2 (3.4)

M sqrt M x 2 M y 2 ; M if H M 180 (3.5)

T = Mz (3.6)

where
αH = resultant H angle x-y plane (radians)
αM = resultant M angle x-y plane (radians)

Note the use of function atan2 and its arguments. Excel/MathCAD

functions atan2(x,y) return the angle (in radians) from the x-axis to a
line between the origin (0,0) and the point (x,y). The corresponding
Matlab/FORTRAN functions have transposed arguments, namely
atan2(y,x). Function atan2(x,y) is superior to function atan(number)
which returns angles (in radians) only between −90° and +90° and
becomes infinite at ±90°.

3.3.1 Example – load conversion

Convert structural loads (Fx, Fy, Fz, Mx, My, Mz) = (828.79, 128.94, −1012.78,
−536.40, 3297.37, −135.07) to geotechnical VHMT loads and compute
resultant angles αH, αM and | αH − αM |. Units are in [kN, m].

αH = 8.8 [deg]
αM = 9.2 [deg]
|αH − αM| = 0.4 [deg]
V = −1012.78 [kN]
H = 838.76 [kN]
M = 3340.71 [kNm]
T = −135.07 [kNm]

The next paragraph applies to single isolated foundations only: H and M

may be non-co-planar for unsymmetrical co-joined foundations subjected to
asymmetric VHM(T) loads.
Good practice (internal, sense check) should compare αH and αM values.
Because H and M are co-planar, αH and αM should be (almost) the same (|αH
− αM| ≈ 0°, M and H both overturning or both restoring) or differ by 180°
30 Loads

(M restoring and H overturning or vice versa). In practice, small differences

are found (a few degrees). These differences are usually due to either (i)
numerical “noise” or (ii) small denominator values in arctangent associated
with α values close to either 90° or 270°.

3.4 GEOTECHNICAL STRESSES AND STRAINS

Compressive stresses and strains are generally taken as positive in geotech-

nical design. This is the other way around from structural analysis, where
typically tension is positive.

3.5 COMMENTARY

Always request clarification if unsure about units, especially ambiguous

ones like MT (or mT), which may mean metric tonnes and not mega tonnes.
Always state units and sign convention in calculations and reports. Other
readers may have different systems.
Chapter 4

Marine geology

4.1 GEOLOGY, SEDIMENT TYPES AND DEPOSITIONAL

ENVIRONMENTS

Knowledge of marine geology is important for offshore engineering. Soil

type and soil properties are the main parameters for geotechnical engineer-
ing, and these are determined by geological processes. Understanding these
processes, and the variations of the processes, aids prediction of soil type
and behaviour, or what to expect.
Facies are distinctive soil/rock units that form under certain conditions of
sedimentation, reflecting a particular process or environment. A 30 m bore-
hole can contain a number of different facies, and their thicknesses can also
vary laterally.
Table 4.1 lists the three different marine sediment types. These types are
not particularly related to marine environments. Most sedimentation takes
place in low-lying areas – usually seas and oceans, but also lakes, rivers,
marshes and plains. Marine and terrestrial depositional environments, with
typical soils, are compared in Table 4.2. Some soils are more hazardous to
installation than others. More marine geology information is given in the
textbooks by Seibold and Berger (2010), by Nichols (2009) and in Section
2.3 of Randolph and Gourvenec (2011).

4.2 LATERAL VARIABILITY TOP LAYERS

Both shallow and intermediate foundations are extremely sensitive to the

soil conditions found in the top few metres below seafloor, and soil condi-
tions can vary laterally quite rapidly – depending upon the depositional
environment. If such conditions are anticipated, then shallow seismic sur-
veys can provide lots of additional information at reasonable cost. Needless
to say, at least one seismic line has to intersect a ground truth borehole in
order to facilitate correlation/extrapolation.

31
32 Marine geology

Table 4.1 Marine Sediment Types (Seibold and Berger, 2010).

Type Examples Occurrence

Clastic/ Products of erosion and Fine-grained
lithogenous weathering of ANY kind of lithogenous
rock – sandstone, sediments (i.e. clays)
mudstone, carbonate rock, are the most
etc. abundant of all
Silica sand, silts and clays marine sediments
(about 70% by
volume)
Biogenic/ Products of organisms and Biogenic sediments are
biogenous their activity widespread and
Carbonate sand, coral reef arrive at seafloor,
breccia, oolites and oozes about 30% by volume
Chemical/ Sediments produced by Not significant by
hydrogenous chemical reactions volume, but
Halite, anhydrite widespread

Table 4.2 Depositional Environments.

Environment Feature Typical Soils

Marine River/delta Fine to medium sand and clay;
some organic material
Tidal flat/lagoon/ Organic clay; sand lenses
barrier system
Beach/continental Silica sands
shelf
Deep sea Normally consolidated clay
Terrestrial Lakes/swamps Organic material/peat finely
interbedded clays/sands
Alluvial fans/flood Fine to coarse angular sand and
plains gravel possibly cobbles and
boulders some organic
material inclined layers
Dunes (aeolian Fine well-sorted sand
sedimentation)
Glacial deposits Overconsolidated (medium
dense to very dense) sands
and gravels till: directly from
ice (firm to hard clay) drop
stones
Outwash plain Sands and gravels normally
consolidated (loose to
medium dense)
Marine geology 33

Because sands are the most challenging to penetrate, and gravels/cobbles/

boulders may cause premature refusal, overconsolidated glacial sand sites
need thorough investigation in order to adequately define the ground condi-
tions. Prime examples include the North Sea, the Baltic and the US East
Coast. Similarly, penetration resistances can exceed the highest expected in
locations where the ground, already variable, has been re-worked by ice.
Such locations include the Volga outwash deposits in the Northern Caspian
Sea, offshore Alaska and the Barents Sea, located off the Norwegian and
Russian northern coasts.

4.3 SEAFLOOR CONDITIONS

The following bathymetric information should be acquired:

(i) water depth

(ii) seafloor slopes
(iii) sand waves, dunes or ripples
(iv) scars (icebergs, trawl boards, jack-up footprints, etc.)
(v) current velocity at or near seafloor
(vi) obstructions
(vii) infrastructure (pipelines, well heads, etc.).

Ad (i) Water depth measurements obtained during a normal geotechnical

investigation using a combination of drill string, echo-sounder and water
pressure sensors are normally for reference purposes only. Hence the water
depth should be accurately related to datum for support foundations. For
deepwater applications, accuracy considerations are not so critical.
Theoretically, the maximum underpressure must not exceed the water
depth +1 atmosphere (100 kPa) in order to prevent cavitation. Hence, in
very shallow water (say, less than 10 m water depth), special installation
techniques must be considered. Besides limited underpressure, sufficient self-
weight penetration (SWP) is needed for a hydraulic seal in order to start the
suction assisted stage. Suction foundations have been installed in very lim-
ited water depth (0.15 m) in Luce Bay (Houlsby et al., 2006; Villalobos,
2007), in the IJsselmeer, Holland, and in Bremerhaven, Germany. For the
Luce Bay project, the foundations (skirt lengths either 1 m or 1.5 m, SWP ≈
0.1 m) were water-filled, probably causing hydraulic leaks from the interior
around the tip to the seafloor outside, and were mitigated by applying
underpressure (Villalobos, 2007). Neither water depths nor installation
techniques for the last two projects are known. A reasonable minimum SWP
value is 0.5 m.
34 Marine geology

(a)

(b)

Figure 4.1 Schematic of (a) sloping seafloor and (b) seafloor ripples.

Ad (ii) excessive seafloor slopes (see sketch in Figure 4.1a), may need
levelling.
Ad (iii) for a “sandy” seafloor, seabed ripple heights and lengths should be
determined. This is because an essentially level seafloor is required for instal-
lation in dense sands – otherwise, local piping may occur if there is limited
self-weight penetration, as sketched in Figure 4.1b.
Ad (v) current velocity is required for scour potential in sands.
Ad (vi) the survey should check for the absence of potential obstructions.
Examples and more details are given in Section 6.2 (Hazards for Intermediate
Foundations).
Chapter 5

Loading conditions and soil drainage

5.1 INTRODUCTION

Effective (not total) stresses affect soil shear strength and volume change
characteristics of geomaterials.
It is misleading to think that all “sands” behave “drained” and all “clays”
behave “undrained” when loaded. These effects are important. Examples
include:

• dilative sands: “undrained” strength can be a factor 3 higher than the

“drained” strength
• contractant sand can have little or no strength if rapidly loaded
“undrained”
• long-term foundation settlement on clay is a “drained” process.

Soil response depends on:

• soil permeability (coefficient of consolidation)

• soil drainage path length (foundation geometry)
• rate of load application (rapid, slow).

Table 5.1 lists typical soil responses for intermediate suction foundations. It
is seen that, for foundation sizing to withstand applied loads, installation
and long-term loading conditions generally do not govern. In addition, the
soil response is often “undrained” for a range of soil types and load condi-
tions. This is for both “sand” and “clay”, and also despite the fact that wave
loading has a longer time period (≈ 10 s) than boat impact or seismic load-
ing (< 1 s). Possible exceptions to the “undrained” rule are italicised in Table
5.1, and analysis methods for these special cases are given in Section 5.2.
The remarks here are solely for “intermediate” caisson foundations: “shal-
low” foundations, due to their shorter drainage path lengths, and the fact
that drainage times are generally proportional to (drainage length)2 tend to
be more “drained” than “undrained”.

35
36 Loading conditions and soil drainage

Table 5.1 Typical Soil Responses for Intermediate Foundations.

Load Type “Sand” ”Clay” Remarks

Installation “drained” “undrained” 0.001 m/s
penetration
speed
Wave/anchor cable “undrained” “undrained” e.g. 100 y extreme
storm or 10 y
summer storm
Boat impact “undrained” “undrained” –
Seismic “undrained” “undrained” –
Long-term “drained” “drained” consolidation
settlement

Note(s): These listings are in order of occurrence, earliest first.

“Sand” could be “partially drained” or “drained” if coarse/gravelly.
“Clay” could be “drained” followed by “undrained” if dead (static) load is significant (e.g. permanent
tension).
Section 9.8.4 (CPT Method Coefficient αu) checks that installation in sand is essentially “drained”.

5.2 DRAINED-UNDRAINED

In order to assess whether or not (shallow or intermediate) foundation

response should be regarded as “drained” or “undrained”, one or more of
the following approaches may be considered.

5.2.1 Non-dimensional velocity – penetration rate

A non-dimensional velocity parameter β based on spudcan and CPT tool
penetration was proposed by Finnie and Randolph (1994):

v D / cv (5.1)
where
v = penetration velocity
D = diameter
cv = soil mass coefficient of consolidation.

Figure 5.1 shows the variation of the non-dimensional bearing modulus

with the velocity of penetration, normalised as vD/cv, where D is the diam-
eter of the foundation and cv is the coefficient of consolidation, taken as 1 ×
10−3 m/s and 5 × 10−5 m/s for the sand and silt respectively. The transition
from drained to partially drained occurs at a non-dimensional velocity of
about 0.01, while the undrained limit is reached at a non-dimensional veloc-
ity of about 30, i.e. if β > 30, then the soil response is “undrained”, and if
β < 0.01, then it is “undrained”. The reduction in bearing modulus of the silt
Loading conditions and soil drainage 37

Figure 5.1 U
ndrained to drained soil response – effect of loading rate on bearing
response during spudcan foundation penetration (Finnie and Randolph, 1994).

under undrained conditions is much more dramatic than for the sand, owing
to a greater tendency for the latter to dilate at low effective stress levels.

5.2.2 Dynamic drainage factor – dynamic loading of

solid piles
Hölscher et al. (2009) and Huy et al. (2009) performed numerical (Biot)
analyses and centrifuge model tests of Statnamic Rapid Pile Load Testing in
sand and derived a dimensionless dynamic drainage factor η, defined as

G T k / R2 (5.2)

where
G = soil shear modulus
T = load duration
R = pile radius
γ = soil unit weight
k = soil permeability.

Figure 5.2 presents their results for closed-ended piles. Typical analyses
for open-ended pipe piles give η ≈ 1, i.e. a fully drained response (Nguyen
et al., 2012).
38 Loading conditions and soil drainage

(a)

(b)

Figure 5.2 U
ndrained to drained soil response – dynamic loading of closed-
ended piles. Centrifuge data (from Hölscher et al., 2009) effect of
loading rate on pile tip capacity (their Figure 10) and pile tip stiffness
(their Figure 8).

5.2.3 Laterally loaded pile

Osman and Randolph (2012) derived an analytical solution for soil consoli-
dation around a laterally loaded pile, idealised as a circular rigid disk sur-
rounded by elastic soil. Drainage occurs at a finite distance away from the
disk. Their equations (verified by 2D plane strain FEA) can be programmed,
but are computationally time-consuming. Figure 5.3 summarises their
Loading conditions and soil drainage 39

(a) (b)

Figure 5.3 U
ndrained to drained soil response – laterally loaded disk (from
Osman and Randolph, 2012). (With permission from ASCE.) Analytical
solutions for (a) excess pore pressure dissipation at disk/soil inter-
face T50 ≈ 0.5; T90 ≈ 5 (their Figure 7) and (b) pile consolidation
displacement (their Figure 9) T50 ≈ 10; T90 ≈ 75 versus dimension-
less time factor T = c h t/r o 2 .

results. It is noted that time factors for pore pressure dissipation are a factor
20 faster than for pile consolidation displacement. Typical analyses for large
diameter caissons (D = 10 m) in dense fine sand (“Best Estimate” parameters
k = 10−4 m/s, E' = 100 MPa, v' = 0.3, ch ≈ 1 m2/s) give dissipation times
t50 ≈ 10 s, t90 ≈ 100 s, i.e. likely to be “partially drained”–“undrained” for
wind or wave loading. For a large diameter monopile foundation (D = 5 m),
because dissipation/consolidation times are proportional to D2, dissipation
times reduce to t50 ≈ 2.5 s, t90 ≈ 25 s, i.e. “drained”–“partially drained”.
Even though ch can vary by one order of magnitude on either side of the
“Best Estimate”, these results suggest that laterally loaded intermediate
foundations may have to be considered as partially drained when either
wind or wave loaded.

5.2.4 Generic
Unlike the Finnie/Randolph and Hölscher/Huy approaches, the Zienkiewicz
et al. (1980) method can be used for both static and dynamic problems,
and it is not limited to one particular foundation type. One dimensional
40 Loading conditions and soil drainage

conditions are assumed – and this may be optimistic. The key diagram is
Zienkiewicz et al.’s Figure 3 (not reproduced here). The method first assesses
dimensionless parameters π1 and π2, using soil permeability k, drainage path
length L and loading frequency ω, and then find which Zone (I, II or III) is
associated with data point (π1, π2). Worked examples include an earth dam
response during an earthquake (“undrained” response valid only for k <
10−4 m/s at depths well below surface) and pore pressure distribution in
seafloor due to wave loading. The Zienkiewicz et al. (1980) method equa-
tions can be programmed in MathCAD or similar.

5.3 CLOSURE

No soil in the ground has a membrane around it. Hence partial drainage is
always a possibility in all soil types. Even if in-situ tests show drained condi-
tions, due to size effects, the soil may respond undrained or partially drained
in the case of large(r) foundations.
Most uncertainty is usually associated with mass cv (coefficient of consoli-
dation) values to be inserted into analytical equations. In addition, Finnie
and Randolph (1994) assume that bearing capacity modulus is proportional
to foundation width B. This may not necessarily be the case: other proposals
include B2, e.g. Hölscher et al. (2009), Huy et al. (2009) and Osman and
Randolph (2012).
Undrained shear strength may be overestimated when derived from in-
situ testing if the soil during testing behaved drained or partially drained.
This is particularly the case for CPT qc using a Nk factor.
If the foundation design is likely to include unconventional soil response,
then the geotechnical investigation/laboratory test program should make
provision for obtaining reasonably reliable parameter values. Some exam-
ples include:

• undrained “sand”: permeability k and undrained shear strength su

(triaxial CU)
• drained “clay”: coefficient of consolidation cv and compression coef-
ficients Cc, Cs and p'c (oedometer).

If reliable triaxial CU data are not available, then possibilities include: (i)
consider a (cautious) drained analysis, (ii) use the DNV (1992) approach
using no or limited dilatancy (dilatancy parameter D = 0 corresponds to
pure shear in p − q stress space, and D = 1/3 for limited dilatancy) and (iii)
use an equivalent undrained shear strength defined by su/σ'vo = 0.25−0.35
for a NC clay.
Chapter 6

Hazards, uncertainties and risk

minimisation

6.1 INTRODUCTION AND CASE HISTORIES

Intermediate foundations may encounter a range of hazards and are arguably

more hazard prone than either shallow or deep (pile) foundations. Examples
include:

1) lower-than-expected penetration resistance during installation in NC clay

2) excessive misalignment during installation in NC clay
3) cylinder buckling during installation
4) sand plug liquefaction during installation
5) underpressures close to/above critical during installation in competent
sands
6) excessive local scour during operation
7) anchor chain trenching during operation.

Notable case histories illustrating each of these seven hazards are given in
the following sections.

6.1.1 Low penetration resistance during installation

in NC clay
Two notable cases were first published in 2002 – Girassol (approximately
1400 m water depth, offshore Angola, Gulf of Guinea, West Africa) and
Laminaria (water depth ≈ 400 m, Timor Sea).
The Girassol soil conditions are very high plasticity clays (Ip ≈ 110%)
with organic and carbonate contents around 10%–15% (Colliat and
Dendani, 2002). The upper 3 m “crust” is underlain by lightly over-
consolidated (LOC) material. The original “Best Estimate” shear strength
profile (kPa, m) was su,DSS = max([6 + 1.26 (z − 3)], 6) and used for the 3
Riser Tower Anchors (RTA) vertically loaded, and for holding capacity
analyses of manifold caissons, FPSO anchors and buoy anchors of various
sizes, all VHM loaded.

41
42 Hazards, uncertainties and risk minimisation

The 3 Girassol RTAs had been painted externally for protection against
corrosion, and their key dimensions were D = 8 m, L = 19.5 m. During the
self-weight penetration stage, actual penetrations were higher than pre-
dicted and ranged from 15 m to 16 m, compared to 12 m predicted, i.e.
only 3.5 m–4.5 m less than the full length (Dendani, 2003). Using over-
pressure, two push-out tests performed on one RTA one week and one
month post installation gave ≈ 40% and 100% friction increases. These
were in relatively good agreement with laboratory thixotropy test data. A
revised long-term axial capacity model was developed, assuming “coring”,
reduced internal friction resistance above internal ring stiffeners, external
friction resistance reductions (accounting for the painted steel surface and
an upper bound St value) and increased thixotropic factor. An additional
175 tonnes of ballast were added to each RTA to obtain the required pull-
out capacity.
All Girassol FPSO and buoy anchors had also been painted and moreover
had relatively large internal ring stiffeners (0.4 m and 0.2 m outstand respec-
tively). Higher-than-anticipated self-weight penetrations were again mea-
sured (Dendani, 2003), and one FPSO anchor was retrieved (i.e. brought
back to deck) to assist in contingency decision making. No clay was seen on
the painted steel, whereas 20 mm to 30 mm clay thickness covered the non-
painted area, confirming that unit friction of painted steel areas should be
based on the lower soil-paint (not soil-steel) adhesion. In addition to this,
another factor contributing to reduced penetration resistance was the ring
stiffener: there was probably a trapped soil wedge below, and a trapped
wedge of water above, the stiffener. As with the RTAs, satisfying in-place
resistance is more critical than understanding installation. Originally, to
meet certifying agency requirements, the FPSO and buoy anchor embedded
lengths were designed taking into account tilt and twist (both ±10o), but
ignoring reverse end-bearing. Re-analysis, incorporating various
combinations of reverse end-bearing (short term), friction “set-up” (longer
term) and without tilt (as installed) gave satisfactory anchor holding capacity
(Colliat and Dendani, 2002).
The Laminaria soil conditions comprised carbonate silty clay/clayey silt
of high plasticity (Ip ≈ 70%) with a carbonate content around 70%. Based
on in-situ T-bar and laboratory CAU triaxial compression data, a simpli-
fied “Best Estimate” undrained shear strength profile was assessed to be
su = 10 + 1.8z (Erbrich and Hefer, 2002). To moor the Laminaria FPSO, 3
clusters, each with 3 anchor piles, were used. Pile diameters D were 5.5 m,
with embedded lengths L up to 12.7 m, and 20 mm wall thickness (WT).
In addition to lug stiffeners, due to the large D/t ratio (275), 8 internal ring
stiffeners, each with 165 mm outstand, were used. Their axial spacings
varied between 2.25 m and 0.95 m (Figure 6.1). During installation, self-
weight penetrations (around 2.5 m) were as predicted. Thereafter, during
suction installation, measured underpressures were approximately 50% of
Hazards, uncertainties and risk minimisation 43

Figure 6.1 L aminara suction pile geometry, D = 5.5 m, L up to 12.7 m, WT = 20

mm, with 8 internal ring stiffeners, outstand = 0.165 m (Erbrich and
Hefer, 2002).

the original Lower Bound estimate. Back-analyses suggested that, like

Girassol, decreased internal penetration resistance was probably due to a
trapped soil wedge below, and a trapped water/soil wedge above each stiff-
ener. Another possibility is a “toothpaste” effect, whereby the inner soil
plug is extruded until the height cannot stand unsupported, when it col-
lapses (McNamara, 2000). More recently, Hossain et al. (2012) performed
centrifuge testing in OC clay to explore three potential failure mechanisms
(Figure 6.2) and provided both a design chart and an equation for back-
flow initiation. Anchor in-place holding capacity was unaffected. This was
because it was a catenary mooring, θlug was low, between 15° and 20°, and
VH capacity was dominated by lateral resistance.
To avoid on-site embarrassment during installation in weak clays, current
design practice is to use reasonable wall thickness to diameter, WT/D, ratios
(at least 1/200), and to vary thickness with depth (i.e. highest WT near the
top, to accommodate highest underpressures and, if possible, avoid using
internal ring stiffeners). In addition, if they are present, internal stiffener
outstand widths are kept as small as possible and are also chamfered. Finally,
a reduced internal soil-wall adhesion or friction factor, αi value, is consid-
ered above the stiffener to assess (Low, Best and High Estimate) penetration
resistances and corresponding underpressures.
Finally, note that both the Girassol and Laminaria projects demonstrate
that (in)sufficient V capacity during installation is not necessarily proof of
44 Hazards, uncertainties and risk minimisation

(a)

(b)

Figure 6.2 S oil flow mechanisms around internal stiffeners during installation of
suction caissons (Erbrich and Hefer, 2002).

reduced capacity under VHM loads. This is due to an unaltered lateral resis-
tance H, and load interaction effects; see Chapter 8, Design basis, Section
8.1 (General principles).

6.1.2 Excessive misalignment during installation

in NC clay
The Na Kika soil conditions comprise normally consolidated Gulf of
Mexico medium plasticity clays. Different undrained shear strength
Hazards, uncertainties and risk minimisation 45

profiles were derived for various design aspects, typically su,DSS ≈ 1.25z. To
moor the FDS (Floating Distribution) semi-submersible in around 1900 m
water depth, a semi-taut mooring system with 4 clusters, each with 4 suc-
tion anchor piles, were used. To withstand factored Tlug = 10,410 kN at
θlug = 25°, final optimised anchor pile diameters D were 4.3 m, with embed-
ded lengths L = 23.8 m, i.e. a penetration ratio L/D = 5.6, as shown in
Figure 6.3a (Newlin, 2003a).
Maximum anchor pile tilt and misalignment values were both ±5°. During
installation, self-weight penetrations averaged 13 m, only slightly above the
12.2 m expected. During the subsequent suction assistance stage, all piles
remained essentially vertical (tilt < 2°), and 14 of the 16 piles had small
misalignments (< 3°). The remaining two piles (L15-P1 and L16-P3) rotated
up to 13°, requiring multiple retrieval and re-installation, sometimes with an
intentional bias (see Figure 6.3b), before they could be accepted. These two

(a) (b)

Figure 6.3 N
a Kika (a) suction pile anchor sketch (Newlin 2003a), and (b) pile twist
versus penetration (from Newlin 2003b). (With permission from ASCE.)
46 Hazards, uncertainties and risk minimisation

anchor piles were re-positioned at least 26 m away from their initial posi-
tion, along the anchor radius circumference (Newlin, 2003b).
The large rotations are unexplained and opinions differ: possible reasons
include non-vertical lug and internal stiffeners, and currents applying lateral
drag to the anchor chain (Lee et al., 2005). Another possibility is unsym-
metrical lifting points used for anchor handling. In all cases, at a given pen-
etration, all pile-soil shear has been utilised axially, so a reasonably small
torque is needed to cause rotation.

6.1.3 Cylinder buckling during installation

Madsen et al. (2012) describe the 2005 buckling incident in Wihelmshaven,
Germany, of an offshore wind turbine (OWT) support foundation in sand,
D = 16 m, L = 15 m and WT = 25 mm, i.e. an unusually large D/WT = 640.
The installation vessel collided with the foundation at 6.8 m penetration,
causing buckling failure. Subsequent installation was not possible, presum-
ably due to a hydraulic leak around the cylinder perimeter between seafloor
and the top of the internal soil plug.
Anecdotally, it is understood that at least two anchor pile projects
buckled during installation – one in normally consolidated Gulf of Mexico
clay (probably due to low D/WT) and the other in the northern North Sea
in competent glacial till (probably caused by high underpressures).
However, due to legal considerations, nothing has ever been published
about their causes, discovery and remedial measures, such that lessons
learnt are limited.

6.1.4 Sand plug liquefaction during installation

Senpere and Auvergne (1982) give details of the first ever commercial anchor
pile project in the Danish sector of the North Sea in around 40 m water
depth. Soil conditions at the 12 locations are sketched in Figure 6.4, and
consisted of around 5–7 m of loose to dense sand (φ = 35–40°) underlain by
1–2 m of soft clay (su = 20 kPa), and stiff clay (su = 70 kPa), from about
6–8 m below seafloor. Suction pile embedded lengths L varied between 8.5 m
and 9.0 m, all with D = 3.5 m and WT = 25 mm. The 200-tonne chain design
load was applied horizontally at seafloor. CPT qc values up to 24 MPa in the
sand were used to assess penetration resistance. The latter included around
50% reduction due to seepage effects. Because of higher qc values (up to
30 MPa with 36 MPa peaks), jetting devices were added at tip level just
before installation.
During suction assistance, refusal of the first anchor pile occurred at
6.6 m depth due to significant plug heave – the internal soil plug surface
had risen by around 3.5 m and had met the top plate. After airlifting the
upper sand plug, suction was re-applied and the required 9 m penetration
Hazards, uncertainties and risk minimisation 47

Figure 6.4 G
orm soil profiles. Anchor pile L = 8.5–9 m, D = 3.5 m (Senpere and
Auvergne, 1982).

(i.e. tipping out in the 70 kPa stiff clay) was achieved. To install the
remaining anchor piles, the top 1.5 m of the soil plug was liquefied by
changing the toe jetting system into one below the pile top. No under-
pressure data were reported.
Based on these data, it is considered that the root cause of pile refusal was
due to higher-than-expected penetration resistances, causing high suction
pressures and high upwards seepage gradients in the sand plug, leading to
48 Hazards, uncertainties and risk minimisation

liquefaction. Some 35 years later, current practice includes acquiring high

quality CPT data well beforehand, using “High Estimate” DNV type kp and
kf coefficients on the CPT qc data and to limit underpressures to those which
give a seafloor critical hydraulic gradient icrit = γsub/γwater. At the limit state, a
50% resistance reduction due to seepage is considered not unreasonable (see
Section 9.11 on installation in sand).

6.1.5 Underpressures close to/above critical during

installation in competent sands
Since Senpere and Auvergne (1982), no cases of liquefying sand plugs have
been reported. However, there is a growing body of documentation, both
from laboratory permeability tests and from foundation installation, both in
the field and centrifuge, of hydraulic gradients i > icrit in medium-dense,
dense and very dense sands. Figure 9.15 (in Chapter 9, Installation, retrieval
& removal) shows normalised field suction penetration data, plus a theo-
retical icrit curve. No sand plug liquefaction occurred. Minor heave occurred
due to sand plug expansion.
The risk of sand plug liquefaction has been largely mitigated by improved
design procedures, sometimes calibrated to measured field data, and
adopting due diligence procedures in the field. More details are found in
Section 9.11.

6.1.6 Excessive scour during operation

Support foundations in sand likely to be scoured, Figure 6.5 gives an exam-
ple, are generally protected, usually with some combination of mats, rock fill
and sandbags. Occasionally scour protection can fail. A gas platform in
around 40 m water depth was founded on four 7 m diameter, 9.9 m long
intermediate suction foundations and originally equipped with rock fill
placed on mats. The soil profile was sand, the upper 3–4 m of which was
loose and mobile. A routine pipeline survey some seven years after platform
installation revealed 3 m to 4 m deep scour pits around each suction founda-
tion. Analyses suggested that in-place VHM resistance was insufficient, and
the scour pits were infilled with rock fill. This was expensive, since the barge
had to be carefully positioned close to a producing platform. Scour protec-
tion is discussed further in Section 12.2.

6.1.7 Anchor chain trenching during operation

A more recent hazard is anchor chain trenching in normally consolidated
clays. Bhattacharjee et al. (2014) described issues for the Serpentina project
offshore Equatorial Guinea. Anchor chain trenches of varying dimensions
Hazards, uncertainties and risk minimisation 49

Figure 6.5 E xample of local scour resulting from environmental events (copyright
Deltares, 2016).

were found in front of all nine suction piles; see Figure 6.6. The piles were
considered unfit for service and were replaced by OMNI-Max free-fall
anchors.
Detailed 3D FEA of trenches by Alderlieste et al. (2016) have shown that,
for a particular Gulf of Guinea project, the anchor pile VH capacity reduc-
tion due to trenching was approximately 7% for the design θlug angles.
More details on this topic, including possible mitigation measures, are to
be found in Section 12.12 on anchor chain trenching.

6.2 HAZARDS

Geological hazards applicable to all foundation types have been discussed

in Sections 4.2 and 4.3. There are additional potential hazards for interme-
diate foundations. These may be conveniently subdivided into natural
(geological and geomorphological) hazards and those resulting from
human activity.

Additional natural geohazards include:

• Cemented layers/rock and coral outcrops. Cemented layers (e.g. beach

rock or caprock or just cemented sand). Unexpected cemented layers
in sediments may cause refusal (foundation installation) or differential
settlement and punch-through risk (foundation performance when
loaded).
50 Hazards, uncertainties and risk minimisation

Figure 6.6 A
nchor chain trench at Serpentina Field, Gulf of Guinea. Surveyed
dimensions (based on Bhattacharjee et al., 2014).

• Undetected boulders in apparently uniform sand (glacial erratics),

dropstones in weak clays, coral rubble in carbonate clay, or flintstones
in chalk. Besides refusal, such large-sized material may compromise
foundation tip integrity (buckling initiation). This is exacerbated
since the intermediate suction foundation WT/D (tip wall thickness/
outer diameter) ratio is usually smaller than that for deep (pipe pile)
foundations.
• Under- or overconsolidation. Soils with non-hydrostatic pore pressure
profiles may consolidate over the foundation lifetime. This is a form of
regional settlement.
• Shallow gas. Escaping gas may (a) decrease the resistance of shallow,
intermediate and deep (pile) foundations, (b) increase the pressure
inside intermediate suction foundations and (c) occasionally cause vio-
lent blowouts during drilling conductor wells.
• Sand waves, giving an additional vertical load on the foundation. This
is generally unlikely to cause stability or settlement challenges – pro-
vided scour protection has been provided.
Hazards, uncertainties and risk minimisation 51

Additional hazards resulting from human activity:

Human activity at sea has resulted in many obstacles/features on (or
under) the seafloor, such as:

• Pipelines, cables, well heads

• Wrecks, archaeological artifacts
• Munitions
• (Jack-up spudcan) footprints
• (Geotechnical) boreholes.

6.3 UNCERTAINTIES

Uncertainties (which are also potential hazards) associated with intermedi-

ate foundations can be subdivided into geotechnical data and design catego-
ries. They may include but are not limited to the following.

6.3.1 Geotechnical data
The main causes of lack of confidence are usually:

• Inadequate soil investigation. Ideally, any site investigation should be

a function of the proposed development/foundation type, and contin-
ued until the ground conditions are known and understood well
enough for the work to proceed safely (Waltham, 2009). However,
intermediate support foundations for oil and gas projects have to be
frequently designed using geotechnical data acquired for a piled plat-
form. This is due to changing scales (see Chapter 7, Investigation
programs).
• Insufficient laboratory testing, especially for sand
• Unforeseen soil. Examples include:
• undetected impermeable seams or layers, hindering formation of
seepage gradient in sand
• undetected peat, causing excessive long-term foundation settle-
ments, etc.

If geotechnical data are insufficient, then there are larger than usual uncer-
tainties about:

• soil variability (over support foundation or platform footprint, per

suction anchor foundation)
• assessment of soil shear strength
• reliability of soil deformation parameters (for support foundations).
52 Hazards, uncertainties and risk minimisation

These uncertainties are transmitted into the subsequent design phase, usu-
ally as a higher coefficient of variance (COV).

6.3.2 Geotechnical design
Selection of an inappropriate ground model and/or parameter values may
result in:

• structural integrity loss (implosion, buckling, handling, etc.)

• insufficient penetration and/or high installation pressures
• excessive self-weight penetration
• excessive tilt and twist
• high (total and differential) settlements.

All of these have been observed in case histories/field experiences, such as

illustrated in Section 6.1.

6.4 RISK MINIMISATION

The aforementioned hazards/uncertainties relate mainly to installation.

Provided the soil type is suitable for installation, insufficient penetration is
avoided by the provision of sufficient preload and/or underpressure,
combined with a high soil resistance estimate. Using additional steel
accommodates high underpressures. Generally excessive self-weight
penetration is not necessarily harmful for long-term performance. This is
because of VHM interaction effects and/or frictional set-up occurring post
installation. However, the installation contractor needs to consider possible
effects in their installation methodology. Tilt, twist and settlement are also
accommodated during the geotechnical design phase.
The main in-place hazards are possible scour and sand waves. Local and
general scour both decrease foundation resistance (and stiffness) if no scour
protection (rock dump or scour mats) are placed. Also, if the foundation is
in abnormal soils (e.g. loose sand which may liquefy during extreme
conditions), then risk minimisation options include application of a higher-
than-usual Factor of Safety or increasing tip penetration depth in order to
encounter more competent soil.

6.4.1 Ground investigation
The principle of any site investigation is that it is continued until the ground
conditions are known and understood well enough for geotechnical engi-
neering design to proceed safely (Waltham, 2009).
Hazards, uncertainties and risk minimisation 53

6.4.2 Geotechnical data
Subject to economic and time constraints, intermediate foundation sites
should be investigated as comprehensively as possible. Investigation compo-
nents include bathymetric, shallow geophysical and geotechnical data, and
should be preceded by a desk study. Insufficient attention is often given to
desk studies where valuable information can be obtained at low cost.

6.4.3 Laboratory testing
Non-routine onshore laboratory tests on samples for suction foundations
normally include:

permeability sands suction pump design, in-place

resistance
thixotropy NC clays retrieval, in-place axial
resistance
undrained triaxial/DSS sands in-place resistance
cyclic triaxial/DSS sands and clays in-place resistance
oedometer clays primary consolidation
settlement

Section 7.4 (on laboratory testing) provides more details.

6.4.4 Geotechnical design
Judicious selection of soil layering, “Best Estimate” and “High Estimate”
parameter values and associated load/material factors account for both
installation and in-place uncertainties. Similarly, platform legs may be
designed to resist possible jack-up spudcan interaction effects and differential
settlements and designing riser–flow line connections to various types of
support foundations to resist high total settlement estimates. Large
uncertainties generally result in a more robust foundation (i.e. additional
steel weight).

6.5 CLOSURE

It has been shown that intermediate foundations are arguably more hazard
prone, especially during the installation phase, than either shallow or pile
(deep) foundations.
Notable case histories of failure (and their mitigation measures) have
been given, the majority of which occurred during installation. Engineers
54 Hazards, uncertainties and risk minimisation

have learnt from their experiences and, due to sharing these, failures have
become less frequent since suction foundations were first employed in the
early 1980s. Potential hazards for intermediate foundations, both natural
and derived from human activity, have also been listed. These may serve as
useful checklists during the necessary risk management/risk minimisation
process. Appropriate ground investigations and subsequent laboratory test-
ing also assist in reducing project uncertainty.
Chapter 7

Investigation programs

7.1 INTRODUCTION

Any investigation program is a function of the proposed structure – shallow,

intermediate or deep foundations, and the anticipated ground conditions.
Programs have the following components:

• desk study
• geophysical (seafloor bathymetry, 3D, shallow seismic and side scan
sonar)
• geotechnical (and laboratory testing).

7.2 DESK STUDY

Desk studies are often the most cost-effective item in the site investigation
process. A desk study may reveal facts that cannot be discovered in any
other way. Despite these merits, the desk study phase tends to be skipped too
often due to lack of awareness. Good practice is to adjust the scope of work
to the intended development, and to bear in mind that more than one
foundation type may need consideration. A good example of this is a
mooring system, where a variety of anchor types are available (e.g. Senders
and Kay, 2002).
Key documents which may serve as useful checklists/templates for
geotechnical desk studies are:

• BSI (1999), Clause 6.2 and Annexes A through F

• AGS (2000), Chapter 3
• Simons et al. (2002), Chapter 2
• Randolph and Gourvenec (2011), Section 3.1.1.

55
56 Investigation programs

7.3 GEOPHYSICAL AND GEOTECHNICAL

General information about planning/requirements can be found in:

• API 2SK (2008), Appendix E.2.1.4, Recommended Sequence for Site

Characterization
• API 2GEO (2011), Section 6.2, Shallow Geophysical Investigation
• ISO 19901-4 (2016a), Section 6, Geotechnical Data Acquisition and
Identification of Hazards.
• ISO 19901-8 (2014b), Marine Soil Investigations
• ISO 19901-10 (2021) Marine Geophysical Investigations

Comprehensive information about offshore geophysical and geotechnical

investigation is to be found in Randolph and Gourvenec (2011), sections 3.2
(geophysical investigation) and 3.3 (geotechnical), and is not reproduced
here. Only items relevant to intermediate foundations are discussed in the
remainder of this chapter.
Geophysical investigations over large areas are required for both anchor
spreads for floating facilities (especially in deep water where the footprint is
large, for subsea field architecture and offshore wind farms. They reveal
seabed features (e.g. sand waves, potential obstructions) and subsurface
ground conditions (e.g. layering, faulting and the presence of shallow gas or
hydrates).
Ideally, the geotechnical investigation for an intermediate foundation
should be more comprehensive than that for a comparable pipe pile
foundation, and comparable in scope to that for a gravity base structure
(Gilbert and Murff, 2001). That is, shallow investigation depths, plus at
least one cone penetration test CPT per suction foundation (platform or
support) or suction anchor (group) location, are necessary. In addition,
at least one geotechnical sampling borehole, preferably located at the
most heavily loaded leg location (platform), is required to obtain
“ground truth” data. Field vane shear tests (with both peak and residual
shear strength measurements) are necessary in normally consolidated
“clay” profiles. Investigation depths should be at least the estimated
penetration depth, plus 1.5 diameters (for large diameter foundations)
and to at least around 5 m below estimated penetration depth (suction
anchors).
Figure 7.1 illustrates the changing scales between shallow and deep
water and also, more importantly, between shallow, intermediate and deep
foundations. As shown in Figure 7.1, for a shallow water site, piled plat-
forms are common. Hence the area investigated is concentrated near the
platform, and the depth investigated is around 100 m below seafloor.
However, for deep water, the scale alters dramatically. Surface vessel
anchoring points and seafloor pipelines cover many square kilometres.
Investigation programs 57

Figure 7.1 C
omparative geotechnical survey coverage for shallow or intermedi-
ate and deep foundations for floating and fixed structures (after Kolk
and Wegerif, 2005).

Moreover, since deep piles are no longer used, the depth investigated is
shallow, at most 30 m below seafloor.
Piled foundations are common in shallow water. In this case, geotechni-
cal engineers basically need to know where the competent layer (e.g. dense
sand or rock) is to get enough pile axial capacity. Provided that the pile
impact driving equipment delivers sufficient energy, installation is not a
problem. However, for shallow or intermediate suction installed founda-
tions, geotechnical engineers are interested in relatively subtle variations
in the soil layering within the top 30 m. For example, the weaker layer
may give in-place resistance problems. Similarly, the harder layer may
prevent installation.
In order to minimise risks associated with the hazards listed in Section
6.2, it is occasionally necessary to conduct two separate geophysical sur-
veys. The first is conventional, covering the complete project area. The
second is made at the exact foundation locations and is a high-resolution
3D seismic scan to reveal boulders, dropstones, coral rubble, etc.
58 Investigation programs

7.4 LABORATORY

Comprehensive information about the subsequent onshore and offshore

laboratory testing programs is again to be found in Randolph and Gourvenec
(2011). The relevant sections are 3.4 (soil classification and fabric studies),
3.5 (laboratory element tests) and 3.6 (physical model tests). They are not
reproduced here. Several of the more esoteric tests mentioned, such as X-ray
examination, palaeontology and geochronology and centrifuge modelling,
are not considered for routine foundation design.
Senders and Kay (2002) give requirements for laboratory testing programs
specifically for intermediate anchor foundations in deepwater soft clays. A
more comprehensive list of laboratory tests, covering soil parameters
required for suction caissons in both sand and clays, is in Sturm (2017). The
former includes routine classification tests (water content, unit weight,
Atterberg Limits, particle size distribution/hydrometer etc.), plus index
strength testing (various types of vane tests, pocket penetrometer and fall-
cone) are included in the former article, but not in the latter article.
Chapter 8

Design basis

8.1 GENERAL PRINCIPLES

8.1.1 Introduction
Intermediate foundation design usually requires close cooperation between
geotechnical and other offshore disciplines, particularly structural engineers.
The following general design considerations indicate how intermediate
foundation design can be distinctive from shallow and pile foundation
design.

1. Both installation and in-place response need to be considered, usually

concurrently.
2. Intermediate foundations must be installed to their design embedment
below seafloor. For support foundations, this ensures sufficient in-place
capacity to resist combined vertical, horizontal and moment (VHM)
loads. Proof of sufficient V capacity during installation is not
necessarily proof of capacity under VHM loads. This is due to load
interaction effects: V capacity decreases as H and/or M increase. This
is unlike conventional piled foundations, where there is almost no
interaction between axial and lateral capacity: in such situations, hard
driving can usually provide proof of sufficient axial capacity. Similarly,
anchors need lug levels at the correct depth below seafloor to optimise
pull-out capacity and ensure near-zero rotation. A metre of under-
penetration may result in a significant capacity decrease.
Figure 8.1 gives two commonly encountered examples of installa-
tion criticality.
3. An appropriate Factor of Safety on resistance has to be assigned a
priori in order to determine foundation geometry. For example, con-
sider the support foundation problem shown in Figure 8.1a. This
assumes pure moment (M0) loading. The second design consideration –
that foundations must be installed to their design embedment – uses

59
60 Design basis

factored moment resistance to find the target penetration. Increasing

the factor of safety from, say, 1.5 to 2.5 increases M0 by a factor of 5/3
(= 2.5/1.5) and radius R by 1.186 (= (5/3)(1/3)). This results in a signifi-
cant change in foundation geometry and adds around 20% to the
target penetration depth.

Another example is the anchor foundation shown in Figure 8.1b. Consider

the case L/D = 5, for which optimum lug level (OLL) is at normalised depth
zcl/L = 0.7, at which there is no rotation and optimum capacity (100% of
maximum) is obtained. If H is applied at, say, zcl/L = 2/3 = 0.67, then the
decreased capacity is ≈ 92% of maximum.

(a)

Shallow Rotational Failure

(b)

Figure 8.1 Intermediate foundation capacity – sensitivity examples: (a) support

foundation – shallow rotational failure, M resistance proportional to
R 3 (Kolk et al., 2001); and (b) anchor pile – pull-out capacity decreases
if lug not at optimum depth (Supachawarote et al., 2004).
Design basis 61

4. Intermediate foundation design usually requires close cooperation

between geotechnical and other offshore disciplines, particularly
structural engineers.
5. Finally, set-up (axial friction increase) needs to be assessed for taut or
semi-taut mooring in normally consolidated clay at relatively short
time intervals (say, 10, 30 and 90 days) after installation.

8.1.2 Installation/retrieval/removal
Installation is by dead weight, generally followed by suction assistance.
Retrieval and removal are usually by overpressure, possibly with crane
assistance. Impact driving, vibratory driving or drilled-and-grouted
techniques are sometimes used to achieve high penetrations (L/D up to 10)
in competent sand and (weak) rock. Since they are well understood, these
techniques are not discussed in detail in this section.
The following general principles need consideration in assessing
installation/retrieval/removal of intermediate foundations:

• Driven foundation installation should be analysed using ISO 19901-4

(2016a) Clause 9 (pile installation assessment).
• Suction foundation installation/retrieval/removal can be analysed by
limit equilibrium methods ensuring equilibrium between design loads
and design resistance.
• Soil resistance models include conventional bearing capacity and CPT
cone resistance.
• Penetration mechanism is axial coring.
• Design loads are unfactored.
• Design resistance should be “Best Estimate” (unfactored) and “High
Estimate” (factored).

As sketched in Figure 8.2, during installation, due consideration shall be

given to the possibility of

• soil plug heave (clay and sand)

• soil plug liquefaction (upwards seepage in sand)
• soil plug reverse end-bearing (clay)
• soil plug separation (clay over sand)
• piping (inwards from seafloor around foundation wall)
• hydraulic fracture (clay over sand).

Where these are critical, more complex analysis approaches are required.
Note the following:

• Calculations using alternative analysis methods should include an

explanation of any possible differences due to the method adopted.
• Wall and top plate thickness shall be adequate to resist the stresses
during installation as well as the axial and lateral in-place loads.
62 Design basis

Figure 8.2 Intermediate foundation – geotechnical failure modes during suction

installation.

Comparing Figure 8.2c and 8.2d, note that plug heave occurs in both cases
and, if large enough, will halt installation. However, the heave is due to the
volume of soil displaced by the foundation steel and water pocket formation
respectively. Also, water pockets can occur only in clay over sand profiles,
whereas soil displacement occurs in all soil profiles.
Design basis 63

8.1.3 In-place resistance
Following ISO 19901-4: 2016a Clause 7 for shallow foundations, the gen-
eral principles given below need to be considered in assessing in-place resis-
tance of intermediate driven and suction-installed foundations:

• Bearing failure constitutes any failure mode that could result in exces-
sive combinations of vertical displacement, lateral displacement, or
overturning rotation of the foundation.
• Foundation resistance can be analysed by various methods, including
limit equilibrium and yield surface.
• Limit equilibrium methods shall determine the shape and location of
the critical failure mechanism. These depend on the design loads, soil
stratification and foundation geometry.
• Yield surface methods ensure that factored VHMT loads lie within (or
on) a VHMT foundation resistance envelope.
• Due consideration shall be given to the possibilities of excessive dis-
placement and deformation of the foundation soil. Where these are
critical, more complex analysis approaches are required.
• Calculations using alternative methods of analysis shall include an
explanation of any possible differences due to the method adopted.
• Design loads need to be assessed with due consideration given to the
design life of the foundation.
• Seafloor gradient and/or installation tolerance has to be taken into account
in design. Tolerable foundation tilt and twist should be specified.
• Undrained calculations shall be adopted where no drainage, and hence
no dissipation of excess pore pressures, occurs during loading. This can
occur as a result of the rate of loading or the impermeable nature of the
soil. In contrast, drained calculations shall be adopted where no excess
pore pressures arise during loading. Analysis of foundations subject to
partial soil drainage during the loading event is complex, and specialist
advice shall be sought in these cases. Impact of structural openings,
stiffeners and protuberances shall be taken into account in design.

H and M loads are usually co-planar for single isolated foundations but may
be non-co-planar for unsymmetrical co-joined foundations subjected to
asymmetric VHM(T) loads (e.g. a braced tripod platform). In the latter case,
resultant H and M angles αH and αM need to be calculated, for which the
equations were given in Section 2.6. If |αH − αM|, the angle between H and
M, is significant, then it is usually cautious (i.e. overestimates in-place resis-
tance) to assume they are co-planar for overturning HM loads, and optimis-
tic if HM loads are restoring. In the latter case, options include applying a
higher factor of safety, using 3D FEA on the most critical load cases, or to
seek specialist advice. More details of the effect of non-co-planar MH loads
are given in the next section.
64 Design basis

8.1.4 In-place resistance – non-co-planar MH loads

Figure 8.3a shows schematically two MH resistance ellipse envelopes.
Figure 8.3b shows the HM sign convention: as drawn; both H and M are
positive and act so as to overturn the foundation. Considering the MH
ellipses in Figure 8.3a, it is seen that the effect of non-co-planar MH loads
is to increase resistance in quadrant Q1 (HM overturning) and to decrease
resistance in quadrant Q4 (HM restoring). The two ellipse shapes suggest
that, for the same foundation geometry and soil profile, the MH resistance

(a)

(b)

Figure 8.3 M H resistance ellipses for non-co-planar loads.

Design basis 65

change is less in Q1 (HM overturning) than in Q4 (HM restoring). Resistance

is unchanged when H = 0 and M = 0 (the two MH ellipses intersect on the
H and M axes). The next two figures give examples for quadrants Q1 and
Q4 respectively.
Figure 8.4 shows a stubby (say, L/D = 1) support foundation subjected
to overturning HM loads (i.e. loaded in quadrant Q1). The likely failure
mechanism is one of shallow rotational failure, with an approximately
spheroidal/circular failure surface passing through the foundation tip.
Since the H component is offset from the centreline, non-co-planar HM
loading changes the coordinates of the critical (lowest FOS, usually
obtained by a grid search) rotation point and increases the failure arc
radius. The latter increases the available M resistance. This is for shallow
rotational failure, but the same argument can also be applied to the inter-
mediate and deep rotational failure modes shown in Figure 2.2a and 2.2b.
Hence non-co-planar HM load increases resistance in quadrant Q1 (HM
overturning).
To illustrate quadrant Q4, Figure 8.5 is for a long (L/D = 6, say) caisson
anchor subjected to pure HM loading (i.e. Vload = 0). Assuming lug level is
at optimum depth below seafloor, the lateral load Hlug equals Hmax, the
maximum available lateral resistance. Because the foundation is rigid,
load Hlug is equivalent to loads H = Hlug and restoring M = - Hlug zlug,opt at

(a)

Figure 8.4 Non-co-planar HM loads: (a) Q1 example (overturning HM loads).

66 Design basis

(b)

Figure 8.4 Continued: Non-co-planar HM loads: (b) example of projecting H and M.

seafloor. This is the (optimum) situation in Figure 8.3. If a non-co-planar

H load is applied at seafloor, then it is obvious that conditions are no lon-
ger optimal, and available H resistance decreases. Again, the argument can
be extended to combined VHM (instead of HM) loading. Therefore, in
quadrant Q4 (HM overturning), non-co-planar HM load decreases avail-
able resistance. This situation is not dissimilar to that of anchor pile twist
(see Section 10.9), where the maximum available V resistance decreases
due to torque T load.
Design basis 67

(a)

(b)

Figure 8.5 Non-co-planar HM loads: (a) Q4 example (restoring HM loads); (b)

example of projecting H and M.
68 Design basis

8.1.5 In-place response
Following ISO 19901-4 (ISO 2016a), the general principles listed need to be
considered in assessing in-place response/serviceability of intermediate
foundations:

• Design can be based on in-place response/serviceability (rather than

in-place resistance) criteria, whereby the foundation deformation is
assessed against allowable movement criteria. The appropriateness of
adopting this approach will depend on the type of structure and its
installation.
• The selection of appropriate soil moduli (especially considering strain
dependency and cyclic loads) is essential in calculation of in-place
responses (i.e. serviceability limit states).

8.2 SIGN CONVENTIONS, NOMENCLATURE AND

REFERENCE POINT

Vertical (V), horizontal (H), overturning (M) and torsional (T) loads are
centric and act at a geotechnical reference point (RP), which is at the mid-
point of the foundation at seafloor level; see Figure 8.6. This is the point of

(a) (b)

(c)

Figure 8.6 S ign conventions, nomenclature, reference point and seafloor stiffness
matrices for analysis of intermediate foundations.
Design basis 69

(d)

Figure 8.6 (Continued)

structural load transfer. Loads H and M are assumed to be co-planar.

Associated displacements, vertical (uV), horizontal (uH), overturning (θM)
and torsional (θT), refer to the same reference point.
Note that the geotechnical reference point (RP) is the same for shallow
and intermediate foundations.
Intermediate suction foundations usually have “stick-up” above seafloor
to allow for general seafloor slope, local seabed variations and steel dis-
placed in soil; see Sections 9.10.2 (clay plug heave) and 9.11.1 (sand plug
heave). In addition, platform foundations may have platform leg(s) offset
from the foundation midpoint, in order to improve barge transport/barge
stability. These features mean that specified loads may not act centrically
and at seafloor level. Since intermediate foundation geometry (generally
outside diameter, stick-up, geotechnical RP and structural RP) varies during
design, this challenge cannot be met by a rigid beam element connecting the
two RPs. In such cases (where RPs differ), structural loads should be trans-
formed to the geotechnical RP. Similarly, foundation stiffness matrices sup-
plied by the geotechnical engineer should be applied by the structural
engineer at the correct offset.

8.3 FOUNDATION STIFFNESS AND FIXITY

Unlike pile foundations, intermediate foundations respond essentially rig-

idly, i.e. a constant rotation θxz with depth below seafloor, under lateral HM
loads (see Table 2.1 and Figure 8.6b).
For structural analysis purposes, the seafloor foundation stiffness matrix
is a simple 4 × 4 symmetrical matrix expressing the relationship between
VHMT loads and corresponding displacements δz, δx and rotations θxz, Ψxy
as shown in Figure 8.6c. Note the following:
70 Design basis

• Both axial V and torsion T loads are assumed uncoupled from the
lateral (HM) component. This is unlike in-place capacity (Section
10.10, Resistance under Combined VHM(T) Loads) where there is
coupling between V, T and HM. Both V and T resistances are adversely
affected by HM loads, V resistance more than T.
• The seafloor foundation stiffness matrix contains coupling between
the horizontal and rotational terms. This is because the rotation point
is located below seafloor when H and/or M is applied
• The rotation point is assumed to be on the foundation centreline – i.e.
small lateral offset is ignored.
• Torsion loads T are usually small and neglected.

Of importance is the 2 × 2 lateral KHM sub-matrix (Figure 8.6d). Occasionally,

some structural jacket analysis programs cannot cope with the off-diagonal
terms. In such cases, an equivalent KHM sub-matrix may be redefined using
lateral spring KH and rotational spring KM, both positioned at the founda-
tion rotation depth h. A rigid beam element connects these springs to sea-
floor. Terms KH, KM and h are simple functions of Kxx, Kxθ and Kθθ (see
Figure 8.7a).

(a)

(b)

Figure 8.7 S eafloor foundation lateral stiffness expressed as (a) horizontal and
moment spring stiffnesses K H and K M , and (b) non-linear stiffness
through springs with K 1 and K 2 .
Design basis 71

Another situation occurs when lateral spring load-deformation (“p-y”)

data need modification. Examples include introducing non-linearity (to
model varying soil stiffness) or lateral y-shifts (due to spudcan installation).
It is easiest to implement these modifications in structural jacket analyses by
using a pair of (non-linear) lateral support (“p-y”) springs. The first spring
has lateral stiffness K1 at seafloor, the second stiffness K2 at depth h2. These
two springs are connected by a vertical rigid beam element. Again, K1, K2
and h2 values are simple functions of Kxx, Kxθ and Kθθ (Figure 8.7b). Lateral
stiffnesses Kxx, Kxθ and Kθθ may be assessed by various methods, including
elastic solutions (see Section 11.2.2 on immediate displacement) or from
non-linear FEA (e.g. Dekker, 2014).
Finally, more complex studies (e.g. wind turbine monopiles under
omni-directional variable cyclic loads) may require the use of a macro
element (a non-linear, force-resultant plasticity model, typically a yield
surface and back-bone curves accounting for displacement accumula-
tion due to soil stiffness). This is specialised non-routine engineering,
generally for detailed case-by-case design using a project specific foun-
dation geometry and soil profile.

8.3.1 Seafloor VHMT loads

Note that:

• axial V load is uncoupled from lateral (HM) load component. In other

words, foundation lateral fixity condition does not affect V load.
• lateral H load is statically indeterminate. Hence, the global shear load
is equally distributed between all platform legs.
• moment load M is large/important for platform foundations. Note
that this is generally overturning for un-braced structures (e.g. a self-
installing platform or jack-up rig) but may be restoring for braced
structures (e.g. a conventional platform or a braced met mast or well-
head tripod).
• torsional (T) load is usually small.

8.3.2 Foundation lateral and rotational fixity

As can be seen from the stiffness matrix in Figure 8.6d, lateral fixity
conditions (δx θxz) influence lateral HM load magnitudes.
Table 8.1 lists some foundation rotational lateral fixity conditions. When
foundations are not “pinned”, they attract both horizontal (H) and moment
(M) loads. Usually, design practice is as follows: during preliminary design,
fully fixed analyses (kxx = kθθ = ∞) are performed in order to maximise M
load, and then to decrease the stiffness values during final design once a
foundation stiffness matrix has become available.
72 Design basis

Table 8.1 Foundation Lateral Fixity, Rotation and M loads.

Rotational Stiffness k θθ
Foundation Fixity Condition Rotation θ xz M Loads
“Pinned” k θθ = 0 M=0
θ xz = maximum
“Intermediate” 0 < k θθ < ∞ 0 < M < maximum
“Fully fixed” k θθ = ∞ M = maximum
θ xz = 0

Project experience suggests that, for intermediate foundations, M load is

relatively insensitive to rotational stiffness kθθ (and kxx).

8.4 LOAD AND MATERIAL FACTORS

Offshore foundations have been historically subdivided into shallow foun-

dations (Table 8.2) and piles (Table 8.3). Currently there is no industry-wide
consensus for (a) definitions of shallow and pile foundations and (b) into
which category intermediate foundations appear in Tables 8.2 and 8.3.
Hence, factors of safety for support foundations appear in (shallow) Table
8.2, whereas factors for suction anchor piles are taken from (pile founda-
tion) Table 8.3. Note that ISO 19901-7:2013 (ISO, 2014a) lumped FOS
values are load angle dependent, more details of which are given in
Figure 8.8. Unlike shallow and deep (pile) foundations, there have been no
full-scale VHM load tests on intermediate foundations to verify in-place
resistance models.
Unlike most USA (API) codes, most European codes differentiate between
environmental (live) and static (dead) loading, and also usually specify dif-
ferent load combinations for analysis. This partial factor design (PFD)
approach, also known as LRFD – Load and Resistance Factor Design, is
more rational than a lumped FOS approach. Figure 8.9 compares both PFD
and lumped FOS approaches for support foundations (using a VHM enve-
lope design) and anchor foundations (VH envelope design). In particular,
note that where dead load opposes live loads, European code partial load
Design basis 73

factors may be less than unity. This is useful for anchor foundations, as
sketched in Figure 8.9d. Partial load factors per guideline or regulation are
included in Tables 8.2 and 8.3. They exclude dead load factors.

8.5 COMMENTARY

Extreme loading conditions relate to probability of occurrence within a

fixed time period – e.g. a 10-year summer storm for mudmats/shallow foun-
dations and 100-year (winter) storm for platform foundations.

8.6 CLOSURE

Intermediate foundation design, especially involving suction installation, is

very different from that for both shallow and deep (pile) foundations. These
major differences have been listed; the most significant of these are possibly:
(a) load interaction, since they are subject to high HM loads, and they must
reach their design penetration below seafloor; (b) even though installation is
usually more critical than in-place resistance, both need to be studied simul-
taneously; (c) design is multidisciplinary, frequently involving structural
engineers to verify steel integrity during installation.
Finally, Figure 8.2 (depicting geotechnical failure modes during suction
installation) is a key figure. It neatly illustrates that more things can go
wrong during installation than when in-place.
74
Table 8.2 P artial Load and Soil Resistance Factors for Offshore Shallow Foundations.

Design basis
Load Soil Resistance

Existing Guideline or Recommended Recommended

Regulation Loading Recommended Partial Partial Resistance Partial Material Overall Factor of Safety b (FOS
(Foundation Type) Condition Load Factor (γ L or φ L ) Factor (φ m ) Factor (γ m ) a ≈ γ L. γ m)

API RP-2GEO As API RP-2A WSD

(2011) (Shallow Not Section 7.3.1
foundations) defined Uplift: 2.0
Section 7.3.3.2.1
API RP-2A, WSD, Bearing: 2.0
21st Ed. (2000) Not Sliding: 1.5
(Shallow defined Section 6.13.4
Foundations)
DNV (1992) 1.30 (env.) 0.77 Bearing: (env.):
(Jack-ups) 1.00 (static) 1.20 (eff stress) Bearing:
Not Table 8.1 1.30 (tot stress) ≈1.56 (eff stress)
defined Sliding: 1.30 1.69 (tot stress)
Table 8.1 Sliding:
1.69
ISO 19902:2007(E) 1.35 (env.) c 1.69 (env.)
19901-4:2016 ISO 19902 Section 1.37 (static)
(Shallow A.9.9.3.1
Extreme
Foundations) 1.10 (static)
ISO 19902
Section 9.10.3.2 1.25 d
1.215 (env.) e 0.80 ISO 19901 Section 1.52 (env.)
ISO 19902 Section 7.3 1.62 (static)
A.9.9.3.1
Operating
1.30 (static)
ISO 19902
Section 9.10.3.2

Notes:
Refer to actual documents for detailed explanations/definitions/equations.
a API applies φ to resistance. DNV and ISO apply γ to tan(φ’) (sands, effective stress) or s (clays, total stress).γ is equivalent to 1/φ in case of clays, e.g. DNV mate-
m m u m m
rial factor φm of 1.30 is equivalent to API resistance factor γm of 0.77.
b FOS assumptions: (i) either all environmental or all static load (ii) clay (not sand) profile.
c ISO 19902: Section A.9.9.3.1 gives default load factor γL = 1.35 is for GoM. According to Section A.9.9.3.3, higher γL values may be in operation elsewhere: e.g. γL =
1.59 (NW Shelf, AUS) and γL = 1.40 (North Sea).
d ISO 19901: material factor γ = 1.25 is for combined (VHM) loading. For pure V loading γ = 1.5 (also in Section 7.3).
m m
e ISO 19902: 0.9 * γ (default load factor γ = 1.35).
L L

Design basis
75
76
Table 8.3 Partial Load and Soil Resistance Factors for Offshore Piles.

Design basis
Load Soil Resistance

Recommended
Existing Guideline or Partial Recommended Overall Factor of
Regulation (Foundation Recommended Partial Resistance Partial Material Safety b (FOS ≈ γ L .
Type) Loading Condition Load Factor (γ L or φ L ) Factor (φ m ) Factor (γ m ) a γ m)

API RP-2GEO (2011)

As API RP-2A WSD
(Piles, Section 8)
1.5
API RP-2A, WSD, 21st Extreme
Section 6.3.4
Ed. (2000) (Platform
2.0
Piles) Operating
Section 6.3.4
API RP-2T (2nd Ed., Axial capacity,
1997) (Tension Leg extreme loading 1.5*B e
Platforms)
Axial capacity,
2.0*B e
Operating
Axial capacity, One
1.5*B e
line damaged
Lateral capacity As API RP-2A WSD
API RP-2SK (3rd Ed., Axial 2.0
Intact
2008) f (Permanent Lateral 1.6
Mooring Anchors) Axial 1.5
ISO 19901-7 (2013)
Damaged
Table 7, Section Lateral 1.2
10.4.3
DNV (1992) (Jack-ups) Bearing:
NPD (1992) (Piles and Extreme 1.30 1.20 (eff. stress) 1.69
Anchors) 1.30 (tot stress)
1.35 (env.) c
1.25
ISO 19902:2007(E) 1.10 (static) 1.69 (env.)
Extreme ISO 19902
19901-4:2016 (Piles) ISO 19902 Section 1.37 (static)
Section 17.3.4
9.10.3.2
1.21 (= 0.9*γ f,E )
(env.) d
ISO 19902 Section 1.50
1.81 (env.)
Operating A.9.9.3.1 ISO 19902
1.95 (static)
1.30 (static) Section 17.3.4
ISO 19902 Section
9.10.3.2

Notes:
Refer to actual documents for detailed explanations/definitions/equations
a API applies φ to resistance. DNV and ISO apply γ to tan(φ’) (sands, effective stress) or s (clays, total stress).γ is equivalent to 1/φ in case of clays, e.g. DNV mate-
m m u m m
rial factor φm of 1.30 is equivalent to API resistance factor γm of 0.77.
b FOS assumptions: (i) either all environmental or all static load (ii) clay (not sand) profile.
c ISO 19902: Section A.9.9.3.1 gives default load factor γ = 1.35 is for GoM. According to Section A.9.9.3.3, higher γ values may be in operation elsewhere: e.g. γ =
L L L
1.59 (NW Shelf, AUS) and γ = 1.40 (North Sea).
d ISO 19902: 0.9 * γ (default load factor γ = 1.35).
L L
e API RP-2T: Bias factor B (recommended minimum 1.5) on API RP2A (WSD) FOS (i.e. FOS=2.25 for extreme/one line damaged reduced environmental loading and 3.0

Design basis
for operating).
f API RP-2SK: factor of safety for maximum anchor load determined from dynamic analysis.

77
78 Design basis

Figure 8.8 Factor of safety versus failure mode for intermediate anchor
foundations.

Figure 8.9 C
omparison of lumped factor of safety (WSD) and partial factor
design (PFD, LRFD) approaches for support and anchor foundations.
Chapter 9

Installation, retrieval and removal

9.1 INTRODUCTION

This chapter is long, emphasising the fact that installation, particularly those
methods using suction, can be more challenging than both in-place resis-
tance (Chapter 10) and in-place response (Chapter 11). Topics relating to
installation using suction assistance, driven, vibratory and drilled and
grouted installation are considered.
Suction installation topics resemble those for shallow foundations,
whereas those for intermediate offshore wind turbine monopile foundations
are essentially the same as piles.

9.2 GENERAL CONSIDERATIONS – SUCTION

ASSISTANCE

Intermediate suction foundations are installed by a combination of self-

weight and suction/pumping, and the top plate vent is open. During instal-
lation, tilt is small, generally less than 5° at final penetration. In addition,
twist is essentially zero. Hence, a simple resistance model is used for pene-
tration, namely an axially coring (unplugged pipe pile) model.
Installation must always be assessed, and is a two-stage process:

• self-weight penetration using foundation dead weight

• suction penetration using pumping.

Retrieval usually also needs to be assessed. In case requirements are not met
(e.g. target penetration depth, foundation verticality or orientation), the
installation contractor has to use overpressure to extract foundation, bump
over (possibly reposition) and reinstall.
Removal is also usually addressed. This is a consequence of site clearance
requirements. Removal is performed using overpressure, since this is less
expensive than soil removal around the foundation outside perimeter and
cutting off the foundation steel, say, 1.5 m depth below seafloor.

79
80 Installation, retrieval and removal

Installation feasibility studies include the following general considerations:

• sufficient water depth to create the maximum underpressure

• sufficient steel thickness to resist maximum underpressure (structural)
• additional underpressure costs less than additional preload (steel weight)
• foundations with a high diameter and low embedment ratio are easiest
to install using underpressure
• generally feasible in clay, sand and sand over clay profiles
• clay over sand and hard glacial till profiles may be problematic
• cemented sand, rock and “gravel, cobbles and boulders” in sand are
generally not feasible.

The conventional procedure for installation analysis assumes static (force)

equilibrium, whereby load = resistance. Load is the sum of foundation
weight and the product of underpressure and cylinder inner cross-sectional
area. The resistance model assumes rigid-plastic ground behaviour and con-
sists of the summation of unit skin friction over the embedded length and
unit end-bearing at the toe and internal stiffeners. Internal friction may be
substantially reduced above ring stiffeners in lightly to heavily overconsoli-
dated clay if soil is remoulded as it passes over the stiffeners, or if it extrudes,
leaving a free-standing column of soil above the stiffener (e.g. McNamara,
2000). In addition, for sand, the effects of water flow are to reduce internal
friction and tip resistance, and to slightly increase external friction resis-
tance (e.g. Houlsby and Byrne, 2005). Overall factors to account for flow
for installation pressure predictions are also presented in Alderlieste and
Van Blaaderen (2015).
Installation assessments assume that both load and resistance act cen-
trally and that the foundation “cores” through the various soil layers during
installation, i.e. it remains vertical. However, foundations may possibly tilt
(become non-vertical) and/or twist (misalign in plan). Platform-type support
foundations, especially if the superstructure is braced, are normally not
problematic regarding tilt and twist. This is provided that they are installed
on an essentially level seafloor (slope angle less than tilt tolerance) and the
suction-assisted penetration depth is sufficient to apply tilt corrections using
varying underpressures per foundation (or possibly foundation compart-
ment). However, isolated (single) foundations are more susceptible to tilt
and twist – possible causes are items causing eccentric resistance such as lug
and chain presence, crane wire torsion and ground inhomogeneity. If tilt
criteria are very strict, a possible solution is to install either an instrumented
braced foundation cluster (or a single foundation containing compart-
ments), and to vary pressures per foundation (or foundation compartment)
in order to minimise/correct tilt during installation. Excessive foundation
twist requires retrieval and re-installation (e.g. Newlin, 2003b).
There is no free-fall (uncontrolled penetration) risk during installation.
During self-weight penetration, a crane or something similar takes the load,
Installation, retrieval and removal 81

and penetration is usually monitored. During the subsequent suction assis-

tance stage, the suction pump extracts water from the interior at a specified
flowrate. Hence, both stages are displacement (not load) controlled. There is
no risk of uncontrolled penetration occurring should less competent soil be
encountered at foundation tip level during installation. This is unlike pile or
jack-up spudcan foundations.
Parameter values for the ground/foundation installation models are usu-
ally estimated in two ways – a “Best Estimate” and a “High Estimate”. The
High Estimate model results are generally used to ensure that the suction
foundation is likely to achieve the required penetration depth. Installation
parameters are not usually appropriate for in-place analyses. Conventional
installation models generally rely upon the results of field measurements,
model tests and theoretical analyses. For clay profiles, relevant measure-
ments are those commensurate with penetrating bodies (i.e. CPT) or large
displacements (e.g. laboratory remoulded or residual undrained shear
strengths, su,rem or su,res). For sand profiles, the field data base size is small,
and actual installation mechanisms are complex and not yet completely
understood. Hence, installation predictions are less reliable for sand than
for clay profiles. Corresponding removal models have even higher parame-
ter values than for installation. This is due to set-up (outer friction increase)
with time: for example, normally consolidated clays regain their original
peak undrained shear strength su – this can be up to a factor 3 higher (pos-
sibly more) than the corresponding su,rem or su,res value used for installation.

9.3 GENERAL CONSIDERATIONS FOR MONOPILE

INSTALLATION – IMPACT DRIVING, VIBRATORY
AND DRILLED AND GROUTED

Most installed monopiles were driven using impact hydraulic hammers

(Figure 9.1). This installation method is often the most versatile (can be
considered for soft ground to weak rock) and cost-effective solution. At the
time of writing, hydraulic hammers with a rated energy going up to 4000 kJ
are available on the market. One limitation of this installation method is the
noise emission and its impact on marine mammals. Depending on the local
regulation, marine mammals protection can require either a very costly
noise mitigation system (e.g. bubble curtain), a limitation of the hammer
size, a restriction of operation during certain periods or monitoring of
marine mammal presence.
Vibratory installation can provide an alternative to reduce the noise emis-
sion during monopile installation as vibro hammers are much less noisy,
such that costly noise-mitigation measures are not necessary. The industry is
gaining more experience with vibro installation, and it is no longer required
to complete the last metre of penetration using an impact hammer, as it used
to be in the past. As an example, the main objective of the joint industry
82 Installation, retrieval and removal

(a)

(b) (c)

Figure 9.1 Impact

driving offshore wind turbine monopiles: (a) cross-over anvils
(IHC, 2010), (b) offshore installation examples (Courtesy of Boskalis),
(c) hammer as used offshore (Courtesy of Boskalis).

Gentle Driving of Piles (GDP) project is to develop a novel technique for

gentle pile driving that simultaneously improves drivability, reduces noise
emission and preserves satisfactory geotechnical performance during opera-
tions; see also Metrikine et al. (2020). Vibration also tends to create less
fatigue in the steel than impact driving. Another advantage of the vibro-
installation is the possibility to lift the monopile with the equipment without
Installation, retrieval and removal 83

Figure 9.2 D
rilling from a jackup using RCD (Reverse Circulation Drilling) tech-
nique (Courtesy of Fugro).

requiring any external guide (e.g. gripper). Vibro-hammering is, however,

not advised in the presence of hard or cemented layers or inclusions and
may also not be efficient or possible in certain types of clay.
In the presence of rock layers or thick flint beds, drilling (see the example
in Figure 9.2) might be required as a mitigation measure (e.g. in case of
potential driving refusal) or as the main installation means. An alternative
hybrid installation method of driving and drilling, the so-called Drive-Drill-
Drive or 3D method, can be considered to reduce the inner friction or the tip
resistance. This can be a cost-effective option by eliminating the need for
grout and reducing the size of the hammer spread, and it offers a contin-
gency option for unexpected ground conditions. Under-reaming (i.e. using
an enlarged drill-bit to drill to the outside pile diameter) can generally be
used to reduce the risk of pile buckling.
However, if the relief drilling operation needs to be performed several
times during the installation, the repeated change of tool (i.e. switching from
hammer to drilling machine) will considerably increase the installation cost.
In addition, due to a large diameter/thickness ratio, monopiles are more
susceptible to extrusion buckling than more classical flexible piles. Driving
or vibro-hammer may therefore represent a risk in the presence of hard lay-
ers, extremely dense sand or very heterogeneous conditions. A drilled and
grouted installation can be a safer and more cost-effective method in some
cases. Drilled and grouted pile installation involves drilling to the target
depth with a diameter larger than the pile, lowering the pile, and grouting
the annulus between the pile and the hole. In the presence of overburden
84 Installation, retrieval and removal

layers or upper fractured rock, a support will be required to avoid hole col-
lapse during drilling. Drilled and grouted installation has seldomly been
used for monopile installation but could become more common as more
complex sites are encountered. Two monopiles of 3.5 m diameters were
drilled and grouted on the Blyth offshore windfarm park; these two piles
were recently decommissioned. More drilled and grouted monopiles of
larger diameter (7 m diameter) should be installed on the French coast in
2022.
Table 9.1 summarises OWT monopile installation feasibility in various
soil and rock profiles. More details are given in Section 9.13.

Table 9.1 Comparison OWT Monopile Installation Feasibility in Hard Ground.

Installation method

Ground Drilled &

profile Impact driving Vibratory grouted Remarks
Very dense Possible Possible Would
sand (D r ≥ require
85%) support of
a casing
beforehand
to reach
bedrock
Hard clay Possible To be Possible, but ASTM 2487
(s u ≥ 400 proven unlikely (2011) /2488
kPa) (2009)
Chalk Possible Possible, Possible CIRIA (2002)
depending on but depending C574
the chalk unlikely on the classification
grade and chalk grade
presence of (see weak
thick flint rock) or if
beds flint bed
are present
Calcareous Possible Impossible Possible Kolk (2000)
Sediments
Weak – MW Feasibility to Impossible Possible ASTM 2487
(1.25 ≤ σ c ≤ be checked (2011) /2488
12.5 MPa) (depending (2009)
on other
aspects such
as rock
porosity and
fracturation)
MS – ES Impossible Impossible Possible
(σ c > 12.5
MPa)
Installation, retrieval and removal 85

9.4 BEST AND HIGH ESTIMATES – INSTALLATION

RESISTANCE

The additional resistance of internal/external stiffening plates should be

taken into account by adding resistance terms accounting for skin friction
and end-bearing on each plate.
Installation assessments must use High Estimate (not Best Estimate) soil
resistances. Figure 9.3 shows this schematically. The top row, Figure 9.3a, is
for in-place resistance assessments (Chapter 10) and plots probability den-
sity function (PDF) against (wind, wave and current) loads and (soil) resis-
tances. The PDF for both load and resistance is the same shape, because
both have coefficient of variation (COV) values of ≈ 0.3. Because loads and
resistances are both factored (see Section 8.4 on load and material factors),
the mean values are sufficiently far apart that the probability of failure (i.e.
factored loads exceeding factored resistance) is suitably small.
The bottom row, Figure 9.3b and 9.3c, consists of the corresponding PDF
graphs for installation (this chapter). Resistance coefficient of variation
(COV) is still around 0.3. However, unlike in-place resistance, loads have an
extremely low COV, around 0.01. This is because suction pressure and dead
weight can be accurately determined. In addition, installation requires fail-
ure (loads exceed resistance). Figure 9.3b sketches the situation if the Best
Estimate soil resistance and load coincide. It can be inferred that there is a
significant refusal risk: for example, at probability p(f) = 1.00 on loads, the
corresponding probabilities on resistances are < 0.95. However, Figure 9.3c
shows that if a High Estimate soil resistance (say p(f) = 0.95) is used, the
corresponding load probability is 0.5, implying that the risk of refusal has
been significantly reduced.
Note that installation assessments should report both High Estimate and
Best Estimate soil resistance profiles, together with the corresponding suc-
tion pressures. This is to avoid on-site embarrassment when field suction
pressures plot below the High Estimate.

9.5 UNDER-PENETRATION AND OVER-PENETRATION

9.5.1 General
This section discusses possible installation measures for both under-penetra-
tion and over-penetration of suction foundations.
The objective is to install the intermediate suction foundation (almost) to
its design (target) penetration depth (TPD) below seafloor. Organisations
have differing approaches regarding target penetration depth tolerance.
These account for installation risks such as depth measurement inaccuracy,
excessive soil plug heave, encountering obstructions or exceeding the allow-
able underpressure. Examples include (a) final penetration depth cannot
exceed, say, 0.3 m above target penetration depth and (b) defining minimum
86 Installation, retrieval and removal

Figure 9.3 P
robability density functions comparing loads and (soil) resistances
for (a) in-place resistance and installation using (b) Best Estimate and
(c) High Estimate soil resistances.
Installation, retrieval and removal 87

final penetration depth as having ∆FOS (defined as factored resistance/fac-

tored loads) value of 1.00, whereas target penetration depth has a higher
∆FOS value (say 1.05). Suction foundations use a combination of self-weight
(and possibly preload) and underpressure to achieve target penetration depth.

9.5.2 Under-penetration
Under-penetration is when self-weight and underpressure alone cannot
achieve target penetration depth. It usually occurs in competent soils, such
as very dense sands or glacial till, where the soil resistance exceeds the High
Estimate value. Increasing the underpressure may cause structural damage.
None of the following measures will work if there is a genuine obstruc-
tion – e.g. a boulder causing refusal. In such cases, retrieval and re-installa-
tion is the only possible measure.
Possible measures to increase the final penetration depth include:

• impact driving
• vibration
• jetting (high pressure)
• water injection (low pressure fluidisation)a
• air lift
• dredge pumpb
• explosions
• free-fall (aka drop fall)c
• cyclic loading
• coat steelb
• friction breakers
• bevel.

Notes:
a water injection: This fluidises sand at tip level. In clays, it lubricates the

clay-steel wall interface.

b dredge pump and coat steel: Centrifugal dredge pumps can remove the

liquefied soil plug contents, and the geotechnical engineer has to design for
disturbed soil around the foundation. Similarly, paint, bitumen or other
coatings significantly reduce unit skin friction. Both measures are likely to
affect axial more than lateral in-place resistance.
c free-fall: More details, including equations and an example, are given in

Section 9.9.3 (on free-fall penetration).

9.5.3 Over-penetration
Over-penetration is when self-weight penetration exceeds target penetration
depth, i.e. when pile weight Wsub exceeds the Low Estimate soil resistance at
TPD. Over-penetration is generally associated with anchor pile foundations in
88 Installation, retrieval and removal

weak clay – the pile weight Wsub is heavy and/or the pile-soil α value is lower
than the Low Estimate (e.g. the soil is usually more sensitive than anticipated).
The soil resistance increases significantly if the top plate meets the clay
soil plug, which is above seafloor due to heave. When this happens, the fail-
ure mode changes from “coring” to “plugged”. The “plugged” : “coring”
installation resistance ratio, ηR, is a function of pile L, D, soil su profile and
pile-soil α. For a normally consolidated clay, at penetration L, ηR is typically
around 2.5 but could vary between 1.5 and 4.5. This resistance increase may
be sufficient to halt self-weight penetration (SWP). In this case (Rplugged ≥
Wsub) we have L < self-weight penetration < Ltot, and the top of the anchor
pile remains above seafloor.
If the resistance increase is not enough (i.e. Rplugged < Wsub) then the pile
would disappear below seafloor during self-weight penetration, and target
penetration depth is exceeded.
Possible measures to decrease the final penetration depth during SWP
phase include:

• closing the top plate vent(s) just before reaching target penetration depth
• adding an external protuberance (pile outstand/collar) at height above
tip equal to target penetration depth
• decrease Wsub such that penetration equals target penetration depth,
and await set-up.

Ad top plate vent closure: This can be done during offshore installation.
Ad adding external protuberance: This needs to be considered during
design. It cannot be done offshore.
Ad decreasing Wsub: This may be achieved using crane hook load tension. It
is a rather extreme measure – the time required depends on the required fric-
tional resistance increase, and the associated strength gain–time relationship.
Alternatively, buoyancy or flotation blocks can be attached to the foundation
to reduce the submerged weight.

9.6 DIFFICULT SOIL PROFILES

The following paragraphs give more details of soil profiles likely to be dif-
ficult/problematic using underpressure. They are listed in easiest (feasible) to
hardest (impossible) order.
Loose sand – easy to install but in-place resistance (liquefaction)
considerations
Clay, sand and sand over clay – generally feasible
Calcareous/carbonate clays require no special considerations. This is
unlike carbonate sands (and silts), which are usually cemented and weaker
than silica/quartz sands.
Sand with occasional clay seams/layers and sand with occasional peat
seams/layers – Thin clay (or peat) seams or layers, if sufficiently thick and
present over the whole foundation footprint, may cause a flow seal and
Installation, retrieval and removal 89

prevent inner friction reduction. Hence, installation may be problematic. At

the time of writing there is no consensus as to what clay thickness actually
causes no flow – estimates vary between 0.01 m and 0.1 m. Complicating
factors are (a) a CPT tool cannot detect clay seams, detection limit is around
2 to 4 tool diameters (b) whether or not the clay is continuous across the
majority of the foundation footprint – the thicker it is, the more likely it is
not a lens and increases the probability of this happening (c) whether or not
the clay becomes punctured by the foundation, causing a hydraulic leak
(inducing fluid flow) at the soil/foundation interface. An additional consi­
deration for peat is long-term creep settlement.
Hard glacial till – these are clays containing sand and gravel. Installation
may be problematic/impossible if there are frequent cobbles or boulders –
foundation tip integrity is likely to be compromised. Design challenges are
usually a high underpressure during installation, a domed (not flat) top, and
a relatively high wall thickness. An example is the Skarv Field suction
anchors (Langford et al., 2012). Soils were soft marine clays underlain by
hard glacial till (su up to 340 kPa). The caissons were 5 m diameter, up to
14.5 m high and about 40–50 mm wall thickness. The domed tops were
designed to withstand the maximum client specified value of 1500 kPa
underpressure. Measured values were up to 1000 kPa during installation.
Carbonate sands – these contain > 90% calcium carbonate (CaCO3).
They are usually cemented and of variable cementation. They are also
weaker than silica/quartz sands. Note that cementation may have been
destroyed by borehole drilling/sampling. Hence the main risk is refusal
(or tilt) during installation, if highly cemented sand is unexpectedly met
under (part of) the foundation tip. In addition, in-place axial skin friction
is low.
Clay over sand – installation may be problematic because, when the foun-
dation tip enters the underlying sand, there may be no water flow to reduce
friction and tip resistance. In such cases, key design challenges are limited
penetration, higher underpressures and a higher-than-usual wall thickness.
Due to limited penetration, in-place resistance may be insufficient. This is
discussed further in Section 9.12 (on installation in mixed sand and clay).
Cemented sand, rock, cobbles and boulders in sand – installation is gener-
ally not feasible, as resistance is too high and there is insufficient self-weight
penetration. Use other installation methods – e.g. impact driving. A com-
mon example is caprock in the Middle East.
Peat – usually of limited thickness and encountered above target penetra-
tion depth. Main design challenges include:

• characterisation (e.g. difficult to accurately determine peat strength)

• likely to seal fluid flow (if sandwiched between sand layers)
• (very) low specific weight (risk for the suction process, both seal and
installation)
• long-term response (creep settlement component)
• possible low pH value (additional cathodes may be necessary).
90 Installation, retrieval and removal

9.7 MAXIMUM PUMP UNDERPRESSURE

Houlsby and Byrne (2005) have listed the following practical limits to the
maximum attainable suction pump underpressure:

• the absolute pressure at which the water cavitates (usually a small

fraction of atmospheric pressure)
• the minimum absolute pressure that the given pump design can achieve
• the minimum relative pressure that the pump can achieve.

In shallow water depth (less than ≈ 30 m), taking allowable underpressure

inside the foundation as atmospheric or vacuum can make an appreciable
difference to installation in difficult soil profiles, especially competent ones.
Installation feasibility may hinge on maximum underpressure – and whether
or not the suction pump can cope with possible cavitation (see Section 4.3).
Hence, unless there is convincing evidence to suggest otherwise, it is cau-
tious to take atmospheric (rather than vacuum) pressure.
Water depth is usually related to Lowest Astronomical Tide (LAT).

9.8 PENETRATION RESISTANCE ASSESSMENT

9.8.1 Introduction
There are two main methods of assessing penetration resistance R:
CPT q c based f k f q c sand and clay ,
(9.1)
q k p q c sand and clay

classical bearing capacity alpha / beta

f su clay , f vo sand ,
(9.2)
q N csu clay , q N qvo sand

Of these, the CPT method is considered to be the more reliable/popular.

This is especially true for sands, where resistance assessments are critical.
This is because the CPT method directly correlates penetration resistance R
to cone resistance qc, whereas two steps are needed for classical bearing
capacity, firstly assessing φ' from relative density Dr (or qc) test data, and
then selecting an appropriate Nq value; in this process Nq increases expo-
nentially with φ'. However, the classical bearing capacity method can serve
as a useful double check for more challenging sites; for example, those
where friction reduction above internal stiffeners has to be considered. A
disadvantage of the CPT method is that two sets of parameter values are
needed (qc, etc. for installation; su for in-place resistance) whereas classical
bearing capacity uses su for both installation and resistance, and it is easier
Installation, retrieval and removal 91

to match installation resistances after 30 or 90 days with in-place axial

resistance assessments.
Penetration is a “coring” mode (Figure 10.1). Hence, axial soil resistance
is the sum of end-bearing on the caisson tip, plus inner and outer skin fric-
tion. Other major assumptions (which are conventional) include (a) static
equilibrium and (b) “friction fatigue” (unit skin friction decrease with
increasing penetration) is not considered, and (c) unpainted steel (d) CPT
method assumes adhesion is the same in self-weight penetration and suction
assistance zones but classical bearing capacity makes a distinction.
The area over which underpressure/overpressure Δu acts is Ai, the internal
area (not the caisson base area A = π D2/4). For no water flow, at depths
below self-weight penetration, Δu is dependent on Wsteel, the foundation
submerged weight, and Δu values are computed assuming static force equi-
librium (load = resistance) from the simple equations:
u R Wsteel /Ai installation (9.3)

u R Wsteel / Ai exaction, i.e. retrieval and removal (9.4)

where
R = penetration (or extraction) resistance
Wsteel = caisson submerged weight (and preload, if any)
Ai = caisson inner area = π Di2/4
Di = caisson inner diameter

Fortunately, since the majority of soil profiles are clay, there is no water
flow, and the resistance versus depth profile is usually simple to calculate.
Entering the penetration resistance curve with Wsteel gives the self-weight
penetration. At greater depths, the suction pressure is computed directly
from the resistance curve using equilibrium Equation (9.3).
However, if either underpressure or overpressure create fluid flow, then
penetration resistance R is a function of Δu. An iterative technique is
required since Δu appears on both sides of both load–resistance equilibrium
equations.
Values of Δu need to be compared with limiting underpressures for
base failure (clay) and liquefaction (sand), details of which are given in
Sections 9.10 and 9.11. In addition, for shallow water depths, the maxi-
mum attainable underpressure (Section 9.7) also has to be checked.

9.8.2 CPT qc method
CPT cone resistance qc data are frequently used to predict skirt penetration
resistance. This is reasonable since the prototype skirt and CPT are of simi-
lar size (as is the case here). Foundation skirt tips are usually flat, whereas
92 Installation, retrieval and removal

the CPT tool is conical. Hence coefficient kp, relating unit end-bearing qw to
CPT qc is a shape factor, and has a value of less than 1.
Penetration resistances R are assessed based on the DNV (1992) method.
For a uniform wall thickness (WT) cylindrical foundation, R is calculated
using the equation

R A pk pq c,L Asi k f q c, zdz Aso k f q c, zdz (9.5)

where
L = foundation tip depth
kf = empirical coefficient relating qc (or qt) to unit skin friction resistance
kp = empirical coefficient relating qc (or qt) to unit end-bearing resistance
Ap = foundation wall tip area = π (D2 − Di2) /4
Asi = caisson inner perimeter = π Di
Aso = caisson outer perimeter = π D
D = foundation outer diameter
Di = foundation inner diameter = D − 2 WT
qc,z = CPT cone tip resistance at depth z below seafloor
qc,L = CPT cone tip resistance at caisson embedded/tip depth L

In Equation (9.1), inner and outer skin friction components have been
separated because of differing fluid flow effects. Both terms are integrated
from seafloor to depth L. The DNV (1992) method uses average CPT cone
resistance qc profiles, plus Best Estimate and High Estimate empirical coef-
ficients kp and kf relating qc to unit end-bearing and skin friction, to derive
foundation penetration resistance. DNV RP-H103 (2014) repeats the DNV
(1992) coefficients, but a complicating factor is that assessment is done on a
single CPT. The DNV (1992) averaging approach is preferred – CPTs in
sand can vary significantly over short distances – provided that sufficient
CPTs have been made.
For sand, DNV (1992) indicates that experience has shown that coeffi-
cients kp and kf for the upper 1 m to 1.5 m should be 25% to 50% lower
than those given in DNV (1992) where local “piping” or lateral platform
movement has occurred.

9.8.3 CPT method coefficients kp and kf

Based on Equation (9.1), “most probable” and “highest expected” penetra-
tion resistances can be calculated. Modern-day practice is gradually pro-
gressing towards probability-based assessments. Hence, Best Estimate and
High Estimate have generally replaced the original DNV (1992) “most
probable” and “highest expected” terms. In addition, the latter DNV term
misleadingly implied that the value will not be exceeded.
Installation, retrieval and removal 93

Table 9.2 Values of kp and kf Coefficients for Sand and Clay.

“Best Estimate” R BE “High Estimate” R HE

Soil type kp kf kp kf Remarks

Clay 0.4 0.03 0.6 0.05 DNV (1992) for
Sand 0.3 0.001 0.6 0.003 “North Sea Conditions”
Clay 0.4 (1) 0.006 0.6 (1) 0.018 Colliard and Wallerand
(2008) for GoG NC
clay, f = k f q T,CPT
Carbonate Most accurate fits from SWP back-analysis Frankenmolen et al.
Silt Unit end-bearing: either q w = 0.6 q T,CPT (2017) for Australian
or q w = 1.0 q T,T-bar North West Shelf
Unit skin friction: f i = 1 kPa (2) , f o = s ur,LE
≈ 1 kPa

Notation:
unit end-bearing resistance qw = kp qc
unit skin friction resistance f = kf qc or kf qT
qT,CPT = total cone resistance(3)
qT,T-bar= total T-bar resistance
sur,LE = low estimate soil remoulded undrained shear strength
Notes:
(1) from DNV (1992) – k values not stated by Colliard and Wallerand (2008)
p
(2) low internal f due to tip stiffener

T,CPT ≈ 1.25 qc for NC clay

(3) q

Table 9.2 gives kp and kf coefficient values for various soil types. For
North Sea conditions (presumably competent dense sands and stiff clays),
coefficients kp and kf were first proposed by DNV (1992). Lower kf coeffi-
cients were suggested by Colliard and Wallerand (2008) for Gulf of Guinea
clay (normally consolidated, sensitive, high plasticity). Since penetration in
normally consolidated clay is governed by skin friction, the Gulf of Guinea
clay kp values are considered to be the same as for North Sea conditions.
Germanischer Lloyd (2005, 2013) requires higher High Estimate coeffi-
cients than DNV (1992). More recently, Frankenmolen et al. (2017) back-
analysed the Prelude project caisson self-weight penetration data, and their
most accurate fit values (neither Best Estimates nor High Estimates) are also
given. Carbonate soils are notoriously difficult to characterise: penetration
resistances were only slightly higher than the Low Estimate but a factor ≈ 5
lower than the High Estimate. Note that the presence of any cemented seams
within the silt would have prevented suction installation. Finally, Table 9.2
shows that there is room for improvement – especially in sand, where end-
bearing is generally the larger resistance component, and the Best Estimate
and High Estimate kp values vary by a factor 2.
Table 9.3 compares North Sea, Gulf of Guinea and South China Sea clay
characteristics. This may be useful when considering extrapolating Colliard
and Wallerand (2008) kp and kf coefficients to other sensitive, normally
94 Installation, retrieval and removal

Table 9.3 Comparison North Sea, Gulf of Guinea and South China Sea Clays.

Gulf of Guinea South China

North Sea clays clays Sea clays
Lunne and St Puech et al. Palix et al.
Soil parameter Units John (1992) (1) (2005) (2013)
General Competent Weak Weak
over normally normally
consolidated consolidated consolidated
low-medium high plasticity high plasticity
plasticity sensitive sensitive
insensitive
qc [MPa] ≈2 ≈ 0.015 z to ≈ 0.017 z to
0.030 z 0.020 z
[q net gradient] [q net gradient]
qT [MPa] n/a n/a n/a
q T /q c [–] ≈ 1.1 ≈ 1.25 n/a
Rf [%] n/a ≈ 0.3 to 0.4 (2) ≈ 2 to 3
Bq [–] n/a ≈ 0.5 n/a
w [%] ≈ 15 to 25 ≈ 150 ≈ 150
wp [%] ≈ 15 to 25 ≈ 40 to 50 ≈ 40
wL [%] ≈ 40 to 65 ≈ 120 to 160 ≈ 80
I p (= w L − w p ) [%] ≈ 25 to 30 ≈ 70 to 120 ≈ 45 to 55
γ [kN/m 3 ] ≈ 20 ≈ 13 to 14.5 ≈ 13 to 14.5
su [kPa] ≈ 150 to 400 ≈ 1.5 z ≈ 1.3 z
[very stiff to
hard]
St [–] ≈ 1.2 to 1.5 ≈ 2 to 5 n/a (T-bar
unreliable, S t
≈ 8.5 at 25
cycles)
s u /σ’ vo [–] ≈ 4 to 10 ≈ 0.4 to 0.5 ≈ 0.35 to 0.3
(s u,DSS /σ’ vc )

Notation:
qc = cone resistance
qT = total cone resistance
Rf = friction ratio
Bq = pore pressure ratio
wp = plastic limit
wL = liquid limit
Ip = plasticity index
γ = unit weight
su = undrained shear strength
St = sensitivity
n/a = not available
z = depth below seafloor
Notes:
(1) Figures 2, 4 and 5 (CPT q ) and Table 1
c
(2) Probably a typographical error: R is typically 3% to 4% in (sensitive) clays
f
Installation, retrieval and removal 95

Figure 9.4 S elf-weight penetration back analyses in NC Gulf of Guinea clay: (a)
adhesion factor α and (b) k f coefficient (Colliard and Wallerand, 2008)
© IHS Markit.

consolidated clay locations (Figure 9.4). The two normally consolidated

clay sites are broadly similar, but the devil is in the detail: the open structure
of the South China Sea clay gives a higher sensitivity St. Low thixotropy and
rate effects were also noted. Unlike the Gulf of Guinea, these have possible
implications for soil cyclic shear strength degradation, and anchor capacity
increases with time. They also underline the necessity of obtaining appropri-
ate ground truth data for use in detailed design.
For soils intermediate between clay and sand (i.e. non-carbonate silt), if
uncertainty exists about drainage, then it is reasonable to take the higher of
clay and sand coefficients – i.e. use sand kp for end-bearing and clay kf for
skin friction. Similarly, kp and kf coefficients intermediate between the tabu-
lated values for clay and sand should be used for highly stratified soils,
interbedded clays and sands, or sand/clay mixtures.

9.8.4 CPT method coefficient αu

For analyses involving water flow in permeable soil (sand), values of foun-
dation tip pore pressure coefficient αu are required. These values depend on
foundation penetration, the soil plug : soil mass permeability ratio and soil
96 Installation, retrieval and removal

layering. Industry practice assumes steady state and isotropic permeability.

Both assumptions are considered very reasonable: steady state is checked at
the end of this section, and Dutch onshore practice (mainly excavation
dewatering) assumes kh = kv in their Pleistocene sands. The results presented
in this section need modification if the foundation is not axisymmetric.
Limiting αu values are 0.5 when penetration ratio L/D is almost zero and 0
when the foundation tip enters an impermeable layer.
As shown in Figure 9.5a, Houlsby and Byrne (2005) used finite element
analyses for foundations with penetration ratio L/D values up to 0.8 to
derive the following approximate equation for isotropic permeability:

u c0 c1 1 exp L / D / c 2 (9.6)

where
c0 = 0.45, c1 = 0.36 and c2 = 0.48

Sand plug heave (i.e. loosening) increases the plug permeability. Assuming
head loss is reduced in inverse proportion to k, and correcting a typographi-
cal error in the denominator, the corresponding Houlsby and Byrne equa-
tion is:
u 1k f / 1 1 1k f (9.7)

where
α1 = c0 − c1 [1 – exp(− (L/D) /c2)]
kf = sand plug : soil mass permeability ratio [–]

Results for kf = 2 and 5 are also given in Figure 9.5a. Using Equation
(9.7), extrapolated αu values are 0.14 and 0.32 for kf = 1 and 3 respectively
at L/D = 1.
Figure 9.6 shows four steady-state flownets obtained for a caisson foun-
dation with L/D = 1, showing equipotentials at intervals of 10% of the
change in pore pressure. A slightly modified version of the finite element
program P72 (Smith and Griffiths, 1998) was used, which solves Laplace’s
equation over an axisymmetric region. A 15-by-15 mesh of 4-noded rectan-
gular quadrilateral elements is used, with no element integration and assem-
bly for the cut-off (foundation wall) elements.
The first flownet, Figure 9.6a, is for an infinite half space of uniform per-
meability. Note that most of the pressure drop occurs within the soil plug
between tip level and seafloor. In addition, the hydraulic gradient is reason-
ably uniform both radially and axially within the soil plug but is a maxi-
mum at seafloor. At tip level, αu ≈ 0.14 (taken by eye between the equipotential
at 0.1 and 0.2), i.e. agrees with Equation (9.7) αu = 0.14.
Installation, retrieval and removal 97

(a)

(b)

Figure 9.5 I nstallation in sand with seepage flow: (a) foundation tip pore pressure
coefficient α u versus penetration ratio L/D and (b) dimensionless flow
parameter F versus L/D. Flow rate into sand plug Q = F (kD∆u/γ sub ).
Parameter k f = soil plug: soil mass permeability ratio. Use the bottom-
most curve (k f = 1) for uniform permeability (Houlsby and Byrne,
2005).
98 Installation, retrieval and removal

(a)

(b)

Figure 9.6 F lownets due to underpressure/overpressure. Axisymmetric steady

state fluid flow. Foundation penetration ratio L/D = 1.0, tip pore pres-
sure coefficient α u and dimensionless flow parameter F: (a) uniform
permeability k, infinite half space, α u ≈ 0.14, F ≈ 0.60. (b) plug k : mass
k = 3, infinite half space, α u ≈ 0.30, F ≈ 1.30.
Installation, retrieval and removal 99

(c)

(d)

Figure 9.6 C
ontinued: Flownets due to underpressure/overpressure. Axisymmetric
steady state fluid flow. Foundation penetration ratio L/D = 1.0, tip pore
pressure coefficient α u and dimensionless flow parameter F: (c) uniform
permeability k, impermeable boundary at 1.1 D bsf. αu ≈ 0.24, F ≈ 0.50. (d) con-
fined aquifer: as (c) but with impermeable suface. α u ≈ 0.32, F ≈ 0.42.
100 Installation, retrieval and removal

In the real world, the soil is usually of non-uniform permeability and the
far field boundaries are not at infinity, and sensitivity analyses of such rou-
tinely encountered cases are given on the remaining three components of
Figure 9.6. Figure 9.6b shows the flownet for the case when kf = 3 (was
kf = 1). Due to the reduced head loss in the sand plug, the tip pore pressure
coefficient αu increases, αu ≈ 0.30 (was αu ≈ 0.14). This means that the plug
seepage gradient reduces (for the same L/D and underpressure Δu). Again,
αu ≈ 0.30 agrees well with Equation (9.7) αu = 0.32.
Figure 9.6c explores the case when the foundation tip approaches an
impermeable boundary. It shows the flownet when the impermeable base is
no longer at infinity, but at 1.1D below seafloor (bsf) (i.e. 0.1D below tip
level). As expected, reduced radial transmissivity below the tip increases the
head drop in this zone, again leading to an αu increase (now αu ≈ 0.24, was
αu ≈ 0.14).
Finally, Figure 9.6d has the same boundary conditions as Figure 9.6c, plus
an impermeable top surface (outside the foundation). This case models a
confined sand aquifer overlain by clay, in which a “water pocket” (see
Section 9.12.2) may be created at the clay/sand plug interface. Since all
water flow comes from the radial far field boundary (the flownet becomes
rectangular), αu increases by a factor 1.3 (now αu ≈ 0.32, was αu ≈ 0.24 for
Figure 9.6c). Surprisingly, the decrease in dimensionless flow parameter F is
only 15% (now F ≈ 0.42, was F ≈ 0.50). This is for steady state flow. The
following commentary includes diffusion analyses of these last two cases
and shows that their excess pore pressure–time curves differ little.
These findings are also applicable to other L/D values. Hence, using infi-
nite half space solutions and uniform permeability for unconfined sand (i.e.
the first three Figure 9.6 situations) are cautious – this case under-predicts αu
and, as shown in the example that follows, also under-predicts soil plug
average seepage gradient iavg and underpressure ∆u. Since unit skin friction
and unit end-bearing decrease with increasing iavg, penetration resistance R
is over-predicted.
As an example, consider a foundation with embedment ratio L/D = 1, L =
10 m and sand γsub = 11 kN/m3 (i.e. icrit = 1.1). Using the infinite half space
solution and uniform permeability kf = 1 (i.e. the Figure 9.6a flownet), the
αu ≈ 0.15 value gives iavg = 0.85 (8.5 m head loss/10 m length) for an under-
pressure ∆u = 100 kPa. Since the hydraulic gradient iavg (0.85) does not
exceed icrit (1.1), ∆u can be increased. When iavg equals icrit, the critical under-
pressure ∆ucrit value is ≈ 130 kPa (100 kPa × 1.1/0.85). If the sand plug
permeability increases to, say, kf = 3, then the results given on Figure 9.6b
are applicable, i.e. αu ≈ 0.33 (was 0.15). The ∆ucrit value is now ≈ 165 kPa
(100 kPa × 1.1/0.67) instead of 130 kPa. This represents a 25% ∆ucrit
increase – provided that a factor 3 sand plug permeability increase can be
justified.
Regarding penetration resistance R, and defining penetration resistance
multiplication factor ηR = (1 − iavg/icrit), consider the preceding example with
Installation, retrieval and removal 101

underpressure ∆u = 100 kPa. Assuming kr = 1, we have iavg = 0.85, giving ηR

= (1 − 0.85/1.1) = 0.22. If in fact kr = 3, then αu ≈ 0.33 gives iavg = 0.67 (6.7
m head loss/10 m length) for the same underpressure ∆u = 100 kPa. This
gives ηR = (1 − 0.67/1.1) = 0.39. Most sand seepage models assume a linear
relationship between R and iavg, i.e. R = 0 when ηR = 0. If so, then, for the
previous example, R has been over-predicted by ≈ 75% (ηR = 0.22 for kr = 1,
but actually ηR = 0.39 if kr = 3).
Values of αu and associated underpressure Δu are generally used to assess
inner friction and tip resistance reduction, and, as shown earlier, iavg needs to
be less than icrit, the critical underpressure. In addition, available underpres-
sure Δu must exceed the underpressure required to overcome the penetra-
tion resistance R; see Section 9.11.2. Note that icrit is usually not iavg, but the
exit seepage gradient (i.e. at seafloor), more details of which are given in
Section 9.11.1. Penetration resistance R methods were summarised earlier
in Section 9.8.1. For water flow, the relevant equation is Δu = (R – Wsteel) /Ai,
where R is a function of Δu.

9.8.4.1 Commentary – steady state

Conventional design practice assumes steady state (not transient) conditions
to analyse the pore pressure conditions in the soil plug and at caisson tip
level. The following paragraphs assess how quickly steady state (equilib-
rium) is achieved and discuss the reasonableness of the steady state flow
assumption.
Consider 100 kPa suction pressure (total stress decrease Δσ) instanta-
neously applied to the seafloor inside a caisson. At time t = 0, underpressures
Δu created in the soil are equal to Δσ. However, Δσ varies throughout the
soil mass. A reasonable approximation is that Δu = 100 kPa throughout the
soil sand plug within the caisson, diminishes with depth below caisson tip
level using a 1:2 (h:v) load spread model (e.g. Δu = 16 kPa at 25 m bsf), and
Δu = 0 elsewhere. With time, the Δu values dissipate to seafloor (Δu = 100
kPa and 0 kPa inside and outside respectively). Linear elastic soil and lami-
nar flow are assumed. The transient flow 2D axisymmetric diffusion equa-
tion was solved using a slightly modified version of finite element program
P80 (Smith and Griffiths, 1998). The following parameter values were
adopted:

L caisson embedded length 10 m

D caisson diameter 10 m
cv soil coefficient of consolidation 5 m2/s

L and D are both reasonable/typical caisson in sand values (L/D = 1)

102 Installation, retrieval and removal

Soil consolidation coefficient cv was derived assuming:

k soil permeability 1e–3 m/s

qc CPT resistance 10 MPa
D soil confined modulus ≈ 5 qc (50e3 kPa)
γwater water unit weight 10 kN/m3
cv = k D/γwater = 1e–3 × 50e3/10 5 m2/s (≈ 1.5·10e8 m2/y).

Program P80 was used to re-analyse the four steady state cases shown on
Figure 9.6 (i.e. uniform, kr = 3, impermeable and aquifer). Figure 9.7 pres-
ents normalised caisson tip Δu pore pressure ratio at tip level (i.e. αu),
together with the corresponding t50 and t90 data points. The steady state (t =
200 s) αu results are slightly higher than the Figure 9.6 αu values. This is
probably due to program P72/Figure 9.6 averaging inner and outer nodal
Δu values, whereas program P80/Figure 9.7 used inner only.
Figure 9.7 shows that:

1. elapsed times vary between 0.03 s and 0.06 s (t50) and 3.5 s and 6.4 s
(t90) to achieve 50% and 90% of steady state Δu values
2. the fastest dissipation (lowest t90) is 3.5 s for case (b) plug (k or) cv:
mass (k or) cv ratio = 3
3. the slowest dissipation (highest t90) is 6.4 s for case (a) uniform
permeability
4. surprisingly, the Δu − t curves for cases (c) and (d) differ little.

Figure 9.7 C
aisson tip excess pore pressure – time. Axisymmetric unsteady state
fluid flow. Foundation L = D = 10 m. Soil mass c v = 5 m 2 /s for cases (a)
through (d). Soil plug c v = 15 m 2 /s for case (b).
Installation, retrieval and removal 103

In all four cases, steady state conditions are achieved rapidly (t90 within
6 seconds). This implies that (for the given geometry and parameter values)
the steady state assumption is very reasonable. Using steady state Δu values
is slightly cautious, since they are only marginally underpredicted. In addi-
tion, using case (a) is also cautious, since it has the “slowest” response.
To put this time (6 seconds) into perspective, assuming ≈ 1 mm/s penetra-
tion rate and 50 mm tip wall thickness WT, the caisson tip will have pene-
trated another 6 mm, i.e. just over 8% of WT.
Dissipation times (t90) are essentially proportional to 1/cv and L2. The
main source of error in cv (cv = k D/γwater) is the uncertainty in soil permeabil-
ity k (typically 2 or 3 orders of magnitude), followed by uncertainty in con-
strained modulus D (up to 1 order of magnitude). Note that, as the soil plug
seepage gradient increases, k increases (soil becomes more permeable), and
D decreases (soil becomes softer), which may possibly keep cv more or less
constant.

9.8.5 Classical bearing capacity method

Penetration resistance R is calculated using the equation:

R Qd Q f Q p fAs qA p (9.8)
where
Qf = skin friction resistance along sides of caisson and any protuberances
Qp= total end-bearing resistance at caisson tip and any protuberances
f = unit skin friction resistance
As = pile inside and outside surface areas embedded in soil
q = unit end-bearing resistance
Ap = caisson tip cross-sectional area (excluding contained soil).

For undrained soil response, inner and outer unit skin friction, f, at any
point along the foundation may be calculated by the equation
f s u (9.9)
where
α = a dimensionless factor (see Tables 9.3, 9.4).

Unit end-bearing q may be computed by the equation

q = N csu (9.10)
where
Nc = bearing capacity factor using Table 9.5. This is from ISO 19901-4
(ISO 2016), not API 2GEO (API 2011), because end-bearing needs to be
more carefully considered: the contribution is larger for penetrating caissons
than for frictional plugged pipe piles.
104 Installation, retrieval and removal

Table 9.4 Recommended Adhesion α Factors.

Purpose Soil/pile adhesion factor α [–] Remarks

Installation, pull-out and α = 0.90 on s u,r (defined by Best
retrieval (time t = 0) Estimate shear strength s u and
sensitivities S t )
Pull-out (time t = ∞) α = 0.70 (self-weight penetration (NGI, 1999)
zone)
α = 0.65 (suction assistance zone)
Removal (time t = ∞) α = 0.90 on s u,peak

Notes: DNV (2005) recommends slightly different (usually higher) α values, e.g. API RP2GEO (2011)
Main Text for pull-out – but they are for driven (not pushed) piles
Pull-out excludes twist considerations, for which α should be decreased slightly

Table 9.5 Recommended N c Factors (ISO 19901-4:2016).

Purpose Shape of Area Bearing Capacity Factor N c [–]

Calculation of pile tip Strip 7.5
penetration resistance
Calculation of critical Circular 6.2 to 9.0 depending on
underpressure causing embedment ratio L/D
soil plug failure N c ≈ min[6.2 (1 + 0.2 L/D), 9]
Calculation of penetration Varies 5 to 13.5
resistance of
protuberances

For frictional soils, unit skin friction, f, may be calculated by the

equation:
f bvo (9.11)
where
β = dimensionless skin friction factor
σ'vo = effective overburden pressure at the depth in question.

Since coring intermediate foundations are assumed similar to driven

open-ended pipe piles unplugged, Table 9.6 may be used for selection of β
values if other data are not available. Unit end-bearing on the pile wall tip
(and any protuberances) qw may be computed using the equation:
q w N qvo, tip (9.12)
where
σ'vo,tip = soil effective in-situ vertical stress at foundation tip level
Nq = dimensionless bearing capacity factor (see Table 9.6).
Installation, retrieval and removal 105

Table 9.6 P
enetration Resistance in Silica Sand. Skin Friction and End-Bearing
Parameters (ISO 19902:2007).

Skin Limit End- Limit unit

friction unit skin bearing end-
Relative factor friction factor bearing
density D r [%] Soil type β [–] f lim [kPa] N q [–] q lim [MPa]
Very loose 0–15 Sand
Loose 15–35 Sand
Loose 15–35 Sand-silt n/a n/a n/a n/a
Medium 35–65 Silt
dense 65–85 Silt
Dense
Medium 35–65 Sand-silt 0.29 67 12 3
dense
Medium 35–65 Sand 0.37 81 20 5
dense
Dense 65–85 Sand-silt
Dense 65–85 Sand 0.46 96 40 10
Very dense > 85 Sand-silt
Very dense > 85 Sand 0.56 115 50 12

Notes:
n/a : not applicable
Unit skin friction f = min(β σ'vo, flim)
Unit wall end-bearing qw = min(Nq σ’vo, qlim)

Intermediate suction foundations are usually not long enough to warrant

limiting f and qw to the limit values given in Table 9.6. In addition, the tabu-
lated Nq values are for circles, not strips. Hence they over predict coring
penetration resistance.

9.9 LANDING ON SEAFLOOR, MINIMUM SELF-WEIGHT

PENETRATION AND FREE-FALL PENETRATION

9.9.1 Landing on seafloor
When landing on a sandy seabed, local scouring at the edge of the founda-
tion and piping around the skirts are almost inevitable unless the lowering
velocity is very low and the areas for water evacuation through the founda-
tion base are large.
Section 7.3.7.6 of DNVGL-RP-C212 (DNV-GL, 2017, 2019) discusses
piping when landing on sandy seabeds:

Open piping channels towards the end of the penetration shall be

avoided in particular if final penetration by suction is necessary. The
process of stopping the piping is complex and is not possible to analyse
with a reasonable accuracy. It involves initially scouring and transport
106 Installation, retrieval and removal

of particles horizontally on the seabed. As the penetration increases

the transportation of the particles involves vertical lifting that requires
higher velocities. Piping is likely to take place at local positions along
the periphery corresponding to the lower spots of the seabed, whereas
along the remaining parts of the periphery a penetration resistance will
gradually build up and cause the piping to stop at some stage. There
appears to be no available reliable method to analyse this process, so
empirical data should be sought after and used for assessment of the
potential for scouring and piping. Surveys after installation are impor-
tant as a basis for deciding whether mitigations become necessary.

In addition, Section 6.2 of DNV-RP-H103 (DNV, 2014b) gives pertinent

advice for analysis and modelling impact landing.
For controlled lowering, a reasonable foundation velocity just above sea-
floor (excluding vessel motions) is 0.2 m/s. This is because most design bases
for shallow and intermediate foundations specify that the velocity should not
exceed 0.2 m/s. This value also is implied in both ISO-19901-4 (ISO, 2016a)
(Annex A.7.6.2.1) and API 2GEO (API, 2011) - Sections 7.14 and A.7.14
discusses installation effects. Both documents are for shallow (not intermedi-
ate) foundations. This (maximum) velocity of 0.2 m/s is a starting point for:

• assessing foundation drainage (shallow foundations)

• designing relief valves (suction caissons) – piping/soil plug base failure
(push-out).

Velocities higher than 0.2 m/s occur during uncontrolled self-weight pene-
tration (both caissons and open-ended pipe piles), and their free-fall pene-
tration is discussed in Section 9.9.3.
Top plate vent design and suction pump design are discussed in Sections
12.6 and 12.7 respectively.

9.9.2 Minimum self-weight penetration

Self-weight penetration should generally be at least 0.5 m below seafloor all
the way around the foundation tip perimeter. This is in order to:

• preclude local piping around the foundation tip

• start the suction assisted installation stage.

Occasionally, when competent soils (e.g. dense sand and very stiff clay) are
present at seafloor, this penetration value may not be achieved.
For a given foundation geometry, possible measures to increase self-
weight penetration are

• decrease foundation D and/or WT (i.e. reduce tip resistance)

• increase preload (e.g. tank or steel pipe pile filled with water, rock,
concrete)
Installation, retrieval and removal 107

• make sure seafloor is as flat as possible (i.e. minimise risk of hydraulic

leaks)
• add loose sand (before installation, perimeter footprint)
• bevel tip (i.e. attempt to reduce tip resistance)
• add sandbags around outer perimeter (after installation, dubious
effectiveness)
• relocate intermediate foundation
• jetting (difficult to control).

Friction breakers and paint are unlikely to be effective in sand. This is

because most of the self-weight penetration resistance is from end-bearing
(not skin friction) resistance.

9.9.3 Free-fall penetration
Free-fall penetration occurs without crane assistance and a heavy founda-
tion, and can also occur with a decreasing soil resistance (“hard over soft”)
profile. Since penetration is uncontrolled, risks need to be minimised. High
free-fall penetration values are associated with heavy intermediate and pile
foundations in clay. There is no free-fall risk during the subsequent suction
assistance stage. This is because suction pumps operate at a semi-fixed flow
rate; hence penetration is displacement, not load, controlled.
Uncontrolled penetration generally exceeds controlled self-weight pene-
tration – this effect is used to install torpedo piles well below seafloor. Free-
fall is best modelled using Newton’s Second Law, forward Euler integration
with time and Archimedes’ Principle for buoyancy effects. Assuming “cor-
ing” in clay and ignoring viscous drag effects on the wall tip, the corre-
sponding pseudo code to model free-fall in soil is:

program free_fall
initialise pile velocity and tip penetration depth;
subscripts 0 and 1 refer to previous and current timesteps
v1 = max(v0, 1e−3 m/s)
z1 = 0 m/s
∆t = 0.01 s
loop while pile velocity v1 > 0:
v0 = v1
z1 = z1 + v0∆t
SF: s
kin friction resistance by integrating su from 0 to
z1
EB: end-bearing resistance using su at z1
SF1 = ∫ Perim α su,z dz
EB1 = Awall su,z1 Nc,strip,deep
RF: vent (nozzle) resistance using v0
108 Installation, retrieval and removal

VN : vent (nozzle) velocity

∆H : vent differential head loss
∆u : overpressure increase
vpile = v0
vn = vpile Ai/An
∆H = [1/(2 g)] [vn2/Cd2]
∆u = ∆H γwater
RF1 = ∆u (Ai − An)
BF: buoyancy force at z1, assuming constant wall area
BF1 = Vsteel [(L − z1) /L] γwater + Vsteel [z1/L] γsoil
F: pile downwards force at z1
a, v: pile acceleration and velocity at z1
F1 = M g − (SF1 + EB1 + RF1 + BF1)
a1 = F1/M
v1 = v0 + a1∆t
end loop
exit with pile penetration z1

The above pseudo code is easily modified to include free-fall in the water
column above seafloor, and viscous effects. Viscosity is velocity dependent,
and various torpedo pile researchers have proposed dimensionless drag
coefficients, both above and below seafloor velocity dependent and shaft/tip
viscosity models (e.g. O’Loughlin et al., 2004). For a given vent area, the
vent resistance Rn is a function of the nozzle velocity and inverse discharge
coefficient squared, i.e. Rn α (vn/Cd)2. Section 12.6 (on top plate vent design)
derives the Rn equation and also gives a worked example.
Both the free-fall model and 1-D drivability (wave equation using lumped
masses and springs) analyse pile dynamics using integration with time. It is
noted that pile buoyancy effects are missing from the original Smith (1960)
drivability model, but it is possible in GRLWEAP (Pile Dynamics, 2010) to
include pile buoyancy effects by changing the pile gravity. Office drivability
practice generally excludes buoyancy – all piles are submerged, whether on
land or offshore. Including buoyancy would make piles slightly harder to
drive; this is because the additional upwards resistance increases soil resis-
tance to driving (SRD).

9.9.3.1 Example – pile free-fall

To illustrate the difference a vent makes to penetration, two free-fall
cases are analysed in a normally consolidated clay profile. Both have the
same foundation geometry (D = 5 m, L/D = 4) and mass (67 tonne), and
have the same seafloor velocity (v0 = 0.2 m/s). Case A is a suction caisson
with two number 1.0 m diameter vents, open during free-fall. Case B is
an OWT monopile with an open top, i.e. without a top plate. Due to vent
Installation, retrieval and removal 109

resistance, it is expected that Case A (suction pile) will penetrate less

than Case B (OWT monopile).
The data, which are subsequently re-used in the Section 12.6 (Top Plate
Vent design) example, for both cases are:

Pile: D = 5 m, L = 20 m, WTside/D = 200, WTtop/D = 100, ρsteel = 7.7 tonne/m3

(Di = 4.95 m, Perim = π (D + Di) = 31.259 m, Awall = 0.391 m2, Ai
= 19.244 m2)
(Vsteel = 8.796 m3, mass M = 67.733 tonne, weight M g = 664.231 kN)
Case A (suction foundation): two vents, diameter = 1.0 m
An = 0.8 m2
Case B (OWT monopile): no vent, vent diameter = 4.95 m
An = 19.244 m2

Soil: su = 2 + 1.5 z [kPa, m], ρsoil = 1.6 tonne/m3, ρwater = 1.0 tonne/m3
Pile – Soil – Fluid:

αi = αo = 0.3, Nc,strip,deep = 7.5

Cd = 0.61, v0 = 0.2 m/s.

Figure 9.8 compares the resulting Case A (coloured red/thick) and Case B
(blue/thin) acceleration and velocity data versus depth. The following para-
graphs discuss the free-fall and self-weight penetration results.

Figure 9.8 F ree-fall penetration example: (a) foundation acceleration – depth and
(b) velocity – depth. Foundation L = 20 m, D = 5 m, M = 67.7 tonne.
Seafloor velocity = 0.2 m/s. Free-fall penetrations: 7.41 m (suction
foundation with top plate vent) and 13.12 m (open top OWT mono-
pile). Self-weight penetration = 7.37 m. Suction foundation accelera-
tions have been increased by a factor 100.
110 Installation, retrieval and removal

9.9.3.2 Example – pile free-fall case A (suction foundation)

When released just above seafloor with velocity v0 = 0.2 m/s, the suction
foundation penetrates to 7.41 m bsf. The red/thick line on Figure 9.8a
shows the free-fall acceleration – depth data (1524 time steps, 15.2 s). For
clarity, accelerations have been multiplied by 100. It is seen that, just
below seafloor, there is an initial acceleration of around 0.02 m/s2. This is
caused by the downwards acting foundation weight (664 kN) exceeding
the upwards acting buoyancy force (86 kN) and vent resistance (143 kN
at vpile = 0.2 m/s), and essentially zero skin friction and end-bearing resis-
tance close to seafloor. Below around 0.2 m depth, acceleration magni-
tudes remain small, reaching around −0.02 m/s2 at final penetration.
There is no acceleration, only a deceleration whose magnitude increases
essentially linearly with depth. This linear decrease is due to both skin-
friction and end-bearing resistance increasing almost linearly with depth,
whereas buoyancy and vent resistances remain almost constant. The red/
thick line on Figure 9.8b shows the corresponding free-fall velocity –
depth data. The maximum velocity is just over 0.4 m/s (around twice v0)
near seafloor and the velocity decreases non-linearly until it eventually
becomes zero at 7.41 m depth. This penetration depth is unlikely to be
radically different had viscous effects been included – velocities are low.
Similarly, penetration is also insensitive to seafloor velocity: the same
value (7.41 m) is assessed for v0 = 2.0 m/s. This is due to the vents slowing
the foundation down.

9.9.3.3 Example – pile free-fall case B (OWT monopile)

For uncontrolled lowering (i.e. free-fall), released at seafloor (again with veloc-
ity v0 = 0.2 m/s), the intermediate foundation will penetrate to 13.1 m bsf.
The blue/thin line on Figure 9.8a shows the acceleration – depth data
(239 time steps, 2.4 s). It is seen that, unlike the suction foundation, accel-
eration is significantly higher – almost 1 g at seafloor. This is due to zero vent
resistance. Acceleration reduces with penetration and becomes zero at around
7 m bsf, and thereafter begins to decelerate, reaching around −15 m/s2 (just
under 2 g) at final penetration.
As expected, due to the higher accelerations, the blue/thin line on
Figure 9.8b shows that velocities are also higher, reaching a peak velocity of
approximately 8.7 m/s. Theoretically, the 20 m long pile would be fully
embedded if v0 is around 20 m/s. However, because viscous effects have been
excluded, this depth is unlikely to be achieved in the field.

9.9.3.4 Example – self-weight penetration Cases A and B

Since the foundation mass is constant, self-weight penetrations are identical
for both cases. For controlled lowering (i.e. using a crane or similar), we
Installation, retrieval and removal 111

have that skin friction (SF) + end-bearing resistance (EB) = pile buoyant
weight. This can be written in pseudo-code notation as:

SF + EB + BF = M g.

The three terms on the left-hand side – SF, EB and BF (buoyancy force) –
are all penetration depth dependent, and BF decreases slightly with increas-
ing penetration into a higher-density material than water. By trial and error
it is found that the SWP is 7.37 m. This value can be checked by verifying
that the forces balance. At 7.37 m depth, we have SF = 520.6 kN, EB = 38.3
kN, BF = 105.3 kN, giving SF + EB + BF = 664.2 kN. As expected, this bal-
ances the 67.733 tonne pile, whose weight M g = 664.2 kN.

9.9.3.5 Example – commentary
For this foundation, the Case B (OWT monopile) free-fall value (13.1 m) is
just over 175% of the self-weight penetration (7.51 m). Incorporating vis-
cous drag effects are likely to reduce the Case B free-fall value. The Case A
(suction foundation) free-fall value (7.41 m) is only 40 mm deeper than the
self-weight penetration (7.37 m).

9.10 INSTALLATION IN CLAY

The key design challenge for installation analyses in clay is usually calcula-
tion of installation and allowable underpressure. The effect on allowable
underpressure when encountering firm to stiff clay layers overlying sand and
weaker clay layers can also be a challenge.
The allowable underpressure may be limited by maximum suction pump
pressure (usually atmospheric or possibly vacuum, see Section 12.7) and
structural considerations (steel cylinder buckling, etc.).

9.10.1 Base failure in clay

If the top plate vent is not properly designed, then excess water pressure
build-up (overpressure) can occur during self-weight penetration. If signifi-
cant, then conventional end-bearing failure at tip level is possible. Section
12.6 gives some top vent design considerations. Assessing maximum allow-
able overpressure is similar to that for underpressure, which is described in
the remaining paragraphs.
Conversely, if the underpressure is too high, then geotechnical failure
(reverse end-bearing) of the soil plug at foundation tip level can occur and
should be checked. Various assessment methods are possible.
112 Installation, retrieval and removal

Maximum allowable underpressure (Δuall) profiles, derived using static

equilibrium of the clay soil plug, can be given by:

uall Fi / Ai N c,circle su,av, tip / m (9.13)

where
Fi = inner skin friction resistance (= α su,av Aint,surface)
Ai = plan view inside area where underpressure is applied
Nc,circle = circular foundation bearing capacity factor ≈ min[6 (1 + 0.2
L/D), 9]
su,av,tip = average su,z between caisson tip (L) and depth (L + D αD,su)
αD,su = su averaging parameter (αD,su ≈ 0.25)
γm = material factor on end-bearing = 1.5

Suction anchor embedment ratios L/D are usually 6 or less for normally
consolidated clays (e.g. Andersen et al., 2005). Equation (9.13), which is
sensitive to inner friction adhesion parameter αi, usually provides L/D values
well in excess of 6.
ISO 19901-4 (ISO 2016, equation A.77) also uses static equilibrium but
recommends a minimum safety factor of 1.5 on Δuall, i.e. on both REB and
Fi. Consequently, ISO 19901-4 (ISO 2016) gives lower Δuall values (and
lower limit embedment ratios) than DNV RP-E303 (DNV, 2018).
Note that using either the Skempton (1951) or the DNV (2005) equation
Nc,circle is likely to be cautious (i.e. they underestimate Δuall) and provide an
additional reserve on material factor γm. Section 10.5.2 gives reasons for
Nc,circle underestimation.
Randolph and Gourvenec (2011) considered the soil plug stability ratio
(the ratio of underpressure to cause plug base failure to the underpressure
required to penetrate the caisson to that depth). They present design curves
of plug stability ratio and embedment ratio, along with an equation for lim-
iting embedment ratio in normally consolidated soil conditions, as shown in
Figure 9.9.
Section 9.10.3 gives an installation example for clay using the DNV
RP-E303 (2005) equation. The pseudo code may be easily modified to
include other base failure options.
At the other end of the scale, low embedment ratios (L/D ≈ 0.5) have not
yet been investigated in detail. This is a typical potential problem in compe-
tent clays, where installation (rather than in-place capacity) governs.
Conventional reverse end-bearing (REB) theory using Nc,circle = 6 may not be
sufficiently cautious. Numerical 3D FEA may be required to verify that the
REB failure model is indeed applicable. The problem may be in fact non-
axisymmetric. A possibility is an essentially circular/spheroidal failure arc,
beginning inside the caisson and subjected to underpressure (i.e. negative
Installation, retrieval and removal 113

Figure 9.9 D
esign curves for suction foundation installation in clay (Randolph
and Gourvenec, 2011).

surcharge), intersecting the caisson at two points on the caisson tip, and
exiting at seafloor outside. Research may be required.

9.10.2 Plug heave in clay

Intermediate foundation penetration causes heave of the inner clay soil
plug at seafloor to match part of the soil volume displaced by the founda-
tion wall and the full volume of any internal stiffeners. Conventional prac-
tice is to assume a 50%–50% split during self-weight penetration but up to
a 100%–0% split in favour of inward flow during suction assisted penetra-
tion (DNV, 2018).
Additional heave should be allowed for if significant free-standing soil is
expected (e.g. above ring stiffeners), or underpressures are higher than Δuall
(maximum allowable underpressure).
114 Installation, retrieval and removal

The former is because a mixture of water and clay can be trapped between
internal ring stiffeners, and that trapped water can give a significant contri-
bution to soil plug heave (Andersen et al., 2005).
The latter is because the amount of soil entering the intermediate founda-
tion is likely to be strongly affected by the proximity to internal base failure.
During the early stages of suction assisted penetration, the factor of safety
against base failure is high, and soil flow around the foundation tip is con-
sidered to be similar to that during the self-weight penetration stage.
However, near the end of penetration, the soil plug is closer to failure requir-
ing mobilisation of reverse end-bearing resistance for stability, and hence a
greater proportion of soil met at the foundation tip is likely to flow inside.
Since installation will fail if the soil plug meets the top plate before the
intermediate foundation has penetrated to target depth, the foundation
overall height should be increased by at least the height of the plug heave.
The additional height is equal to (part of) the foundation steel volume
divided by its footprint area, plus the full volume of any internal stiffeners.
For a 50%–50% split during both self-weight penetration and suction-
assisted penetration stages, and no internal stiffeners, the additional height
is given by ∆Hheave/L ≈ 2 WT/D (Romp, 2013). Hence, for wall thickness/
diameter ratios between 1/300 and 1/100, height increases could be up to
2% of the embedded length. For large intermediate foundations (D = 5 m, L
= 30 m), ∆Hheave could be around 0.75 m. More soil plug heave can be
accommodated by domed (than flat) top plates.

9.10.3 Clay installation/retrieval example

To provide example resistance/underpressure–penetration depth diagrams,
the classical bearing capacity method for clay (f = αsu and q = Ncsu) has been
programmed to provide a simple clay installation (and retrieval/removal)
model. It can be modified/improved to accommodate non-uniform founda-
tion geometry, multiple clay layers and different opinions.
Assumptions include:

uniform caisson diameter D and wall thickness WT (i.e. no protuberances)

reverse wall tip end-bearing (i.e. same resistance during installation and
retrieval)
DNV RP-E303 (2005) base failure equation (γm = 1.5) during suction-
assisted installation.

Using these assumptions, the equations to find penetration resistances

Rsuction, Rretrieval and corresponding underpressure ∆usuction, retrieval overpres-
sure ∆uretrieval and allowable underpressure ∆uall at any tip penetration L are:

R suction Fi Fo Q w (9.14)
Installation, retrieval and removal 115

R retrieval = R suction (9.15)

usuction R suction Wsub,steel / A plug (9.16)

uretrieval Rretrieval Wsub,steel / A plug (9.17)

uall,suction Fi / A plug N c,circle su,av, tip / m,install (9.13, 9.18)

uall, retrieval Fi / A plug N c,circle su,av, tip / m, retrieval (9.19)

where
Fi Asi i su, zdz (9.20)

Fo Asoo su, zq c, zdz (9.21)

Q w = A wall N c,strip,deep su,av, tip (9.22)

and
αi = clay-steel inner adhesion factor (unit friction fi = αisu)
αo = clay-steel outer adhesion factor (unit friction fo = αosu)
Aplug = caisson internal (suction) and clay plug area
= π Di2/4
Asi = caisson inner perimeter
Aso = caisson outer perimeter
Awall = caisson wall tip area
L = caisson tip penetration into clay
Lcan = caisson total length
Nc,strip = strip foundation bearing capacity factor (clay)
≈ min[5 (1 + 0.2 L/D), 7.5]
Nc,strip,deep = deep strip foundation bearing capacity factor (clay) ≈ 7.5
R = soil resistance
su,av,tip = average su,z between caisson tip (L) and depth (L + Di αD,su)
(note: Di, not D)
αD,su = su averaging parameter (αD,su ≈ 0.25)
Wsub,steel = caisson submerged weight (including ballast)

The corresponding pseudo code for the above installation in clay is:

program install_clay
initialise and evaluate constants:
2 WT, Asi = π Di, Aso = π D, Ap = π (D2 - Di2) /4 and
Di = –
Aplug = π Di2/4
116 Installation, retrieval and removal

∆L = 0.1 m
nL = Lcan/∆L + 1
zero matrix ANS (size nL rows by 12 columns)
k = 0

loop tip penetration L from 0 to Lcan in ∆L steps:

k = k + 1
L = (k – 1) ∆L
Fi: inner skin friction resistance by integrating su
from 0 to L
Fo: outer skin friction resistance by integrating su
from 0 to L
Fi = Asi αi ∫ su,z dz and Fo,clay = Aso αo ∫ su,z dz
Qw: end-bearing resistance using su at depth L
Qw = Awall Nc,strip,deep su,L
Rsuction = Fi + Fo + Qw
Rretrieval = Rsuction
∆usuction = (Rsuction − Wsub,steel)/Aplug
∆uretrieval = (Rretrieval + Wsub,steel)/Aplug
su,av,tip = average su,z between caisson tip (L) and
depth (L + Di αD,su)
∆uall,install = Fi/Aplug + Nc,circle su,av,tip/γm,install
∆uall,rerievall = Fi/Aplug + Nc,circle su,av,tip/γm,retrieval
store in kth row of matrix ANS:
k, L, and L/Lcan in columns 1 through 3
Fi, Fo, Qw, Rsuction and Rretrieval in columns 4 through 8
∆uall,install and ∆usuction in columns 9 and 10
∆uall,retrieval and ∆uretrieval in columns 11 and 12

end loop on tip penetration L

exit with matrix ANS

matrix ANS (nL rows, 12 columns) contents:

col# 01: k matrix ANS row number [–]

col# 02: L caisson tip penetration into sand [m]
col# 03: L / D caisson penetration/diameter ratio [–]
col# 04: F i caisson inner wall friction resistance [kN]
col# 05: F o caisson outer wall friction resistance [kN]
col# 06: Q w caisson wall tip resistance [kN]
col# 07: R suction soil resistance during penetration [MN]
col# 08: R retrieval soil resistance during retrieval [MN]
col# 09: ∆u all,install allowable water (under) pressure [kPa]
col# 10: ∆u suction water (under) pressure during penetration [kPa]
col# 11: ∆u all,retrieval allowable water (over) pressure [kPa]
col# 12: ∆u retrieval water (over) pressure during retrieval [kPa]
Installation, retrieval and removal 117

Installation plots:
col# 07 (Rsuction) versus col# 02 (L) installation resistance – depth
col# 10 (∆usuction) and col# 09 (∆uall,install) versus col# 02 (L) underpressure
– depth

Retrieval plots:

col# 08 (Rretrieval) versus col# 02 (L) retrieval resistance – depth

col# 12 (∆uretrieval) and col# 11 (∆uall,retrievall) versus col# 02 (L) overpressure
– depth

9.10.3.1 Example – clay installation and retrieval

Consider a typical lightly overconsolidated clay (su = 5 + 2 z) in which a 4 m
diameter D anchor pile caisson has to be installed to 17 m below seafloor
(bsf) using self-weight and suction assistance. Check for base failure during
suction assisted phase using γm = 1.5 on plug base heave failure. Also assess
(immediate, time t = 0) retrieval, but with a slightly lower γm value (say 1.25)
to assess overpressure

Pile: D = 4 m, Lcan = 17 m, WT = D/200 = 20 mm, Wsub,steel

= 750 kN
(Di = 3.96 m, As,i = π Di = 12.44 m, As,o = π D =
12.57 m, Awall = 0.25 m2, Aplug = 12.32 m2)
Soil: clay, su,z = 5 + 2 z [kPa, m]
Pile – Soil:
αi = αo = 0.30
αD,su = 0.25
γm,install = 1.5 and γm,retrieval = 1.25.

Using the data just presented, at the end of penetration (L = 17 m), we have

k = 171
L = 17 m
L/D = 4.25

and

su,av,L = average su from 0 to L (17 m) = su at L/2 (8.5 m) = 5 kPa +

2 kPa/m × 8.5 m = 22 kPa
∫ su,z dz = average su from 0 to L (17 m) × L = 22 kPa × 17 m = 374 kN/m
su,tip = su at L (17 m) = 5 kPa + 2 kPa/m × 17 m = 39 kPa
su,av,tip = average su from L to L + 0.25 Di (17 m to 17.99 m) = 40 kPa

and
118 Installation, retrieval and removal

Fi = Asi αi ∫ su,z dz = 12.44 m × 0.30 × 374 kN/m = 1396 kN

Fo = Aso αo ∫ su,z dz = 12.57 m × 0.30 × 374 kN/m = 1410 kN
Qw = Awall Nc,strip,deep su,tip = 0.25 m2 × 7.5 × 39 kPa = 73.1 kN
Rsuction = Fi + Fo + Qw = 1396 kN + 1410 kN + 73.1 kN = 2.88 MN
Rretrieval = Rsuction = 2.88 MN
∆usuction = (Rsuction − Wsub,steel) /Aplug = (2.88 MN − 0.75 MN) /12.32 m2 =
173 kPa
∆uretrieval = (Rretrieval + Wsub,steel) /Aplug = (2.88 MN + 0.75 MN) /12.32 m2 =
295 kPa
Δuall,install = Fi/Aplug + Nc,circle su,av,tip/γm,install
= 1396 kN/12.32 m2 + 9.0 × 40 kPa /1.5 = 113 kPa + 240 kPa =
353 kPa
Δuall,retrieval = Fi/Aplug + Nc,circle su,av,tip/γm,retrieval
= 1396 kN/12.32 m2 + 9.0 × 40 kPa /1.25 = 113 kPa + 288 kPa
= 401 kPa

Figure 9.10 shows two resistance and pressure profiles. Figure 9.10a is for
penetration. It is seen that, at final tip penetration (17 m bsf), the required
underpressure ∆usuction (blue/thin line) is ≈ 175 kPa. The corresponding
allowable suction pressure Δuall (red/thick line) is just over 350 kPa. Since
Δuall exceeds ∆usuction, there is no major risk of soil plug base failure during
suction assisted penetration. The self-weight penetration value is 7.6 m, and
is more than sufficient to start the suction assisted stage. The high self-weight
penetration (nearly half the embedded pile length) is not uncommon for
anchor piles in weak clay, and self-weight penetration would decrease for αi
and αo values in excess of 0.3.
Figure 9.10b shows the corresponding retrieval profiles. These were
obtained using the same αi = αo = 0.3 value and the same soil su profile, but
with a γm = 1.25 (was 1.5). Because there is (reverse) end-bearing, the
retrieval and penetration resistance profiles are identical. Unlike penetra-
tion, retrieval resistance (and overpressure) drops once the pile starts to
move upwards. Hence the overpressure profiles (both required and allow-
able) are valid only at/near the penetration depth at which retrieval com-
mences. Since pile submerged weight opposes retrieval (instead of assisting
penetration), the overpressure magnitude is higher than the corresponding
underpressure: at 17 m bsf, ∆uretrieval is now 295 kPa (compared to ∆usuction
= 173 kPa). Despite the higher pressure, there is still no base failure risk at
all depths, except apparently above 3 m bsf. This is unlikely to occur – this
zone is well above the SWP depth (7.6 m bsf), and a vessel crane is used to
keep the suction anchor vertical and lift the it from the seafloor.

9.10.3.2 Commentary
Interested readers are encouraged to perform sensitivity analyses. For a
given su profile, key input variables include αi, αo and Wsub,steel. The model
assumes a uniform pile geometry. In practice, anchor piles are rarely
Installation, retrieval and removal 119

Figure 9.10 I nstallation and retrieval resistance in clay. Example using f = αs u method.
Anchor pile D = 4 m, L = 17 m in LOC clay (s u = 5 + 2z [kPa, m]).

uniform: internal ring stiffeners (usually an increased WT at/near lug level)

are used to transfer chain loads, and inner protuberances are used to reduce
the cylinder WT and steel stresses. In such cases, αi (and WT) is no longer
constant and has to be depth dependent. The pseudocode is capable of han-
dling such modifications.
At the time of writing there is no (ISO, DNV-GL) advice for selecting a
suitable γm value for overpressure assisted retrieval. The example presented
here used a lower γm value (1.25) for retrieval than installation, i.e. 1.5
according to DNV RP-E303 (2005). This lower γm,retrieval value implies that
the probability of plug base failure is lower during retrieval because of dif-
ferent design parameters. The main reason is usually adopting higher (HE or
120 Installation, retrieval and removal

BE) α values for retrieval and lower (LE or BE) α values for installation.
Another reason may be that, during the preliminary design phase, when
both installation and retrieval assessments are made, installation data are
available from previous projects nearby, thereby reducing the uncertainty
bandwidth. Other organisations may have other γm,retrieval opinions. In any
case, γm,retrieval values of less than 1.5 should be justified.
Since the tip resistance contribution is small, and unit friction f = αsu, it is
obvious that various α and su combinations can be used to compute the
same f (and hence assess resistances and pressures). In the example, su has
been kept constant.
Even though there is limited risk of plug base failure during suction-
assisted penetration, base failure could be possible during the preceding self-
weight penetration stage if high overpressures occur in the “water plug” due
to undersized vents. Section 12.6 (Top Plate Vent Design) below includes an
example calculation.

9.10.4 Friction set-up in clay

Set-up is the increase of skin friction resistance with time in normally con-
solidated and lightly overconsolidated clays. Friction increases with time
after installation due to thixotropy, dissipation of excess pore water pres-
sures and soil consolidation. Due to their smaller displacement ratio and
installation by suction, pore pressure dissipation is expected to occur more
quickly for suction foundations than for driven pipe piles.
Friction resistance component is usually large compared to total end-
bearing resistance. An “alpha” method (i.e. maximum friction is a propor-
tion of the undrained shear strength) is employed, with different alpha
values for the self-weight and suction assisted penetration zones. The shear
strength reduction factor (α) at time t = 0 is generally close to the reciprocal
of the soil sensitivity (St) value.
More information can be found in the following references:

ISO 19901-4:2016 Section A.8.1.3.2.5: changes in axial capacity in clay

with time (piles)
ISO 19901-4:2016 Section A.11.5.2.2.4: increase of side friction with time
(suction anchors)
DNV RP-E303 (2005) Sections 4.7 and 4.8: give outside and inside set-up
factors for suction anchors. Since factor values are low estimates, they
should be increased for retrieval and removal assessments.

For suction foundations in clay, two resistance estimates are usually pro-
vided at various time periods after installation:

• a Low Estimate for pull-out axial resistance

• a High Estimate for retrieval/removal.
Installation, retrieval and removal 121

Such friction set-up is for normally consolidated clays. Set-up magnitude

will be less (and could possibly decrease) for over consolidated clays. This is
due to the different pore water pressure regimes during and after
installation.

9.10.5 Boulders in clay
Foundation tip integrity (buckling initiation) may be adversely affected by
large-sized material. Such materials include boulders in glacial till, drop-
stones in weak clays, coral rubble in carbonate clay, and flintstones in chalk.
Like sand and gravel, boulder angularity can vary from angular (sharp edges
with relatively plane sides with unpolished surfaces) to rounded (smoothly
curved sides and no edges); see Figure 9.11.
This section outlines a reasonable model and design procedure should
there be a possibility of encountering boulders (or similar) during suction
foundation installation in clay. Assumptions made include (a) the boulder is
spherical with diameter Dboulder and (b) a constant coefficient of friction μ
models the boulder–steel interface and (c) diameter Dboulder < D should the
boulder be within the pile. The design procedure consists of three stages: (i)
check for refusal, then, if there is no refusal, (ii) check foundation tip integ-
rity and modify foundation tip geometry if necessary. Then (iii), repeat
stages (i) and (ii) if the geometry has changed. Since the forces acting on the
pile tip are needed for all stages, these are first derived in the following
paragraphs.

Figure 9.11 B
oulders, Whiterose Development, Grand Banks, Offshore
Newfoundland. PanGeo (2010). This area is unsuitable for suction
foundations - ground conditions include hard clays and very dense
granular glacial till with boulders.
122 Installation, retrieval and removal

9.10.5.1 Forces on boulder and pile tip

Figure 9.12a shows top and plan view of the boulder and the pile wall con-
tact point. This point, which is usually offset from the boulder centreline, is
defined by an angle θxz measured anti-clockwise from the horizontal. Hence
the pile wall is precisely on the boulder centreline when θxz = 90°, and the
pile is just grazing the edge of the boulder when θxz = 0° (boulder inside the
pile) or 180° (boulder outside the pile).
Figure 9.12b shows the corresponding forces acting on the boulder, both
inside and outside the pile. Rs, the soil resistance may be calculated by stan-
dard bearing capacity formulae (Rs = ANcsu). Unlike shallow foundations,
the boulder is “deep” (depth > 4 Dboulder, say) and Rs is omnidirectional (i.e.
independent of load inclination).
Knowing αxz, Rs and Wsub,boulder, the unknown boulder force Fn and angle
βxz can be derived using the polygon of forces, and hence the required pile
tip lateral and vertical forces Fx and Fz, and their resultant Rp, can be found
(see Figure 9.13).
Pile tip forces Fx, Fz and Rp for any pile tip penetration L and angle θxz are
given by the equations:
If 0 ≤ θxz ≤ 90°

Fx = Fnfac1 (9.23)

Fz = Fnfac2 (9.24)

Figure 9.12 B
oulder in clay – left, boulder and pile wall contact point, and right,
forces acting on the boulder.
Installation, retrieval and removal 123

Figure 9.13 B
oulder in clay – boulder polygon of forces and equivalent pile tip
forces.

If 90 o ≤ θxz ≤ 180°

Fx = Fnfac3 (9.25)

Fz = Fnfac4 (9.26)

where

fac1 cos xz sin xz (9.27)

fac2 sin xz cos xz (9.28)

fac3 cos xz sin xz (9.29)

fac4 sin xz cos xz (9.30)

xz 180o xz (9.31)

124 Installation, retrieval and removal

and
Rp F x
2
Fz 2 (9.32)

and

2


Fn R s 2 1 2 – Wsub,boulder fac1 Wsub,boulder fac2 / 1 2 (9.33)

Tn Fn (9.34)

R s = Aboulder N c,circle su, tip (9.35)

N c,circle min 6 1 0.2 L / Dboulder , 9 (9.36)

Wsub,boulder Vboulder sub,boulder (9.37)

Aboulder Dboulder 2 / 4 (9.38)

Vboulder 4 / 3 Dboulder / 2
3
(9.39)

Figure 9.14 shows how pile tip forces Fx, Fz and Rp vary with θxz (Example
1 following, L = 17 m, μ = 0.6), and Table 9.7 lists force variations for

oulder in clay – pile tip forces versus θ xz . Example 1: D boulder = 2

Figure 9.14 B
m at L = 17 m.
Installation, retrieval and removal 125

Table 9.7 Boulder and Pile Tip Forces – Sensitivity to Friction Coefficient μ.

μ δ F x,max F z,max R p,min R p,max R p,max /∆R boulder

[–] [deg] [kN] [kN] [kN] [kN] [–]
0.2 11 963 937 937 982 1.048
0.4 22 903 938 937 972 1.037
0.6 31 827 938 938 964 1.028
0.8 39 749 938 937 958 1.022
1.0 45 674 937 937 953 1.017

Notation:
μ = boulder-pile friction coefficient
δ = boulder-pile friction angle, atan(μ)
Fx,max = maximum pile tip lateral force
Fz,max = maximum pile tip axial force
Rp,min= minimum pile tip resultant force, Rp =√(Fx2 + Fz2)
Rp,max= minimum pile tip resultant force, Rp =√(Fx2 + Fz2)
Rs = soil resistance on boulder, Nc Aboulder su
Wsub,boulder = boulder submerged weight,Vboulder γsub,boulder
∆Rboulder = net soil resistance on boulder, Rs - Wsub,boulder
Notes:
(1) Refusal Example 1 – weak clay with 2 m diameter dropstone
(2) R = 995kN
s
(3) W
sub,boulder = 57 kN
(4) ∆R
boulder = Rs - Wsub,boulder = 938 kN

friction coefficient μ varying between 0.2 and 1.0. This figure and table are
key items. Note that:

(a) Fx, Fz and Rp kinks at θxz = 90° are due to the sign change in Tn
(b) maximum lateral Fx has almost the same magnitude as maximum
axial Fz. (Example 1: Fx,max = 827 kN, Fz,max = 938 kN)
(c) axial Fz = Fn when θxz = 90°
(d) lateral Fx = Tn when θxz = 0° and 180°
(e) maximum axial Fz occurs when θxz = atan(1/μ) (Example 1: 59° for
μ = 0.6)
(f) resultant Rp magnitude is essentially constant (Example 1: Rp,min = 938
kN, Rp,max = 964 kN)
(g) maximum resultant Rp.max ≈ ∆Rboulder (Example 1: Rp,max = 964 kN,
Rs = 995 kN, Wsub,boulder = 57 kN, ∆Rboulder = Rs – Wsub,boulder = 938 kN,
Rp,max/∆Rboulder = 1.028)

Findings (f) and (g) make possible a reasonably accurate simplification for
both refusal and integrity assessments. That is, ∆Rboulder (= Rs − Wsub,boulder)
the net soil resistance, can be applied as a force Rp.max on the pile tip at any
angle θxz. – anywhere from vertical to horizontal. Since the force initiating
lateral buckling is 50% of that for axial buckling (Aldridge et al., 2005), this
means that lateral is more important than axial.
126 Installation, retrieval and removal

9.10.5.2 Refusal
Encountering boulders causes “spikes” in soil resistance–depth profiles. This
is similar to CPTu encountering gravel in sand, where the qc value temporar-
ily increases as the obstruction is pushed aside. However, underpressure ∆u
increases whereas CPT u2 decreases. Refusal occurs if ∆u exceeds allowable
Δuall,install.

9.10.5.3 Refusal example 1 – weak clay with dropstone

Example 1 re-uses the installation data example given in Section 9.10.4.
Check for pile refusal during suction-assisted phase should a 2 m diameter
boulder (dropstone) be encountered within the LOC clay at final tip pene-
tration. To maximise the axial resistance increase, assume that the initial
point of contact is on top of the spherical boulder. Take force Rp = resistance
∆Rboulder

Relevant data are:

Pile: D = 4 m, Lcan = 17 m, WT = D/200 = 20 mm, Di = 3.96 m, Aplug =

12.32 m2,
Wsub,steel = 750 kN
Soil: clay, su,z = 5 + 2 z [kPa, m]
Pile-Soil: αi = αo = 0.30
Boulder: Dboulder = 2 m (Aboulder = 3.14 m2, Vboulder = 4.19 m3),
γsub,boulder = 16 kN/m3 (Wsub,boulder = 536 kN), μ = 0.6.

At the end of penetration (L = 17 m) and no boulder (from the Section 9.10.4

example), we have:

Rsuction = Fi + Fo + Qw = 1396 kN + 1410 kN + 73.1 kN = 2.88 MN

∆usuction = ( Rsuction − Wsub,steel) /Aplug = (2.88 MN − 0.75 MN) /12.32 m2 =
173 kPa
Δuall,install = Fi/Aplug + Nc,circle su,av,tip /γm,install
= 1396 kN/12.32 m2 + 9.0 × 40 kPa/1.5 = 113 kPa + 240 kPa =
353 kPa

For a 2 m diameter boulder at the pile tip, we have:

su,tip = su at L (17 m) = 5 kPa + 2 kPa/m × 17 m = 39 kPa

Rboulder = Nc,circle su,tip Aboulder = 9.0 × 39 kPa × 3.14 m2 = 1102 kN
∆Rboulder = Rboulder − Wsub,boulder = 1102 kN − 536 kN = 566 kN
Rsuction,boulder = Rsuction + ∆Rboulder = 2.88 MN + 566 kN = 3.45 MN
∆usuction,boulder = (Rsuction,boulder − Wsub,steel) /Aplug = (3.45 MN − 0.75 MN)
/12.32 m2 = 219 kPa
Installation, retrieval and removal 127

It is seen that, if a single boulder is met at final tip penetration (17 m bsf),
the required underpressure ∆usuction,boulder rises by 46 kPa to 219 kPa (was
173 kPa). The corresponding allowable suction pressure Δuall was 353 kPa.
Hence, the design is satisfactory – there is no refusal (∆usuction,boulder ≤ Δuall).
Had the boulder been encountered at a shallower depth, or the initial con-
tact point was not directly on top of the boulder, then both Rboulder, and
∆usuction,boulder would be smaller. Note that ∆Rboulder, the net soil resistance,
subtracts Wsub,boulder from Rboulder.
Foundation tip integrity assessments will subsequently demonstrate that
the tip WT has to be increased from 20 mm to at least 35 mm to preclude
local buckling.

9.10.5.4 Refusal example 2 – competent clay/glacial till with

boulders
Example 2 is similar to Example 1 but with three clay layers, 1 m of 10 kPa
soft drape, underlain by 4 m of firm to stiff clay (su = 40/100 kPa), and very
stiff 175 kPa glacial till containing occasional boulders below 5 m depth.
The stiffer clays require a stubbier pile (D = 5 m, L/D = 2). Check for refusal
should two 2 m diameter boulders be met at final tip penetration.

Pile: D = 5 m, Lcan = 10 m, WT = D/100 = 50 mm, Di = 4.9 m,

Aplug = 18.86 m2,
Wsub,steel = 1750 kN
Soil: clay, su,z = 10 kPa, z ≤ 1 m; su,z = 40/100 kPa, 1 < z ≤ 5 m;
su,z = 175 kPa, z > 5 m
Pile-Soil: αi = αo = 0.50
Boulder: same as Example 1.

Because both pile and soil data are different, a separate calculation using the
Section 9.10.4 program is needed. The fine details are not presented. At tar-
get penetration depth (L = 10 m) and no boulder, we have:

Rsuction = Fi + Fo + Qw = 8.97 MN + 9.15 MN + 1.02 MN = 19.14 MN

∆usuction = (Rsuction − Wsub,steel) /Aplug = (19.14 MN − 1.75 MN) /18.86 m2
= 0.92 MPa
Δuall,install = Fi/Aplug + Nc,circle su,av,tip /γm,install
= 8.97 MN/18.86 m2 + 8.5 × 175 kPa/1.5 = 0.47 MPa + 0.99 MPa
= 1.46 MPa

For two boulders at the pile tip, we have:

su,tip = su,z at L (10 m) = 175 kPa

Rboulder = Nc,circle su,tip Aboulder = 8.5 × 175 kPa × 3.14 m2 = 4.67 MN
∆Rboulder = Rboulder − Wsub,boulder = 4.67 MN − 536 kN = 4.14 MN
128 Installation, retrieval and removal

Rsuction,boulder = Rsuction + 2 × ∆Rboulder = 19.14 MN + 2 × 4.14 MN = 27.42

MN
∆usuction,boulder = (Rsuction,boulder − Wsub,steel) /Aplug = (27.42 MN − 1.75 MN)
/18.86 m2 = 1.36 MPa

It is seen that, if two boulders are met at final tip penetration (10 m bsf),
the required underpressure ∆usuction,boulder is 1.36 MPa (was 0.92 MPa). The
corresponding Δuall value is 1.46 MPa. There is again no refusal (∆usuction,boulder
≤ Δuall), but is marginal (∆usuction,boulder/Δuall = 0.93). This implies that γm,install
is closer than usual to, but still exceeds 1.5. Major challenges are likely had
both boulders been larger than 2 m diameter. Foundation tip integrity assess-
ments will subsequently demonstrate that, for the current pile and soil data,
WT has to be increased from 50 mm to 100 mm to preclude local buckling
and buckle propagation. Underpressures are extremely high, and structural
steel checks are required. A domed top is necessary.

9.10.5.5 Foundation tip Integrity

If pile tip refusal is unlikely, then foundation steel tip integrity assessments
should be made. According to the analytical methods of Aldridge et al.
(2005), a local buckle will occur at the pile tip when the lateral force Fx or
a near-axial force Fz values are:
Fx 1.4 y WT 2 (9.40)

Fz 2.8 y WT 2 (9.41)

where
y pile yield stress 345 MPa for steel (9.42)

This local buckle will propagate if:

D / WT
3

5 1 soil 2 E pile / E soil (9.43)

where

soil soil Poisson’s ratio 0.5 for clay (9.44)

E pile pile Young’s modulus 210e3 MPa for steel (9.45)

E soil soil Young’s modulus 400 su for clay (9.46)

Since both examples do not refuse, pile tip integrity assessments should be
made. Again, take forces Fx and Fz equal to net boulder resistance ∆Rboulder.
Example 1 and 2 ∆Rboulder values were 0.566 MN and 6.14 MN respectively.
Installation, retrieval and removal 129

Table 9.8 Pile Tip Integrity Assessment Examples.

Refusal Example 1 Refusal Example 2

– weak clay with – competent
dropstone clay/glacial till

Original Revised Original Revised

Parameter Units WT WT WT WT
D [m] 4 4 5 5
WT [mm] 20 35 50 100
WT/D [–] 1/200 1/100 1/100 1/50
s u,tip [kPa] 39 39 175 175
F x,boulder = ∆R boulder [MN] 0.57 0.57 4.67 (1) 4.67
F z,boulder = ∆R boulder [MN] 0.57 0.57 4.67 4.67
Local Buckle – Equation (9.40) F x,buckle = 1.4 σ y WT 2
F x,buckle [MN] 0.19 0.59 1.21 4.83
Local x-buckle? [Yes/No] Yes No Yes No
Local Buckle – Equation (9.41) F z,buckle = 2.8 σ y WT 2
F z,buckle [MN] 0.39 1.54 2.42 9.66
Local z-buckle? [Yes/No] Yes No Yes No
Buckle Propagation – Equation (9.43) (D/WT) 3 < 5 (1 – ν soil 2 ) E pile /E soil
(D/WT) 3 [–] 8000e3 n/a 1000e3 n/a
5(1 – ν soil 2 )E pile /E soil [–] 50.5e3 n/a 7.9e3 n/a
Buckle propagation? [Yes/No] Yes n/a Yes n/a

Notation:
n/a not applicable - no local buckling
Notes:
(1) point load from a single boulder
(2) pile: σ = 345 MPa, E
y pile = 210e3 MPa
soil = 400 su,tip, νsoil = 0.5
(3) soil: E

If local buckling occurs, then increment WT in 5 mm steps until buckling

stops to find the minimum required WT. Table 9.8 presents results. It is seen
that:

(a) Example 1. Buckling (and propagation) occurs with 20 mm WT:

Fx,boulder (0.57 MN) is larger than Fx,buckle (0.19 MN). However, with a
revised WT value of 35 mm, there is no local buckling: Fx,boulder (0.57
MN) no longer exceeds Fx,buckle (0.59 MN). Propagation is not appli-
cable – there is no buckling.
(b) Example 2 has the same outcome as Example 1: the original WT (50
mm) needs to be increased to at least 100 mm, where Fx,boulder (4.67
MN) is less than Fx,buckle (4.83 MN).

9.10.5.6 Commentary – suction pile

Boulders in sand cause suction pile refusal. Other installation methods (e.g.
impact driving) are necessary. Hence there is no comparable text to be found
in Section 9.11 (Installation in Sand).
130 Installation, retrieval and removal

The friction coefficient μ value (whether static or sliding), and where the
point of contact is on the boulder, are both irrelevant: it is reasonable to
assume that the Rp magnitude remains essentially the same. It is also consid-
ered sufficiently accurate to use the net force (∆Rboulder = Rs − Wsub,boulder) for
both axial and lateral integrity assessments.
The model described here has assumed that the bearing capacity factor of
a sphere equals Nc,circle. Ball penetrometer research indicates higher Nc values
for deep (full flow) conditions, namely 11 < Nc,ball < 15 depending on the
interface α (0 − 1) value (Randolph et al., 2000). No solutions are available
for “wished-in-place” near-surface spheres at shallow depth, but Nc values
are also expected to be higher than those for shallow circular foundations.
As usual, opinions differ, but 6 < Nc,circle < 9 used herein is considered reason-
able: major model inaccuracies include assumptions of shape (e.g. an ellipse
instead of sphere) and ULS (the boulder is pushed aside before achieving the
peak soil resistance).
In practice, since the geometry has changed for both the examples, Stage
(iii) (= Stage (i) re-check for refusal followed by Stage (ii) reassess pile tip
integrity) should be carried out. Stage (i) implies additional ”no boulder”
installation analyses with (a) increased Wsub,steel and (b) increased WT values.
However, increasing Wsub,steel further increases Δuall, which is beneficial. The
”no boulder” soil resistance Rsuction value will not increase markedly, since
the Qwall increase will be offset by a smaller Fi value. It is also advisable to
perform analyses with other alpha values (αLE, αBE, αHE), and to check that
there is no risk of over-penetration (particularly for the weak clay Example 1)
due to increased foundation weight, Wsub,steel.

9.10.5.7 Commentary – anchor chain

Since the anchor chain will contact a (spherical) boulder if the boulder is
within a distance Dboulder/2 of the chain edge, Vchain, the soil volume swept
out by a (straight, constant θlug) chain, is given by:
Vchain = Achain Wchain (9.47)

where
Achain cross sectional area
1 / 2 z lug / tan volume z lug Dboulder / 2 (9.48)

Wchain plan width Dchain Dboulder (9.49)

Dchain chain effective width in bearing
2.5 nominal chain diameterr (9.50)

Vchain may be compared with soil volumes swept out by an anchor pile.
Plugged implies that the soil plug contains internal stiffeners/protuberances,
Installation, retrieval and removal 131

whereas coring is for an essentially constant WT pile with no pad-eye stiff-

eners etc. Values of Vpile,plugged and Vpile,coring are given by:

Vpile, plugged / 4 D Dboulder L Dboulder / 2

2
(9.51)

Vpile,coring / 4 D Dboulder D Dboulder L Dboulder / 2 (9.52)

2 2

Typically, Vchain ≈ 0.5 Vpile,coring, but, as usual, swept volumes are sensitive
to pile, chain and boulder geometry. Nevertheless, assuming boulders are
randomly distributed in 3D space, then, even though the probability of
boulder contact is greater for the pile than the chain, due diligence implies
that the chain–boulder contact should also be analysed. The major issue is
that θlug will increase, decreasing the FOS on axial pull-out failure. Whatever
analysis option is selected, the crux is to select parameter values giving the
required p (soil normal/bearing resistance per unit chain length) and f (soil
frictional resistance per unit chain length) (Vivatrat et al., 1982) and to uni-
formly distribute p and f over an embedded chain length, ∆Lchain modelling
the contact zone. A reasonable ∆Lchain value is Dboulder /4. Using the pile tip
model, p = (Rs − Wsub,boulder) /∆Lchain and f = μp. Analysis options include (a)
introducing a competent soil layer at the boulder depth and (b) modifying
the chain analysis program. These options are left to interested engineers to
investigate.

9.11 INSTALLATION IN SAND

9.11.1 Sand plug liquefaction, piping and heave in

sand
Liquefaction, piping and heave are closely interlinked. Underpressure cre-
ates an axial upwards flow/seepage in the sand plug. When this is sufficiently
high, inner friction and tip resistance have both been sufficiently reduced
and penetration occurs. If the underpressure exceeds that causing liquefac-
tion and/or piping, then penetration is halted and installation fails (liquefac-
tion possibly results in sand in the suction pump and piping drops the
underpressure to zero, and resistance increases). Due to the decreased effec-
tive stress within the sand plug, soil expansion occurs, and this causes addi-
tional sand plug heave. Additional side-effects of seepage/heave include
increased permeability and reduced strength of the soil plug, both due to an
increased voids ratio. All are discussed in more detail here.
Hence, the key design challenges for installing intermediate foundations
in sand are:

(a) to sufficiently reduce inner friction and foundation tip resistance such
that penetration can occur
132 Installation, retrieval and removal

without incurring

(b) liquefaction failure of sand plug

(c) piping failure around foundation wall
(d) excessive sand plug heave

and, if in shallow water,

(e) the maximum underpressure should not exceed the lowest astronomi-
cal tide (LAT) water depth (Section 9.7).

Ad (a) reduced inner friction and foundation tip resistance. Unlike clay,
most resistance is from the foundation tip, and the primary effect of under-
pressure is to reduce this tip resistance rather than to increase the load
(which is a secondary effect).
Ad (b) sand plug liquefaction failure. Underpressure is traditionally lim-
ited to that which gives zero vertical effective stress, i.e. seafloor icrit (critical
hydraulic gradient) = γsub /γwater.
The foregoing implies that underpressure cannot exceed this critical value.
However, there is some evidence from carefully executed laboratory perme-
ability tests (e.g. Fleshman and Rice, 2014; Panagoulias, 2016; Panagoulias
et al., 2017), and from foundation installation, both in the field and centri-
fuge (summarised by Panagoulias, 2016; Panagoulias et al., 2017) that i >
icrit in medium-dense to very dense sands. Figure 9.15 shows suction

Figure 9.15 N
ormalised suction penetration curves (Panagoulias et al., 2017)
(With permission of SUT, London)
Installation, retrieval and removal 133

penetration data for six North Sea locations, together with the theoretical
(Houlsby and Byrne, 2005) icrit curves. No sand plug liquefaction occurred.
In addition, no excessive sand plug heave was reported in any of the six
cases considered. Hence, provided that due care is taken, gross failure is
unlikely at underpressures (slightly) exceeding this critical hydraulic gradi-
ent. However, at the time of writing, a credible model has yet to be pub-
lished, and installation procedures have not yet been generally accepted.
Upward seepage, even if at pressures less than critical, also creates some
loosening/density reduction within the soil plug (Tran, 2005). This leads to
internal sand heave, which, if not allowed for by stick-up/over height, may
prevent the foundation penetrating to its design depth. In addition, Tran’s
small-scale model tests showed that no major adverse hydraulic conditions
(only 10% heave) occurred, even with L/D ratios up to 2.0 and critical gra-
dient > 1. Another factor mitigating severe sand plug liquefaction is proba-
bly passive arching in aged sands – arching theory (for the no-flow case) has
been given by Paikowsky (1989).
The loosening/density reduction effect needs to be accounted for in in-
place resistance assessments – the sand plug strength is reduced from origi-
nal in-situ values.
Ad (c) piping failure. More important than heave is piping. If sand “boil-
ing” occurs, piping channels may be locally formed, especially along the
wall/soil plug interface. These will create a hydraulic leak – the water plug
beneath the top plate is in direct contact with the seawater via a continuous
channel along the foundation inner and outer wall. In this case, the suction
pump is merely removing seawater. There is no water flow within the soil to
reduce inner friction and end-bearing resistance. Also, a mixture of sand and
water may come into the suction pump. In this case, installation will fail.
From the very limited amount of data available, it appears that piping
failure is more likely when the foundation is restrained from penetration.
The basic reason is that the foundation tip can no longer penetrate into
weakened soil. Hence, obstructions in sand, such as cobbles or boulders, can
quickly initiate local piping.
Ad (d) excessive soil plug heave. This is greater in sand than in clay (for a
given foundation geometry). This is because there is an additional compo-
nent of sand expansion due to stress decreases from upwards flow and other
effects. If high installation pressures are anticipated, and domed top plates
are required, these can take more soil plug heave than flat top plates. In the
Gorm Field caisson installation (Senpere and Auvergne, 1982) excessive
sand heave and piping failure was observed. Excess sand had to be removed
by water jetting to enable the caissons to reach design penetration depth. No
installation details were released – e.g. underpressure magnitudes.
Ad (e) maximum underpressure should not exceed the LAT water depth.
This is a potential problem only in shallow water. European examples
include the southern North Sea, Irish Sea and the Baltic.
134 Installation, retrieval and removal

Due diligence/mitigation measures for minimising the risk of liquefaction,

piping, soil plug heave, etc., include:

• instrumentation
• careful installation
• retrieval/reinstallation
• pump flowrate (as high as possible, e.g. Tran, 2005)
• increased preload (gives smaller underpressure)
• reverse flow (i.e. overpressure) at end of installation. This is similar to
hydraulic sand fill placement causing compaction. Reverse flow
pushes back the sand plug and increases plug relative density/strength,
but a potential disadvantage is that it increases/creates a void under
baseplate.

9.11.2 Models for sand

As noted earlier, the effect of water flow has to be considered, and penetration
resistance methods can be either the CPT qc or beta type (see Section 9.8.1).
A variety of analytical and empirical models have been developed and applied
to assess installation resistance in sand. At the time of writing, there is no
single industry-wide accepted method. More information can be found in the
following references (in chronological order):

Erbrich and Tjelta (1999) Statoil/Kvaerner beta

Houlsby and Byrne (2005) Oxford beta
Andersen et al. (2008) NGI beta and CPT qc
Senders and Randolph (2009) UWA CPT qc
Chatzivasileiou (2014) SPT CPT qc
Sturm et al. (2015) NGI beta.

Installation data for fewer than 20 projects are in the public domain (e.g.
Panagoulias, 2016; Alderlieste and Van Blaaderen, 2015). Hence, the models
described here (and associated parameter values) are likely to evolve/
improve as more data become available and additional back-analysis is per-
formed. In addition, attempts are made to numerically investigate this phe-
nomenon; see e.g. Martinelli et al. (2020).
Figure 9.16 compares critical underpressure as a function of embedment
ratio based on published equations. Only two (Houlsby and Byrne, 2005;
Sturm et al., 2015) account for a lower soil plug permeability. All equations
are for a water flow source at seafloor (not for a source at a near-infinite
radius), i.e. they cannot be used for confined sand aquifers.
Installation, retrieval and removal 135

Figure 9.16 C
omparison critical suction pressures (based on Senders and
Randolph, 2009). Critical underpressure, ∆u crit /(γ'D) versus embed-
ment ratio L/D, k r : inner soil plug/outer soil permeability ratio, [–].

9.11.3 Sand installation example

To provide example installation resistance/underpressure–penetration depth
diagrams, the DNV CPT qc resistance equations can be programmed to pro-
vide a simple model for sand installation. The resulting model is capable of
modifications/improvements to accommodate different opinions.
Assumptions are numerous and include:

uniform caisson D and WT (i.e. no protuberances)

no silo effect within sand plug
DNV CPT qc method (i.e. qw = kpqc and f = kfqc)
uniform permeability sand (i.e. no sand plug loosening)
outer skin friction: no increase due to flow
inner skin friction and wall tip end-bearing: linear decrease proportional
to 1 − isand
critical ∆u based on average (not exit) hydraulic gradient isand
Houlsby and Byrne (2005) αu − L/D equation.

Using these assumptions, the equations to find penetration resistance Rsuction

and corresponding underpressure ∆usuction at any tip penetration L are:

R suction Fi,suction Fo,suction Q w,suction (9.53)

136 Installation, retrieval and removal

usuction R suction Wsub,steel / A plug (9.54)

where
Fi, suction Fi,no _ flow 1 isand (9.55)
Fo,suction = Fo,no _ flow (9.56)
Q w,suction Q w,no _ flow 1 isand (9.57)

Fi ,no _ flow Asik f ,sand q c, zdz (9.58)

Fo,no _ flow Asok f ,sand

q c, z dz (9.59)

Q w,no _ flow = A wallk p,sandq c,L (9.60)

isand 1 u usuction / subL (9.61)

u 0.45 0.36 1 exp L / D / 0.48 (9.62)

Since iteration is required to balance loads and resistances, we have two

more equations:
Lsuction A plug usuction Wsub,steel (9.63)

R suction = Lsuction (9.64)

where, noting that subscripts “suction”, “liq” and “no_flow” mean “with
suction assistance and fluid flow”, “required for full sand plug liquefaction”
and “with suction assistance but no fluid flow” respectively:
αu = excess pore pressure ratio at tip level
= Δusuction (tip)/Δusuction
Δuno_flow = water (under) pressure – no fluid flow
Δuliq = water (under) pressure – full liquefaction
Δusuction = water (under) pressure – fluid flow
γsub = sand plug submerged unit weight
γwater = (sea) water unit weight
Aplug = soil plug area
Ap = caisson wall tip area
Asi = caisson inner perimeter
Aso = caisson outer perimeter
D = caisson outer diameter
Di = caisson inner diameter
Installation, retrieval and removal 137

Fi,no_flow = caisson inner wall friction resistance – no fluid flow

Fi,liq = caisson inner wall friction resistance – full liquefaction
Fi,suction = caisson inner wall friction resistance – fluid flow
Fo,no_flow = caisson outer wall friction resistance – no fluid flow
Fo,liq = caisson outer wall friction resistance – full liquefaction
Fo,suction = caisson outer wall friction resistance – fluid flow
isand = sand plug (vertical) average seepage gradient
kf,sand = DNV empirical coefficient – unit skin friction resistance in sand
kp,sand = DNV empirical coefficient – unit end-bearing resistance in sand
L = caisson tip penetration into sand
Lcan = caisson total length
Lsuction = load on caisson steel – fluid flow
qc,z = CPT cone tip resistance at depth z below seafloor
qc,L = CPT cone tip resistance at caisson tip depth L
Qw,no_flow = caisson wall tip resistance – no fluid flow
Qw,liq = caisson wall tip resistance – full liquefaction
Qw,suction = caisson wall tip resistance – fluid flow
Rno_flow = soil resistance – no fluid flow
Rliq = soil resistance – full liquefaction
Rsuction = soil resistance – fluid flow
Wsub,steel = caisson submerged weight (including ballast)
WT = caisson wall thickness

The corresponding pseudo code for the installation in the sand model is:

function αu_fun: evaluate αu at any caisson tip penetration

L and diameter D
ref Houlsby and Byrne, 2005
in: L, D // out: αu
c0 = 0.45, c1 = 0.36 and c2 = 0.48
αu = c0 – c1 [1 – exp {(-L/D) /c2}]
end

function isand_fun: evaluate isand at any caisson tip penetration L, diameter

D, underpressure ∆u and soil submerged unit weight γsub

in: L, D, ∆u, γsub // out: isand

αu = αu_fun(L, D)
isand = [1 − αu ∆u/ (γsub L) ]
isand = min(isand, 1)
end
138 Installation, retrieval and removal

program install_sand
initialise and evaluate constants
– 2 WT, Asi = π Di, Aso = π D, Ap = π (D2 - Di2) /4 and
Di = D
Aplug = π Di2 /4
∆L = 0.1 m
nL = Lcan /∆L + 1
zero matrix ANS (size nL rows by 21 columns)
k = 0

loop tip penetration L from 0 to Lcan in ∆L steps:

k = k + 1
L = (k − 1) ∆L
Fi,no_flow: inner skin friction resistance by integrating qc
from 0 to L
Fo,no_flow: outer skin friction resistance by integrating qc
from 0 to L
Fi,no_flow = Asi kf,sand ∫ qc,z dz and Fo,clay = Aso kf,sand ∫ qc,z dz
Qw,no_flow: end-bearing resistance using qc at depth L
Qw,no_flow = Awall kp,sand qc,L
Rno_flow = Fi,no_flow + Fo,no_flow + Qw,no_flow
∆uno_flow = (Rno_flow − Wsub,steel) /Aplug
Fi,liq = 0
Fo,liq = Fo,no_flow
Qw,liq = 0
Rliq = Fi,liq + Fo,liq + Qw,liq
∆uliq = (Rliq − Wsub,steel) /Aplug
αu = αu_fun (L, D)
iterate using ∆usuction to balance resistance Rsuction and
load Lsuction
isand = isand_fun (L, D, ∆usuction, γsub)
Fi,suction = Fi,no_flow (1 − isand)
Fo,suction = Fo,no_flow
Qw,suction = Qw,no_flow (1 − isand)
Rsuction = Fi,suction + Fo,suction + Qw,suction
Lsuction = Aplug ∆usuction + Wsub_steel
exit iteration with ∆usuction, etc., when Rsuction = Lsuction
store in kth row of matrix ANS:
k, L, and L/Lcan in columns 1 through 3
Fi,no_flow, Fo,no_flow, Qw,no_flow, Rno_flow and ∆uno_flow in columns 4
through 8
αu and isand in columns 9 and 10
Fi,liq, Fo,liq, Qw,liq, Rliq and ∆uliq in columns 11 through 15
Fi,suction, Fo,suction, Qw,suction, Rsuction, Lsuction and ∆usuction in
columns 16 through 21
Installation, retrieval and removal 139

end loop on tip penetration

exit with matrix ANS

matrix ANS (nL rows, 21 columns) contents:

col# 01: k matrix ANS row number [–]

col# 02: L caisson tip penetration into sand [m]
col# 03: L / D caisson penetration/diameter ratio [–]
col# 04: F i , no_flow caisson inner wall friction – no fluid flow [kN]
col# 05: F o,no_flow caisson outer wall friction resistance – no fluid [kN]
flow
col# 06: Q w,no_flow caisson wall tip resistance – no fluid flow [kN]
col# 07: R no_flow soil resistance – no fluid flow [MN]
col# 08: ∆u no_flow water (under) pressure – no fluid flow [kPa]
col# 09: α u excess pore pressure ratio at tip level [kN]
col# 10: i sand sand plug (vertical) average seepage gradient [–]
col# 11: F i,liq caisson inner wall friction resistance – full [kN]
liquefaction
col# 12: F o,liq caisson outer wall friction resistance – full [kN]
liquefaction
col# 13: Q w,liq caisson wall tip resistance – full liquefaction [kN]
col# 14: R liq soil resistance – full liquefaction [MN]
col# 15: ∆u liq water (under) pressure – full liquefaction [kPa]
col# 16: F i,suction actual caisson inner wall friction resistance – fluid [kN]
flow
col# 17: F o,suction actual caisson outer wall friction resistance [kN]
– fluid flow
col# 18: Q w,suction actual caisson wall tip resistance – fluid flow [kN]
col# 19: R suction actual soil resistance – fluid flow [MN]
col# 20: L suction load on caisson steel – fluid flow [MN]
col# 21: ∆u suction actual water (under) pressure – fluid flow [kPa]

Plot col# 07 (Rno_flow), col# 14 (Rliq), and col# 19 (Rsuction) versus col# 02 (L)
resistance – depth
Plot col# 08 (∆uno_flow), col# 15 (∆uliq), and col# 21 (∆usuction) versus col#
02 (L) underpressure – depth

9.11.3.1 Example – sand installation

Consider a typical North Sea type profile (1.5 m of loose sand, Dr ≈ 0.3,
underlain by dense sand, Dr ≈ 0.8) in which a 10 m diameter, 12 m length
caisson has to be installed. Input data are:
Pile: D = 10 m, Lcan = 12 m, WT = D/200 = 50 mm, Wsub,steel = 6000 kN
(Di = 9.90 m, As,i = π Di = 31.10 m, As,o = π D = 31.41 m, Ap = 1.565 m2,
Aplug = 77.0 m2)
Soil: sand 0 – 1.5 m: qc = 2 z; >1.5 m: qc = 10 + 1.5 z [MPa, m], γsub =
12 kN/m3
Water: γwater = 10 kN/m3
140 Installation, retrieval and removal

Pile – Soil:
kp,sand, kf,sand = 0.3, 0.001 (BE) and kp,sand, kf,sand = 0.6, 0.003 (HE)

For the analysis using DNV qc HE coefficients, midway through penetra-

tion, at L = 6 m, we have:

k = 61
L=6m
L/D = 0.6

and

Fi,no_flow = Asi kf,sand ∫ qc,z dz = 31.10 m × 0.003 × 72.5 MN/m = 6770 kN

Fo,no_flow= Aso kf,sand ∫ qc,z dz = 31.41 m × 0.003 × 72.5 MN/m = 6840 kN
Qw,no_flow = Awall kp,sand qc,L = 1.565 m2 × 0.6 × 19 MPa = 17800 kN
Rno_flow = Fi,no_flow+ Fo,no_flow+ Qw,no_flow = 6770 kN + 6840 kN + 17800 kN
= 31.4 MN
∆uno_flow = (Rno_flow - Wsub,steel) /Aplug = (31.4 MN − 6 MN) /77.0 m2 = 330 kPa
αu = c0 − c1 [1 − exp {(−L/D) /c2}] = 0.45 − 0.36 [1 – exp {(−L/D) /0.48}]
= 0.193
Δusuction= 72.1 kPa by iteration, verified below (Rsuction = Lsuction)
isand = (1 − αu) Δusuction/ (γsub L) = 0.807 × 72.1 kPa/(12 kN/m3 6 m) = 0.81

and

Fi,liq = 0 kN
Fo,liq= Fo,no_flow = 6840 kN
Qw,liq = 0 kN
Rliq = Fi,liq + Fo,liq+ Qw,liq = 0 kN + 6840 kN + 0 kN = 6.84 MN
∆uliq = (Rliq − Wsub,steel) /Aplug = (6.84 MN − 6 MN) /77.0 m2 = 10.9 kPa

and

Fi,suction = Fi,no_flow (1 − isand) = 6770 kN (1 − 0.81) = 1300 kN

Fo,suction = Fo,no_flow = 6840 kN
Qw,suction = Qw,no_flow (1 − isand) = 17800 kN (1 − 0.81) = 3420 kN
Rsuction = Fi,suction + Fo,suction + Qw,suction = 1300 kN + 6840 kN + 3420 kN =
11.5 MN
Lsuction = Aplug ∆usuction + Wsub,steel = 77.0 m2 × 72.1 kPa + 6000 kN = 11.5 MN
∆usuction = (Rsuction − Wsub,steel)/Aplug = (11.5 MN − 6 MN) /77.0 m2 = 72.1 kPa

Figure 9.17 shows results for 10 m diameter caissons tipping out in dense
sand (Dr ≈ 80%). Figure 9.17a is for Best Estimate DNV kp and kf values. It
is seen that, if there is no flow, the Best Estimate underpressure ∆u (solid
blue/thin line) at 12 m penetration is moderate, just under 300 kPa. Since
there is no flow, there is no risk of the soil plug liquefying (the major risks
are cavitation if the WD is less than 30 m and steel cylinder implosion if the
Installation, retrieval and removal 141

Figure 9.17 I nstallation in sand. Example using DNV CPT approach. Caisson 10 m diam-
eter. 1.5 m of loose sand (qc = 2z [MPa,m], Dr ≈ 30%) underlain by dense
sand (qc = 10 + 1.5z [MPa,m], Dr ≈ 80%). (a) BE kp and kf, (b) HE kp and kf.

cylinder WT is inadequate). However, seepage significantly decreases pene-

tration resistance R and ∆u, and these are given by the red/thick lines. At
12 m penetration, ∆u is low, around 100 kPa (was ≈ 300 kPa). The dashed
blue/thin lines are for the extreme case of full liquefaction (i.e. outer skin
friction only – both inner skin friction and wall end-bearing are zero). Unless
full liquefaction occurs (see the High Estimate” bottom of figure), the full
liquefaction lines (plug hydraulic gradient i = icrit) lie below the actual seep-
age lines (i < icrit). For the Best Estimate kp and kf values, the actual seepage
lines never intersect the full liquefaction line.
142 Installation, retrieval and removal

Figure 9.17b shows the corresponding High Estimate resistance and pres-
sure depth diagrams. The situation changes dramatically: full plug liquefac-
tion first occurs at around 11 m bsf (i.e. L/D ≈ 1.1), and the final ∆u ≈ 175
kPa (was ≈ 100 kPa) at 12 m penetration. In such cases, where liquefaction
occurs only for the last metre of penetration, and only for High Estimate kp
and kf values, it is tempting to accept the risk and adopt suitable due dili-
gence procedures (Section 9.11.1) during installation.

9.11.3.2 Commentary
Large diameter caissons (D ≈ 10 m) are necessary to achieve embedment
ratios around L/D ≈ 1 to 1.5 in competent sands. This is largely due to the
fact that load is proportional to D2, whereas resistance is proportional only
to D.
All profiles are straight lines because CPT qc varied linearly in each layer.
In the real world, greater variability is expected.
The model can be modified for non-uniform caissons, other αu and isand
relationships. Unlike the clay installation example (Section 9.10.3), retrieval/
removal has been omitted, and interested engineers can easily include this.
In this case, note that (unlike clay), all Qwall terms practically become zero.
Again, Wsub,steel counteracts overpressure. Hopefully, friction “set-up” is
allowed for by judicious kf selection.
SWP was 1.5 m for both BE and HE assessments. This coincides with the
top surface of the underlying dense sand. Since this depth exceeds (say) 0.5
m, this is satisfactory and is sufficient to start the suction-assisted stage. Had
the surface loose sand layer been absent, then SWP ≈ 0, and there is a high
risk of local piping occurring around the can tip. Special measures (listed in
Section 9.9.2) need to be considered.

9.11.4 Friction set-up in sand

Short term (retrieval) set-up is generally not considered in sand but should,
however, be considered for removal.
It is generally accepted that skin friction (but not end-bearing) resistance
of driven piles in sand increases with (logarithm) time. Mechanisms (see
Jardine et al., 2005; Lehane et al., 2005) include

• creep destroying “arching”, thereby increasing soil radial effective stress

• aging increasing radial stiffness and/or dilation
• corrosion and/or aging (cementation) causing increase in pile-soil inter-
face friction angle.

Opinions differ as to rate – see Figure 28 of Jardine et al. (2005) and

Figure 18 of Schneider et al. (2007).
Installation, retrieval and removal 143

Additional information is given in:

  Jardine et al. (2005) sand and clay

  Lehane et al. (2005) sand
  Jardine et al. (2006) sand
  Schneider et al. (2007) sand
  Karlsrud et al. (2014) field tests piles sands and clays
  Lehane et al. (2014) centrifuge tests caisson un-aged sands.

With the exception of Lehane et al. (2014), all these studies were for
driven pile (not intermediate) foundations.

9.11.5 Back analysis of installation data in sand

Installation in sand is not yet well understood. The number of well-docu-
mented field projects and centrifuge tests in the public domain is limited
(Panagoulias, 2016). Consequently models are cautious and/or not optimal.
Back analysis of field installation data (e.g. Chatzivasileiou, 2014;
Frankenmolen et al., 2017) is considered to be an excellent way forward to
improve models, reduce the spread between Best Estimate and High Estimate
DNV type kp and kf coefficients, and hence enable foundations in sand to be
more confidently designed for embedment ratios L/D in excess of one.
However, back analysis is challenging – ideally at least three unknowns
(typically parameters kp, kf and αu) have to be derived using a single known
(penetration resistance R). Also, Section 9.8.4 (CPT Method Coefficient αu)
showed that αu is a function of foundation penetration, soil plug : soil mass
permeability kf and sand layer thickness. To overcome these complications,
back-analyses should adhere to the following rules:

1. Use the same model as for installation prediction.

2. Include the suction assisted zone. This allows for the possibility of
altered kp, kf values in more competent deeper sands and/or different
degrees of disturbance.
3. Keep the αu model as simple as possible. It is considered reasonable to
use the Houlsby and Byrne (2005) αu − L/D relationship (or similar)
for an infinite half space of uniform permeability. This reduces the
number of unknowns from three to two.

The challenge (more unknowns than knowns) is dissimilar to that of CPTs

– from CPT qc and Rf it is possible to derive up to two parameters, but it is
virtually impossible to derive (say) soil type and strength using only qc.
Research may eventually provide an inexpensive and robust solution to
increase the number of knowns – for example, skirt pore pressure measure-
ments. Push-out data (kp = 0) reduces the number of unknowns by one. Hence,
specifying (say) 5% extraction and back-analysis may prove beneficial.
144 Installation, retrieval and removal

9.11.6 Observational method in Peck (1969)

The observational method (Peck, 1969; Eurocode/CEN 2004) may be useful
for projects where foundation embedment ratios L/D in excess of one in
sand are being considered. Due diligence installation procedures have been
given in Section 9.11.1.
Since the observational method is a departure from conventional offshore
design practice, details are given. The observational method was described
by Peck (1969) and forms part of Eurocode geotechnical design (CEN,
2004).
Peck (1969) notes:

In brief, the complete application of the (observational) method embod-

ies the following ingredients.

(a) Exploration sufficient to establish at least the general nature, pat-

tern and properties of the deposits, but not necessarily in detail.
(b) Assessment of the most probable conditions and the most unfa-
vourable conceivable deviations from these conditions. In this
assessment, geology often plays a major role.
(c) Establishment of the design based on a working hypothesis of
behaviour anticipated under the most probable conditions.
(d) Selection of quantities to be observed as construction proceeds and
calculation of their anticipated values on the basis of the working
hypothesis.
(e) Calculation of values of the same quantities under the most unfa-
vourable conditions compatible with the available data concerning
the subsurface conditions.
(f) Selection in advance of a course of load or modification of design
for every foreseeable significant deviation of the observational find-
ings from those predicted on the basis of the working hypothesis.
(g) Measurement of quantities to be observed and evaluation of actual
conditions.
(h) Modification of design to suit actual conditions.

CEN (2004) proposes the following (Section 2.4, 4P):

If no reliable calculation model is available for a specific limit state,

analysis of another limit state shall be carried out using factors to ensure
that exceeding the specific limit state considered is sufficiently improba-
ble. Alternatively, design by prescriptive measures, experimental models
and load tests, or the observational method, shall be performed.
Installation, retrieval and removal 145

9.11.7 Observational method in CEN (2004)

1. When prediction of geotechnical behaviour is difficult, it can be
appropriate to apply the approach known as “the observational
method”, in which the design is reviewed during construction.
2. The following requirements shall be met before construction is
started:
• acceptable limits of behaviour shall be established;
• the range of possible behaviour shall be assessed and it shall be
shown that there is an acceptable probability that the actual
behaviour will be within the acceptable limits;
• a plan of monitoring shall be devised, which will reveal whether
the actual behaviour lies within the acceptable limits. The moni-
toring shall make this clear at a sufficiently early stage, and with
sufficiently short intervals to allow contingency actions to be
undertaken successfully;
• the response time of the instruments and the procedures for ana-
lysing the results shall be sufficiently rapid in relation to the pos-
sible evolution of the system;
• a plan of contingency actions shall be devised, which may be
adopted if the monitoring reveals behaviour outside acceptable
limits.
3. During construction, the monitoring shall be carried out as planned.
4. The results of the monitoring shall be assessed at appropriate stages
and the planned contingency actions shall be put into operation if
the limits of behaviour are exceeded.
5. Monitoring equipment shall either be replaced or extended if it
fails to supply reliable data of appropriate type or in sufficient
quantity.

9.11.8 General
Installation studies are difficult and opinions/models/results differ widely.
Key challenges usually include:

(a) unit inner skin friction and end-bearing reduction in the sand aquifer
due to upwards seepage in soil plug
(b) water pocket in soil plug at the clay/sand interface. This is because
installation stops if separation occurs and the clay plug meets the
foundation top plate (see Figure 8.2d)
(c) hydraulic fracture of the clay plug (see Figure 8.2f). This is perceived
to be generally associated with competent clay and moderate
underpressures.
146 Installation, retrieval and removal

Background reading includes:

• Tran (2005) sand and silt layers

• Watson et al. (2006) clay over sand, clay plug heave
• Senders et al. (2007) clay over sand and clay plug hydraulic fracture
model
• Romp (2013) clay over sand, clay plug heave and hydraulic
fracture (cracking)
• Plug heave see Section 9.10.2 (on plug heave in clay) and
Section 9.11.2 (on models for sand).

9.11.9 Water pocket model

This section is about a model for water pocket development (until now
unpublished). This is an improved version of the original (1996) Fugro con-
cept but has been greatly improved for water flow. It was used on a project
considering 15 m diameter caissons as FPSO anchors under 24 MN lateral
load.
The following questions arise when installing suction caissons in mixed
clay and sand layers: (a) Is there a possibility of clay plug–sand plug separa-
tion, causing a water pocket to be formed at the sand–clay interface? (b) Is
there any reduction in inner friction and tip resistance in the sand layers?
The model is sketched in Figure 9.18 and is analogous to a piston (the

ater pocket. Water pocket, height ∆H pocket , at bottom of clay plug.

Figure 9.18 W
Clay plug height increase ∆H heave = ∆H pocket . Foundation penetration
L sand into sand layer.
Installation, retrieval and removal 147

impermeable clay plug) being pulled out of a cylinder embedded in perme-

able material (the sand aquifer). The soil profile consists of clay (height Hclay)
underlain by sand (Hsand). The caisson is penetrating Lsand into the sand aqui-
fer. If there is fluid flow Qflow from the sand, a water pocket, height ∆Hpocket,
will form with an underpressure ∆ubase. The upwards-acting differential
force, magnitude (∆utop − ∆ubase) Aplug, pushes the clay plug upwards. There
will be no water pocket if Qflow is zero. Conversely, according to the model,
a water pocket will start to form as soon as Qflow > 0.
Model assumptions include the following:

(a) The clay plug is impermeable. Hence, there is no fluid flow (i.e. no
head drop) within the clay plug
(b) The sand response is fully drained. Hence there are no excess pore
pressures (effective stress = total stress, pore pressures are steady
state).

The objective is to find the sand aquifer flowrate Qflow at any penetration
Lsand. From this, and knowing the time to penetrate the sand aquifer, ΔHpocket,
the required water pocket height, i.e. clay plug heave, ∆Hheave, can be
computed.
The two steady state flow approximate equations are:

(a) sand annulus + plug : flowrate Qflow − Δubase

Q flow 8 / 13 water [ ksandHsand / HLD ubase (9.65)

sand annulus + plug: upwards hydraulic gradient iplug − Δubase

iplug 1 – 1 / HLD / Lsand water ubase (9.66)

There are also two equilibrium equations and another two linking sand
resistances to the hydraulic gradient in the sand plug:
(b) forces on caisson steel
utop A plug Wsub,steel Fi,clay Fo,clay Fi,sand Fo,sand Q w,sand (9.67)

forces on clay plug

utop A plug ubase Fi,clay Wsub _ clay (9.68)

friction and wall end-bearing resistances in sand

Fi,sand Fi,sand,no _ seep 1 ibase (9.69)

Qw,sand Q w,sand,no _ seep 1 ibase (9.70)

148 Installation, retrieval and removal

Figure 9.19 W
ater pocket. Steady state radial fluid flow – sand plug and aquifer
annulus components are connected with no change in hydraulic head.

Assumptions, limitations and reasonableness of these equations include:

Ad (a) and (b): as sketched in Figure 9.19, the sand layer has been sub-
divided into two components, an annulus (height Hsand, inner and outer
radii D/2 and 10 D, with a cut-off height Lsand equal to the tip penetra-
tion) representing the outer portion of the confined sand aquifer, and the
sand plug (diameter D, height Lsand). The volume of sand in the aquifer
directly below the sand plug is missing, for which it is assumed that a
frictionless pipe transmits all radial flow out of the aquifer annulus into
axial upwards flow in the sand plug. Steady state axisymmetric fluid flow
FEA were made of the annulus for varying Lsand/Hsand ratios (0–1) to
obtain Qflow − Δubase and Qflow − iplug relationships. Both are considered
reasonable approximations and are generally cautious (i.e. they overpre-
dict Qflow and ΔHpocket), as shown in Figure 9.20. Better (slightly more
accurate) project specific relationships are possible (e.g. partially pene-
trating well theory or FEA), but, as will be shown, heave assessments are
strongly dependent on ksand.
These considerations suggest that there cannot be installation resistance
reduction in sand without a water pocket – there must be flow. No-flow
resistances apply if there is no seepage.
Ad (c): force equilibrium - caisson. The caisson is penetrating into the soil;
hence the downwards-acting underpressure and caisson weight are opposed
by upwards acting soil resistances. Note that the clay resistances remain
constant, but in the sand they are a function of tip penetration Lsand and iplug.
In addition (unlike free-fall penetration), the buoyancy force is in water, not
soil penetration dependent, and therefore constant.
Ad (d): force equilibrium – clay plug. Again, caisson penetration is
assumed. Hence, inner friction resistance Fi,clay and underpressure Δubase Aplug
both act downwards opposing caisson motion.
Installation, retrieval and removal 149

Figure 9.20 W
ater pocket. Comparisons Q flow versus Δu base and i plug versus Δu base
approximate relationships with axisymmetric FEA at H sand /D = 0.4
and 1.0. Diameter D = 18.5 m, k sand = 1e-4 m/s.

Ad (e): caisson friction and end-bearing in sand. Opinions differ, and a

simple linear reduction of both inner friction and tip resistance (and no
outer friction resistance reduction) with sand seepage gradient iplug has been
assumed. Since inner friction and end-bearing both become zero when
iplug = 1, iplug cannot exceed unity. Other end-bearing possibilities include
50% when iplug = 1, but ΔHpocket is more sensitive to Qsuction and ksand than
resistance.
There are six variables and six equations. Hence the five unknowns (ibase,
Δutop, Δubase, Fi,sand and Qw,sand), all of which are Lsand dependent, can be
derived for a given Qflow using the equations:

Q flow 8 / 13 ksand H sand Lsand / HLD

Fo, clay Fi,sand, no _ seep Fo,sand, no _ seep Q w,sand, no _ seep Wsub,steel Wsub, clay /

1 1 / HLD Fi,sand, no _ seep Q w,sand, no _ seep A plug Lsand water ]
(9.71)

ubase Q flow / 8 / 13 water ksandHsand / HLD (9.72)

iplug
min 1 1 / HLD / Lsand water ubase ,1 (9.73)
150 Installation, retrieval and removal

utop ubase Fi,clay Wsub,clay / A plug (9.74)

Fi,sand Fi,sand,no _ seep 1 iplug (9.75)

Qw,sand Q w,sand,no _ seep 1 iplug (9.76)

If the aquifer flow rate Qflow is negative, then there is no water pocket, and
the no-seepage values should be used, namely:

Q flow = 0 (9.77)

ubase 0 (9.78)

iplug = 0 (9.79)

utop Fi,clay Fo,clay Fi,sand,no _ seep Fo,sand,no _ seep Q w,sand,no _ seep
Wsub,steel / A plug (9.80)

Fi,sand = Fi,sand,no _ seep (9.81)

Q w,sand = Q w,sand,no _ seep (9.82)

where, noting that subscript “no_seep” means “with suction assistance but
no seepage/fluid flow”:

αHLD = (1 + 2.5 Hsand Lsand/D2)

ΔHpocket = water pocket height (= clay plug heave ∆Hheave)
Δutop = water (under) pressure at clay plug top
Δubase = water (under) pressure in water pocket
γsub,clay = clay plug submerged unit weight
γwater = (sea) water unit weight
Aplug = soil plug area
D = caisson diameter
Fi,clay = caisson inner wall friction – clay
Fo,clay = caisson outer wall friction – clay
Fi,sand = caisson inner wall friction – sand, fluid flow
Fi,sand,no_seep = caisson inner wall friction – sand, no seepage
Fo,sand = caisson outer wall friction – sand, fluid flow
Installation, retrieval and removal 151

Fo,sand,no_seep = caisson outer wall friction – sand, no seepage

Hclay = clay layer and clay plug thickness
Hsand = sand aquifer thickness
iplug = sand plug (vertical) average seepage gradient
kf,clay, kf,sand = DNV empirical coefficients – unit skin friction resistance in
clay and sand
kp,clay, kp,sand = DNV empirical coefficients – unit end-bearing resistance in
clay and sand
ksand = sand aquifer permeability
Lsand = caisson tip penetration into sand aquifer
Nk = empirical CPT cone factor
NLsand = number of penetration depths Lsand (= 0 through Hsand)
penrate,sand,avg = caisson average penetration rate into sand aquifer
Qbase = sand plug flowrate (into water pocket)
Qflow = sand aquifer flowrate (into base of sand plug)
Qflow,avg = average Qflow during aquifer penetration
Qsuction = suction pump flowrate
Qw,sand = caisson wall tip resistance – sand, fluid flow
Qw,sand,no_seep = caisson wall tip resistance – sand, no seepage
Rno_seep = soil resistance without suction assistance
Rsuction = soil resistance with suction assistance
tpen,Lsand = time taken to penetrate sand aquifer to Lsand
Vwater = water plug volume at time tpen,Lsand
Wsub,steel = caisson submerged weight (including ballast)
Wsub,clay = clay plug submerged unit weight = γsub,clay Aplug Hclay

Steps to find ΔHpocket are as follows:

1. Perform steady state axisymmetric fluid flow sums to obtain Qflow −

Δubase and Qflow − iplug relationships for varying Lsand/Hsand ratios (0–1);
see Figure 9.20. Use these relationships for the first two equations in
step 2 following.
2. Set up six equations and solve for 6 unknowns (see pseudo code that
follows).
3. Repeat step 2 for selected Lsand values to obtain Qflow variation.
4. Hence, obtain average flowrate Qflow,avg whilst the caisson penetrates
the sand aquifer.
5. Knowing the average caisson penetration rate penrate,sand,avg and Aplug,
values of Vwater (water pocket volume), the required ΔHpocket value can
be computed. See the water pocket model example that follows.
152 Installation, retrieval and removal

The corresponding pseudo code for the water pocket model is:

program water_pocket
initialise and evaluate constants; assume DNV CPT qc model
for resistances
F*,clay: skin friction resistances by integrating qc from 0 to
Hclay
Fi,clay = π Di kf,clay ∫ qc,z dz and Fo,clay = π D kf,clay ∫ qc,z dz
penrate,sand,avg: caisson average penetration rate
Aplug = π Di2/4
penrate,sand,avg = Qsuction /Aplug
Wsub,clay = Aplug Hclay γsub,clay
NLsand = 21
zero matrix ANS (size NLsand rows by 10 + 4 = 14 columns)
k = 0
loop tip penetration Lsand from 0 to Hsand in k equal steps:
k = k + 1
Lsand = max(Lsand, 0.001 m)
FI,sand,no_seep and FI,sand,no_seep: skin friction resistances by
integrating qc from Hclay to Hclay + Lsand
Qw,sand,no_seep: end-bearing resistance using qc at depth (Hclay
+ Lsand)
Fi,sand,no_seep = π Di kf,sand ∫ qc,z dz and Fo,sand,no_seep = π D kf,sand
∫ qc,z dz
Qw,sand,no_seep = Awall kp,sand qc,z
αHLD = (1 + 2.5 Hsand Lsand/D2)
Qflow, Δubase, iplug, Fi,sand, Qw,sand and Δutop using Equations (9.71)
through (9.76)
Rno_seep = Fi,clay + Fo,clay + Fi,sand,no_seep+ Fo,sand,no_seep+
Qw,sand,no_seep
Fo,sand = Fo,sand,no_seep
Rsuction = Fi,clay + Fo,clay + Fi,sand+ Fo,sand+ Qw,sand
store in kth row of matrix ANS:
Lsand, Lsand /Hsand, Rno_seep and Rsuction in columns 1 through 4
Qflow, iplug, Δutop and Δubase in columns 5, 6, 9 and 10
Fi,sand and Qw,sand in columns 7 and 8
end loop on tip penetration Lsand
k = 0
loop tip penetration Lsand from 0 to Hsand in k equal steps:
k = k + 1
Qflow,avg = ∑ Qflow (i = 1 to k)
Qflow,avg = Qflow,avg /k
Lsand = ANS(k,1)
tpen,Lsand = Lsand/penrate,sand,avg
Installation, retrieval and removal 153

Vwater = tpen,Lsand Qflow,avg

ΔHpocket = Vwater/Aplug
store in kth row of matrix ANS:
Qflow,avg, tpen,Lsand, Vwater, ΔHpocket in columns 11–14
end loop on tip penetration Lsand
exit with matrix ANS

matrix ANS (NLsand rows, 14 columns) contents:

col# 01: L sand caisson tip penetration into sand aquifer [m]
col# 02: L sand / H sand penetration/aquifer height ratio [–]
col# 03: R no_seep soil resistance without suction assistance [MN]
col# 04: R suction soil resistance with suction assistance [MN]
col# 05: Q flow sand aquifer flowrate (into base of sand plug) [m 3 /s]
col# 06: i plug sand plug (vertical) average seepage gradient [–]
col# 07: F i,sand caisson inner wall friction – sand, suction [MN]
col# 08: Q w,sand caisson wall tip resistance – sand, suction [MN]
col# 09: Δ u,top water (under) pressure at clay plug top [kPa]
col# 10: Δ u,base water (under) pressure in water pocket [kPa]
col# 11: Q flow,avg average flowrate Q whilst in aquifer at L sand [m 3 /s]
col# 12: t pen,Lsand time taken to penetrate sand aquifer to L sand [s]
col# 13: V water water pocket volume at time t pen,Lsand [m 3 ]
col# 14: ΔH pocket water pocket height (= clay plug heave [m]
∆H heave )

Plot col#14 (x, ΔHpocket) versus col# 01 (y, Lsand) water pocket height (=
clay plug heave) – penetration.

9.11.9.1 Example – water pocket model

The following example is for competent soil. DNV CPT High Estimate
kp,sand, kf,sand sand data, a high sand permeability k, and a typical suction
pump flowrate Qsuction (sufficient to obtain ≈ 1 mm/s penetration rate) are
used to assess high water pocket height values. Input data for a typical
North Sea type problem (basically 4 m of 70 kPa clay, underlain by sand qc
≈ 20 MPa) are:

Pile: D = 10 m, WT = D/200 = 50 mm, Wsub,steel = 6000 kN

(Di = 9.90 m, As,i = π Di = 31.10 m, As,o = π D = 31.41 m, Ap = Awall
= 1.565 m2, Aplug = 77.0 m2)
Soil: clay Hclay = 4 m, qc,clay = 1 + 0.1 z [MPa, m], γsub = 9 kN/m3
sand Hsand = 6 m, qc,sand= 20 + 0.5 (z - Hclay) [MPa, m], ksand = 1e-4 m/s
Pile – Soil – Fluid:
γwater = 10 kN/m3
kp,clay, kf,clay = 0.6, 0.05 and kp,sand, kf,sand = 0.6, 0.003
Qsuction = 83.3e-3 m3/s (300 m3/hour)
154 Installation, retrieval and removal

Sand permeability can range between ksand = 1e-4 m/s and ksand = 1e-6 m/s;
HE taken.
The sand layer has been analysed at nLsand = 21 equally spaced depth
intervals.

Pseudo code constants are:

Fi,clay = 7.46 MN and Fo,clay = 7.54 MN

Wsub,clay = 2.77 MN
penrate,sand,avg = 1.08 mm/s

Halfway through the sand layer, at Lsand = 3 m, we have

Fi,sand,no_seep = 5.81 MN and Fo,sand,no_seep = 5.87 MN

Qw,sand,no_seep = 20.2 MN
αHLD = 1.45
Qflow = 7.08e-3 m3/s Equation (9.71)
Δubase = 88.6 kPa Equation (9.72)
iplug = 0.916 Equation (9.66)
Δutop = 222 kPa Equation (9.68)
Fi,sand = 0.488 MN Equation (9.69)
Qw,sand = 1.69 MN Equation (9.70)

and

Fo,sand = 5.87 MN
Rno_seep = Fi,clay + Fo,clay + Fi,sand,no_seep + Fo,sand,no_seep + Qw,sand,no_seep = 46.8 MN
Rsuction = Fi,clay + Fo,clay + Fi,sand+ Fo,sand + Qw,sand = 23.1 MN
Qflow,avg = 6.43e-3 m3/s
tpen,Lsand = 2270 s
Vwater = 17.95 m3
ΔHpocket = 0.23 m

Figure 9.21 graphs the results. Because of the model complexity, the basic
components are also plotted to gain confidence before discussing the key
plot (ΔHpocket versus Lsand) on Figure 9.21b.
The two right hand underpressure versus penetration plots on Figure 9.21a
show that:

• self-weight penetration is just under 2 m into the 70 kPa clay. This is

sufficient to start the suction assisted penetration stage.
• underpressure at the base of the clay is just over 100 kPa.
• within the underlying sand, a water pocket is created at the start of
penetration, and this is the reason why the blue line (with pocket)
plots below the red/thick line (no pocket). Underpressures increase by
Installation, retrieval and removal 155

around 100 kPa to just under 200 kPa when the tip penetrates the sand,
increasing to 222 kPa halfway through (Lsand = 3 m) and around 260
kPa at the base of the sand aquifer (Lsand = 6 m). These magnitudes are
normally reasonable for structural engineers. The red/thick lines denote
results had no water pocket been formed. In this case, underpressures
are significantly higher (around 350, 500 and 700 kPa respectively), for
which shell buckling and/or top plate design would be of concern.

On Figure 9.21b, it can be seen that:

• aquifer flow rates, just over 20 m3/h for the first 3 m of sand penetra-
tion, are creating the water pocket. They are less than 10% of Qsuction
(300 m3/h).

Figure 9.21 W
ater pocket example. Foundation D = 10 m, WT = 40 mm. Soil
profile: 4 m of 100 kPa clay over 20 MPa sand, k = 1e-4m/s, Q suction =
250 m 3 /hour: (a) soil resistances and underpressures versus
penetration.
156 Installation, retrieval and removal

Figure 9.21 C
ontinued: Water pocket example. Foundation D = 10 m, WT = 40
mm. Soil profile: 4 m of 100 kPa clay over 20 MPa sand, k = 1e-4m/s,
Q suction = 250 m 3 /hour: (b) water pocket flowrate and height versus
penetration and time.

• a water pocket is created immediately, at the start of penetration

into the sand. The water pocket height ΔHpocket increases almost lin-
early with penetration: at 3 m penetration, ΔHpocket = 0.232 m and
the final ΔHpocket = 0.503 m.
• applying Qsuction (300 m3/h), 3 m and 6 m penetration into the sand
take ≈ 0.77 and 1.53 hours respectively.

Finally, the four plots on Figure 9.21c show that

• because there is a water pocket, iplug, the sand plug seepage gradient is
non-zero and increases with increasing penetration. This is caused by
increasing underpressure. A critical state “liquefaction” (iplug = 1) is
present from just over 5 m penetration into the sand.
• both inner friction and wall tip components decrease due to seepage,
and are zero below 5 m.
• the differential water pressure (Δutop - Δubase) is ≈ 135 kPa and con-
stant. This is as expected – when (Δutop − Δubase) Aplug is the force mov-
ing the clay plug upwards.
Installation, retrieval and removal 157

Figure 9.21 C
ontinued: Water pocket example. Foundation D = 10 m, WT = 40
mm. Soil profile: 4 m of 100 kPa clay over 20 MPa sand, k = 1e-4m/s,
Q suction = 250 m 3 /hour:(c) water pocket flowrate, sand plug gradient,
sand resistances and underpressures versus penetration.

Geotechnical engineers should make sensitivity analyses to familiarise them-

selves with complex models. By changing original parameters one by one,
their effect on ΔHpocket could be found, and results are given in Table 9.9.
It is seen that ΔHpocket (which is the same as ∆Hheave) is very sensitive to
Qsuction and ksand. ΔHpocket is inversely proportional to Qsuction and directly
proportional to ksand. Minimum ΔHpocket occurs when caisson penetration
rate is maximised (i.e. get through the sand as fast as possible). This implies
that Qsuction should be as large as possible in “clay over sand” soil profiles.
These ΔHpocket values were at 6 m penetration into the sand, i.e. a caisson
embedded length L = 10 m (L/D = 1.1). In practice, because of possible liq-
uefaction from ≈ 4 m into the sand, a lower L value, probably around 8 m
(giving L/D = 0.8) would be studied in more detail. This would also give less
onerous clay plug heave (≈ 0.3 m) and underpressures (≈ 200 kPa).
Table 9.9 also shows that ΔHpocket is moderately sensitive to caisson geom-
etry (D < 7.5 m are unlikely to be effective) and to the sand layer thickness
Hsand (aquifer inflow Qflow is proportional to transmissivity T = ksand Hsand).
158 Installation, retrieval and removal

Table 9.9 Water Pocket Model – Sensitivity Analyses.

Revised/
Revised original
Original and revised ΔH pocket ΔH pocket
Parameter(s) parameter value(s) [m] [–] Remarks

Q suction 300 600 [m3/h] 0.251 0.5 Original: 1 mm/s

penetration rate
k sand 1e-4 1e-5 [m/s] 0.05 0.1 Original: HE
k sand 1e-4 1e-6 [m/s] 5.1e-3 0.01 Original: HE
k p,clay 0.6 0.4 [–] 0.273 0.543 Original: HE
k f,clay 0.05 0.03 [–] Revised: BE
k p,sand 0.6 0.6 [–]
k f,sand 0.003 0.001 [–]
D 10 15 [m] 0.791 1.573 WT = D/200,
W sub,steel α D 2
WT 50 75 [mm] no liquefaction
in sand
W sub,steel 6000 13500 [kN]
D 10 7.5 [m] 0.345 0.686 WT = D/200,
Wsub,steel α D2
WT 50 37.5 [mm] liquefaction from 1
m in sand
W sub,steel 6000 3375 [kN]
H clay 4 6 [m] 0.560 1.113 liquefaction from
2.5 m in sand
H clay 4 3 [m] 0.478 0.951 no liquefaction
in sand
H sand 6 9 [m] 0.838 1.665 liquefaction from
4 m in sand
H sand 6 3 [m] 0.197 0.391 no liquefaction
in sand
Notes:
BE: Best Estimate, HE: High Estimate
Water pocket height ΔHpocket = clay plug heave ∆Hheave
Original ΔHpocket = 0.503 m, Hsand = 6 m, i.e. ΔHpocket/Hsand = 8%
ΔHpocket is ≈ proportional to tip penetration in sand; Hpocket values are when tip reaches the base of
the sand layer.

9.11.10 Reverse punch-through failure

The reverse punch-through failure mode is illustrated in Figure 9.22 for a cais-
son being installed into clay over sand. Unlike the water pocket model, the clay
layer extends below the caisson tip. There is a possibility of reverse punch-
through (uplift of the complete clay plug) occurring, even though the caisson
tip level is within the clay. Conventional punch-through is studied in detail for
“clay over sand” profiles below spudcans (ISO 19905-1:2016 and Figure 2.1a).
Installation, retrieval and removal 159

Figure 9.22 R
everse punch-through failure model: Forces acting on the clay plug.
No reverse-end-bearing (Cotter, 2009).

The model assumes zero reverse end-bearing (REB), and no water pocket
development, at the clay/sand interface. There are two equilibrium equa-
tions: forces on the caisson steel (during installation) and forces on the com-
plete clay plug (reverse punch-through). Note that Wsub,steel assists penetration
but Wsub,clay,RPT opposes reverse punch-through. From these equations, the
corresponding underpressures Δu can be computed. Reverse punch-through
failure is possible when ΔuRPT > Δuinstall. The four equations are:
uinstall A plug Wsub,steel Fi,clay,adhesion Fo,clay,adhesion Q w,clay (9.83)

uRPT A plug Fi,clay,adhesion Fi,clay,shear Wsub,clay,RPT (9.84)

uinstall Fi,clay,adhesion Fo,clay,adhesion Q w,clay Wsub,steel / A plug (9.85)

uRPT Fi,clay,adhesion Fi,clay,shear Wsub,clay,RPT / A plug (9.86)

where
σvo = total soil vertical stress at depth z = (γsub,clay + γwater) z
γsub,clay = clay submerged unit weight
γwater = (sea) water unit weight
Aplug = soil plug area
Ap = caisson wall tip area
As,i = caisson inner perimeter
As,o = caisson outer perimeter
Fi,clay,adhesion = caisson inner wall friction – clay adhesion above tip
Fo,clay,adhesion = caisson outer wall friction – clay adhesion above tip
Fi,clay,shear = caisson inner wall friction – clay shear below tip
Hclay = clay layer and clay plug thickness
kf,clay = DNV empirical coefficient – unit skin friction resistance in clay
kp,clay = DNV empirical coefficient – unit end-bearing resistance in clay
160 Installation, retrieval and removal

Lclay = caisson tip penetration into clay layer

Nk = empirical CPT cone factor
Qw,clay = caisson wall tip resistance – clay
qc,clay,z = CPT cone tip resistance at depth z
Rinstall = soil penetration resistance
RRPT = soil reverse punch-through resistance
su,z = clay equivalent undrained soil shear strength at depth z
Wsub,clay,RPT = punched through clay plug submerged weight =
γsub,clay Aplug Hclay
Wsub,steel = caisson submerged weight (including ballast)
Δuinstall = under pressure for caisson installation
ΔuRPT = under pressure causing reverse punch-through

The corresponding yield function can be derived from Cotter Equation

2.6.12 and is
yfRPT Y grad1 intercept1 X (9.87)

where
grad1 = α π (Hclay/D) + π Nc,strip (WT/D)
intercept1 = (Hclay/D) (π WT/D - π/4)
X = Wsub,steel/(γsub,clay D3)
Y = su,av/(γsub,clay D)
su,av,Hclay = average su in clay layer
Nc,strip = bearing capacity factor at caisson tip

Due to assuming Δuinstall acts over π D2/4, not π Di2/4, equation 9.12.19 is
slightly approximate. Reverse punch-through will not occur if yfRPT < 0.

9.11.10.1 Example – reverse punch-through

The following example is for competent soil. DNV CPT qc data are used to
assess underpressures Δuinstall and ΔuRPT. The input data are consistent with
those used for the water pocket example. The data are for a typical North
Sea type problem (basically 4 m of 70 kPa clay over sand) and are:

Soil: clay Hclay = 4 m, qc,clay,z = 1 + 0.1 z [MPa, m], γ = 19 kN/m3,

γsub = 9 kN/m3, Nk = 17
(equivalent su,z = (qc,clay,z – σvo)/Nk ≈ 60 + 5 z [kPa, m])
Pile: D = 10 m, WT = 50 mm, Wsub,steel = 6000 kN
(Di = 9.9 m, As,i = π Di = 31.1 m, As,o = π D = 31.4 m, Ap = 1.563 m2,
Aplug = 76.977 m2)
Pile –Soil:
γwater = 10 kN/m3
Installation, retrieval and removal 161

kp,clay, kf,clay = 0.6, 0.05 (High Estimates)

The calculations to find and compare underpressures Δuinstall and ΔuRPT are
straightforward. At Lclay = 2 m (half way through the clay), we have:

qc,clay,z = 1000 + 0.1 Lclay = 1200 kPa (at 1 m depth)

Fi,clay,adhesion = As,I Lclay kf,clay qc,clay,z = 3421 kN
Fo,clay,adhesion = As,o Lclay kf,clay qc,clay,z = 3456 kN
su,z = (q.c,clay,z − σvo) /Nk = 73.1 kPa (at {Hclay − Lclay} = 3 m depth)
Fi,clay,shear = As,I (Hclay − Lclay) su,z = 4548 kN
qc,clay,z = 1000 + 0.1 Lclay = 1200 kPa (at 2 m depth)
Qw,clay = Ap kp,clay qc,clay,z = 1125 kN
Wsub,clay,RPT = γsub Aplug Hclay = 2771 kN
Rinstall = Fi,clay,adhesion+ Fo,clay,adhesion + Qw,clay = 8002 kN
RRPT = Fi,clay,adhesion+ Fi,clay,shear = 7969 kN
Δuinstall = (Rinstall − Wsub,steel) /Aplug = 26.0 kPa
ΔuRPT = (RRPT + Wsub,clay,RPT) /Aplug = 139 kPa

Since Δuinstall < ΔuRPT, reverse punch-through is not occurring at 2 m

penetration.
Figure 9.23 graphs resistance and underpressure versus penetration
for both High Estimate and Best Estimate kp, kf coefficients. The red/thin
and blue/thick lines are for reverse punch-through and installation
respectively.
Figure 9.23a corresponds to the High Estimate hand calculation, and the
left-hand plot (resistances versus penetration) shows the following:
Since Qw,clay is small, installation resistance is almost zero at seafloor.
However reverse punch-through resistance is essentially constant, around 10
MN. This is as expected – unit friction values in the adhesion and shear zones
are not dissimilar, and the punched-through plug weight remains constant.
At ≈ 2 m penetration, reverse punch-through resistance equals installa-
tion. However, this does not imply that reverse punch-through is occurring
from this depth onwards – this is due to steel and plug submerged weights
assisting and opposing failure respectively.
The corresponding Figure 9.23a right hand plot (underpressures versus
penetration) reveals the following:
Underpressures become equal at 3.98 m depth, very close to the base of
the clay at 4 m. If Hclay had been 3 m, then Δuinstall < ΔuRPT at base of clay
Hclay (i.e. no reverse punch-through failure at all would occur). If Hclay = 5 m
(was 4 m), then Δuinstall = ΔuRPT at 4.65 m depth, again very close to the base
of the clay at 5 m.
Underpressure at the base of the clay is just over 100 kPa.
SWP, self-weight penetration, is just under 1.5 m into the 70 kPa clay.
This is sufficient to start the suction assisted penetration stage.
Figure 9.23b plots the corresponding results using Best Estimate kp, kf
coefficients. Since these are lower than the High Estimate values:
162 Installation, retrieval and removal

(a)

(b)

Figure 9.23 R
everse punch-through example. Foundation D = 10 m, WT = 50 mm,
W sub = 6 MN. Soil profile: 4 m of ≈ 70 kPa clay over sand. Soil resistances
and underpressures versus penetration: (a) High Estimate k p, k f = 0.6,
0.05. Reverse punch-through from 3.98 m and SWP = 1.47 m, (b) Low
Estimate k p, k f = 0.4, 0.03. No reverse punch-through. SWP = 2.48 m.

Both resistances decrease, and SWP increases by 1 m to 2.48 m.

Both Δuinstall and ΔuRP drop.
There is no possibility of reverse punch-through. The same conclusion
would apply had Hclay, the clay layer thickness, been either 3 m or 5 m.
The High Estimate and Best Estimate yfRPT values were −0.01 and −0.39
respectively.

9.11.10.2 Commentary
Based on these results, the probability of reverse punch-through occurring is
extremely small. The model excluded water pocket type fluid flow effects in
the sand. If they had been included, then increases would occur in the down-
wards-acting force on the punched-through plug base (currently assumed to
Installation, retrieval and removal 163

be zero) and hence also increase the underpressure ΔuRPT. This would make
reverse punch-through even more unlikely.
The reverse punch-through example presented here was for competent
clay. Had weak NC clay been studied, then (as with using lower kp, kf coef-
ficients in the example) the increased SWP values would decrease Δuinstall and
the likelihood of reverse punch-through. This exercise is left to the interested
reader.

9.12 INSTALLATION IN (WEAK) ROCK

9.12.1 Impact driving
Driving is not the most common way to install a foundation in a full massive
of weak rock. Traditionally, weak rock concerns have been encountered
mainly for jacket pile or anchor pile installation in cemented carbonate sand
(mainly Persian Gulf and Australian projects) or in chalk (CIRIA, 2002).
Case studies related to pile driving in other types of weak rock refer mainly
to driving through limited thicknesses of few meters (e.g. Puech et al., 1990);
see Figure 9.24.

Figure 9.24 Impact driving in rock (Puech et al., 1990).

164 Installation, retrieval and removal

The effect of impact driving on the rock mass will depend on rock char-
acteristics (e.g. strength, porosity, mineralogy and in-situ fractures), but also
on the foundation geometry. Stevens et al. (1982) considered that driving in
rock will severely fracture the rock layers and reduce the rock to granular
material. The shaft friction used in soil resistance to driving predictions
(SRD) is therefore defined using sand parameters and end-bearing propor-
tional to the rock unconfined compressive strength (UCS).
In chalk, the above approach proposed by Stevens et al. (1982) proved to
be inappropriate (Dührkop et al., 2017 and Wood et al., 2015) for low to
medium density chalk. A zone of remoulded material is observed around the
pile annulus when driving open-ended piles in these types of chalk. SRD is
linked to the grade and density of the chalk and needs to account for “fric-
tion fatigue” (unit skin friction decreases with increasing penetration) and
set-up effects. Figure 9.25 shows evidence of remoulding and fracturing
around a pile after driving in chalk.
In practice, the ground will need to accommodate the volume of steel pen-
etrating. At the start of driving, the easiest way will be to use the available
void above the ground, leading to ground uplift around the foundation and

Figure 9.25 C
halk-pile interface model – St. Nicolas pile tests, D = 0.76 m,WT = 44.5 mm
(Muir Wood et al., 2015).
Installation, retrieval and removal 165

creating wedges of fractured material. When the pile starts to penetrate, the
ground will accommodate the steel penetration through a combination of:

• Rock crushing around the pile wall. Depending on the rock type, the
crushed material will be either in a pasty or remoulded state (e.g. for
grade C and D chalk) or as a granular material.
• Radial fractures up to a few foundation diameters from the founda-
tion edge. In the presence of natural fractures, blocks will be pushed
radially and fractures will tend to close.

Parameters such as rock porosity, geological strength index (GSI) or pres-

ence of underlying weaker layers are not considered in usual empirical
methods to derive soil resistance to driving. These parameters will, however,
play a fundamental role in the way the ground will accommodate the foun-
dation penetration.
Driving installation of intermediate foundations in weak rock will not differ
much from classical flexible pile installation. It should however be noted that
for large diameter foundations, steel integrity will need to be ensured by either
the use of thick plates (driving shoe) or internal stiffeners (see Figure 9.26).
Depending on the foundation diameter and risk of buckling (see Section
9.13.5), the bottom plate thickness can be in the range of 70 to 130 mm.

9.12.1.1 Pile driving refusal

There are no clear criteria to define drivable or undrivable conditions. The
experience shows that calcareous weak rock or cemented sand can be driven
as long as a cone penetrometer can penetrate. High-density chalk will tend
to behave as limestone and would not necessarily be easy to drive through.
API RP 2A (API, 2000) provides some guidance on the maximum blow
count not to exceed to ensure pile integrity. Other contractor specifications
go a bit further and include lower criteria in terms of maximum percentage
of yield stress to not overcome in rock. Criteria limiting the reflection ratio
measured by pile driving monitoring are also used.
It should be noted that chalk can experience significant increase of shaft
resistance with time (i.e. set-up). Driving interruption of few hours can
increase the soil resistance to driving and therefore create a refusal condition
by increasing the blow count and reflection ratio.

9.12.1.2 Risk of buckling
Open-ended piles are vulnerable to tip distortion. An initial imperfection
from the theoretical pure cylindrical geometry can progressively grow with
increasing pile penetration when the stiffness of the surrounding soil exceeds
the elastic pile stiffness (Figure 9.27). Care is required during fabrication,
transportation and installation to avoid creation of an initial imperfection
166 Installation, retrieval and removal

Figure 9.26 I chthys pile tip stiffening arrangement, D/WT ratio over 110 (Erbrich
et al., 2017).

that can lead to buckling. Aldridge et al. (2005) present equations, and both
Erbrich et al. (2010) and Aldridge et al. (2005) describe conditions for prop-
agation of an initial imperfection during pile penetration (i.e. extrusion
buckling). The propagation of a small initial deformation requires that the
soil be both strong enough and stiff enough to deform the pile further.
For monopile or large diameter anchor piles, the ratio between the diam-
eter and the pile tip thickness (i.e. D/WT ratio) is becoming very high com-
pared to classical flexible piles (see Figure 9.28). Figure 9.13 shows that for
D/t greater than 40, buckling was observed. Risk of insufficient hoop stiff-
ness or heterogeneous conditions across the diameter are significantly
greater for stubby large diameter foundation. The ring stiffness is inversely
proportional to cube of D/t ratio.
Installation, retrieval and removal 167

Figure 9.27 P
lastic hinge mechanism assumed to cause local tip buckling (Aldridge
et al., 2005).

Figure 9.28 P
ile tip geometries, oil and gas full blue dots and monopiles empty
dots, red circles correspond to recorded tip damage (Randolph, 2018).
168 Installation, retrieval and removal

Particular care needs to be accounted for in the case of heterogenous con-

ditions around the pile tip (e.g. inclined hard layer, boulders or flint beds).
Presence of strong nodules in a softer matrix can initiate localised damage if
not easily pushed aside or split (Holeyman et al., 2015). The buckling risk
can be partially mitigated by considering a driving shoe (i.e. increase pile tip
thickness up to 130 mm) end adapted driving procedure.
Classical monitoring with strain gauges and accelerometer placed at the
top of the pile are not sufficient to issue real-time warning at the start of
buckling initialisation.
A careful reading of monitored information could indicate, post-installa-
tion, where the buckling started. A modification of the impedance from the
starting damaged point should be required to match the measured signal.
Proper real-time monitoring would require sensors to be placed at the toe of
the pile, but the installation of sensors would risk weakening the pile (e.g.
creation of a groove to install fibre-optic strain gauges). The use of fibre
optics without a groove (i.e. only fixed with glue) is under development and
seems to be a promising approach.
Special procedures (i.e. decision tree) can be applied in homogenous or
well-known stratigraphy to interpret in real time the soil resistance to driv-
ing (SRD) evolution and take appropriate mitigation measures (e.g. drilling
inside the foundation down to the target penetration depth). However, in
case of heterogeneous material, it will be difficult to differentiate if the
observed increase in blow count (i.e. increase of tip resistance) is due to a
modification of the rock strength or an increase of the tip area induced by
deformation. In practice, the occurrence of buckling is often discovered only
when drilling inside the pile is required.

9.12.2 Drilling
9.12.2.1 Drive drill drive
Drilling can be used as a contingency measure when driving is expected to
be difficult (as the drilling tool is generally already mobilised together with
the hammer) or in case of unexpected pile refusal (e.g. presence of boulder).
After the end of driving, the drilling head is inserted inside the pile and drill-
ing starts from the seafloor. Drilling can either stop after the particularly
difficult layer (or boulder) or be continued to the pile target depth. In the
latter case, hole stability needs to be guaranteed. The driving process is then
continued until reaching the target depth or the next difficult layer.
Drive Drill Drive (DDD) operations often require under-reaming capabil-
ity, i.e. the possibility for the drilling head to enlarge its diameter once it
reaches the pile tip. Drilling more than the pile diameter removes resistance
beneath the tip in a particularly hard layer (or boulder).
The switch of tool between the hammer and drilling tool is time consum-
ing. Drive Drill Drive (DDD) installation is often scheduled only if one pass
Installation, retrieval and removal 169

of drilling is expected to be sufficient. If several passes are expected, a drilled

and grouted process might become more cost effective.

9.12.2.2 Drilled and grouted

When driving is not an option or when the risk of pile buckling is too high,
an alternative can be to drill and grout the foundation. The drilled hole
diameter is equal to the foundation diameter plus two times the grout annu-
lus. The foundation is then lowered in the hole and the void between the
foundation and the ground is grouted from the bottom to the top.
Open drilling can start from the seabed when the rock mass is stable
enough. In most cases, upper sediments or degraded rock are present at the
seabed, and drilling needs to be advanced using temporary (i.e. retractable)
or sacrificial (i.e. stay in place after drilling operation) casing supporting
unstable ground until the grouting operation.
Figure 9.29 shows an example of a drilled and grouted installation
sequence using a casing to equalise the overburden layer and support the top
drilling equipment.

Figure 9.29 D
rilled and grouted installation sequence with sacrificial casing
(Courtesy of Fugro).
170 Installation, retrieval and removal

The main risks during drilling are:

• Hole collapse: meaning retrogressive failure of the wall resulting in

damage to, or loss of drilling equipment and requiring relocation of
the foundation;
• Over-break: meaning local instability of the hole wall, likely resulting
in an increase in hole diameter and potentially a deterioration of the
rock condition local to the hole wall;
• Decrease of hole diameter: preventing the foundation being lowered
and any temporary casing being removed (e.g. swelling rock);
• Hole inclination out of tolerances;
• Drilling rate much lower than expected (e.g. in presence of large boul-
der or plastic clay if drilling tool is not designed for the conditions
encountered).

Before drilling, the rock mass is in a state of equilibrium. After drilling,

instability can occur if:

• the post drilling redistributed stress-state exceeds rock strength.

• a network of closely spaced discontinuities could lead to the sliding of
blocks. The presence of at least three discontinuities passing through
the drilled hole is necessary for a block to slide (two discontinuities
could be sufficient, but the risk of getting an unstable block would be
quite low). The risk of sliding will depend on the orientation and prop-
erties of joints and the rock properties.
• rock-chemical interaction leads the rock to weaken or swell. Drilling
fluid needs to be carefully selected, particularly in shale.

In all cases, the time during which the hole is kept unfilled needs to be
minimised.
Unfortunately, there are not many references applicable for shallow hole
stability for foundations (i.e. low drilling fluid pressure due to relatively low
depth relative to e.g. well-boring). Most of the published work covers well-
bore drilling, tunnelling or mining activities. For cylindrical geometry, zero
drainage boundary conditions and homogenous rock mass, hole stress con-
ditions can be derived based on elastic solutions.
For weak rock where discontinuities are present, the complexity of the
stress distribution and geometry of the discontinuities requires FE (Finite
Element) or DE (Discrete Element) methods. It should, however, be noted
that FE and DE methods will not capture the chemical interactions and time
dependency.
The main difficulty remains to get a detailed enough idea of the geometry
of nearby discontinuities. Offshore, the available information is often lim-
ited to the logging of joints and fractures on 100 mm samples from a single
borehole. It might be necessary, where faults are prevalent, to perform
Installation, retrieval and removal 171

several boreholes per foundation, enlarge the depth of boreholes (to com-
pare geological markers) and/or consider downhole logging such as natural
gamma-ray, televiewer and calliper. These tools can provide valuable infor-
mation on the nature, frequency and inclination of discontinuities (crossing
the borehole).
During grouting operations, it will be necessary to avoid a grout flow
inside the foundation. This can be done by either creating a grout plug at the
bottom of the foundation before filling the annulus, through a physical
boundary (e.g. internal membrane) or through a mechanical closure (e.g.
using external packers).
In the presence of fractured rock or karstic voids, the volume of grout to
be injected can largely exceed the annulus volume (volume between the
foundation and the rock). Grouting operations offshore remain a delicate
operation that is difficult to control.

9.12.3 Vibratory
Pile installation using vibratory means would not be considered if the stra-
tigraphy includes cemented material or hard (even small) layers.

9.13 PRESENTATION OF INSTALLATION ASSESSMENT

Presentation of suction foundation installation analyses usually includes:

• soil layering and parameter values

• foundation geometry
• preload values
• Best Estimate and High Estimate resistances versus depth
• Best Estimate and High Estimate underpresssures versus depth, includ-
ing allowable/limiting pressures for base failure and liquefaction.

Figure 9.10 has given an example for an anchor pile in LOC clay. Figure 9.17
presents a typical example for caisson installation in sand. Note the
following:

• Double plot (resistance R and underpressure derivative Δu versus

depth z): the latter plot cannot be used if the preload alters. Entering
the R – z profile with Wsteel value gives self-weight penetration. Entering
the Δu - z plot with the target penetration depth gives the maximum
underpressure (usually).
• For clays, the Δu - z plot also shows Δuall, the allowable suction pres-
sure to prevent base failure.
• For sands, the Δu - z plot also shows Δuliq, the underpressure causing
sand plug liquefaction.
172 Installation, retrieval and removal

• Other allowable pressures may be necessary, for example water depth

and structural criteria
• High Estimate results have been omitted for clarity – normally both
Best Estimate and High Estimate results are provided.

9.14 RETRIEVAL AND REMOVAL RESISTANCE

ASSESSMENTS

9.14.1 Suction foundations
Procedures for retrieval and removal are similar to those for installation
(Section 9.2 on penetration resistance assessment), except that:

• set-up may have occurred in normally consolidated clay (Section

9.10.3) and also in sand (Section 9.11.3)
• overpressure (downwards seepage in soil plug) increases inner skin
friction in sand
• end-bearing resistance at tip level is neglected, provided that there is
no passive suction below the top plate and the vent is open
• resistance (both friction and end-bearing) may occur above internal
friction breakers
• dead weight counteracts overpressure
• special techniques (e.g. Broughton et al., 2002) are needed to apply
overpressure beneath underbase grouted foundations.

9.14.2 OWT monopiles
Driven monopiles cannot be removed in their entirety – their pull-out capac-
ity (resistance plus submerged weight) is significant. For example, a 5 m
diameter 30 m embedded length monopile pull-out capacity could easily
exceed 60 MN, depending on the exact soil profile and other factors. There
are, however, on-going developments to investigate removal of such piles,
see e.g. Balder et al. (2020). Marine legislation generally stipulates that piles
be cut off at a certain depth (typically between 3 m and 5 m) below seafloor.
This is generally achieved by explosives, mechanical or abrasive cutters.
Jetting may be necessary to remove the soil beforehand. Pile cutting may be
either internal or external; usually the former approach is selected, in view
of the soil volumes to be excavated. Grout overflow or other seafloor
obstructions may hinder access.
Drilled-and-grouted monopiles cannot be removed entirely either. For the
decommissioning phase, the requirement is generally to cut the monopile 1
m below seabed and to fill it with rock. Only a few offshore windparks have
already been decommissioned. Few relatively small drilled and grouted
Installation, retrieval and removal 173

monopiles (around 3.5 m diameter) were successfully cut 1 m below seabed

using an internal jetting system (e.g. Blyth monopiles).
Vibratory driven monopiles can also be cut off using the same driven pile
methods. Theoretically, vibratory methods could also be used, using a higher
energy vibratory hammer due to set-up in sand. However, at the time of
writing, this method has not been applied offshore.

9.15 CLOSURE

This lengthy chapter has covered intermediate foundation installation. Most

sections are devoted to suction-assisted foundations. This is because they are
more challenging and complex than monopile foundations, for which design
methods are well understood and accepted by industry.
At the time of writing there is no general agreement/consensus as to which
design methods/models are appropriate for suction foundations. In terms of
increasing maturity, installation and penetration resistance methods are
more mature for clays (especially NC clay), than sands, followed by mixed
sand/clay profiles. Reasonably accurate models are given for these profiles.
However, installation in clay is less problematic than in sand – the sand
model accuracy needs to be improved (e.g. Figure 9.17) and field data sug-
gest that icrit may be exceeded (e.g. Figure 9.15). In addition, little is known
about “set-up” factors for skin friction in sand. Hence, more research, cali-
brated using high-quality field data, is required on installation in sand,
where plug liquefaction and plug heave are critical issues. At the moment,
not exceeding the limiting icrit = γsub/γw value gives L/D values between 1.0
and 1.5, the latter usually associated with low relative density layers near
seafloor. Geotechnical engineers prefer to embed foundations as deep as
possible: soils get more competent with depth, and cost savings accrue.
Hopefully, the observational method, together with the research described
here, will eventually lead to achieving higher intermediate foundation
embedments L/D in sand.
Pseudo code for three simple programs (and verification input/output
data) have been provided: install_clay, install_sand and water_pocket for
clay over sand. These simple building blocks can be modified and combined
for solving real-world suction assisted foundation installation problems.
Chapter 10

In-place resistance

10.1 INTRODUCTION

This chapter covers in-place resistance for both support type and anchor
pile type intermediate foundations.

10.2 LOADING CONDITIONS AND SOIL RESPONSE

In routine design, T (torsion) loads are small compared to significant HM

(lateral) loading. In addition, offshore extreme loading conditions are such
that soils – even sands – can usually be modelled with an undrained response
(Chapter 5).

10.3 IN-PLACE FAILURE MODES

Figure 2.2 (Section 2.2, on in-place resistance modes) showed six failure
modes for intermediate foundations under combined VHM loads in und-
rained soil. All assume co-planar HM (horizontal load and moment) and
negligible T (torsion). Figure 2.3 showed the corresponding failure modes
for drained soil response. All modes are suitable for preliminary/routine
design calculations, but possibly need to be complemented by FEA for final
design.
Figures 2.2 and 2.3 are both for combined VHM loading. Figure 10.1
(Section 10.5 on maximum axial resistance) gives the corresponding failure
models for pure axial (pull-out) V load.
Resistance envelope models for VHM(T) loading are discussed further in
Section 10.10.

175
176 In-place resistance

10.4 TENSION CRACKS AND GAPPING

Tension cracks (also known as gaps) can occur behind the trailing edge
(active side) of foundations under lateral load. This has been realised as a
potential problem for sheet pile walls for many years. The crack depth
depends upon many factors, including the lateral load magnitude and the
in-situ stress profile.
The occurrence of either tension cracking or gapping reduces VHMT
resistance. For offshore foundations, if a tension crack develops to skirt tip
level, then in-place undrained resistance capacity at tip level (including
reverse end-bearing) may be severely affected by loss of suction in fine-
grained soil. In addition, in the zone above the tension crack, a small reduc-
tion in bulkhead net resistance (passive minus active) occurs.
Tension cracking is not generally considered for normally consolidated
clay subjected to short-term loads (Randolph and Gourvenec, 2011).
Jeanjean et al. (2006) found no evidence of gapping in more competent OC
clay (su ≈ 25 kPa at 9 m depth). Gapping is, however, expected to be non-
reversible in cemented formations or rock. In the case of drilled and grouted
piles, the gap will most likely occur between the steel pile and the grout
annulus. This gap should normally be limited to the upper 1–2 meters due
to the limited displacements.
For intermediate anchor foundations in normally consolidated clay, ten-
sion cracking is not usually considered. This is because locating the lug
slightly below optimum depth causes restoring loads and decreases the risk
of forward rotation. However, tension cracking and gapping may need con-
sideration for intermediate support foundations under overturning loads
and/or in stiff clays.
Mitigation measures for foundations include a flexible mat (a “gap
arrester” or “mud liner”) around the perimeter (e.g. Keaveny et al., 1994;
Mana et al., 2013). This inhibits a tension crack by preventing water supply,
and hence suction loss, at the soil–foundation interface.

10.5 MAXIMUM AXIAL RESISTANCE - SUCTION PILES

10.5.1 Failure modes for maximum axial tensile

resistance
This section discusses possible axial failure modes, and selection of the most
appropriate mode. The following sections discuss appropriate unit friction
and unit end-bearing parameter values for undrained “clay”, undrained
“sand” and drained “sand” respectively. Skin friction and end-bearing con-
tributions to axial resistance are assumed to be uncoupled.
Distinctions should be made between:
V loads (tensile or compressive, covered in Section 3.2 on intermediate
foundation types) soil response (undrained or drained, covered in Section
5.2 on drained-undrained-partially drained).
In-place resistance 177

Figure 10.1 F ailure models for maximum axial tensile resistance, V max , in und-
rained soil. Reverse end-bearing is included only in the “plugged”
model (Senders and Kay, 2002).

Figure 10.1 shows the three possible failure modes – “plugged”, “coring”
and “leaking” – under tensile V loads. Identification of these modes have
been based on previous work. Fuglsang and Steensen-Bach (1991) reported
results of 1 g model and 40 g centrifuge pull-out tests on 65 mm and 80 mm
diameter L/D = 2 caissons in kaolin clay. To derive end-bearing and inner/
outer skin friction components, they used three model caisson types:
“plugged” (Fo + Qbase), “coring” (Fo + Fi) and “coring without the clay plug”
(Fo). They verified that there was indeed reverse end-bearing failure. In sev-
eral 1 g “plugged” tests, but none of the centrifuge tests, brittle tensile failure
was observed – the clay at the caisson tip abruptly failed in tension. This was
possibly the “leaking” mode – the tests were carried out underwater.
Randolph and House (2002) considered the same three “plugged”, “cor-
ing” and “leaking” failure modes as Fuglsang and Steensen-Bach (1991),
using instead the term “base-vented” when a hydraulic short circuit (i.e.
“leaking”) prevents underpressure at the caisson base. For the “coring”
mode, reverse end-bearing resistance (Qwall) under the caisson wall is
included. Although admittedly small, it is optimistic to include this compo-
nent. Senders and Kay (2002) presented the same three models, but with
component Qwall omitted from the “coring” mode. The following paragraphs
give fuller details and update the aforementioned work.

10.5.1.1 Tensile V loads (anchor foundations) in

“undrained” soil
If the top plate vent is permanently sealed, then reverse end-bearing (“pas-
sive suction”) can be relied upon for a limited time period, such as peak
pull-out loads during a storm. In this case, axial resistance (Vmax) equals
outer friction and reverse end-bearing. This is the “plugged” failure mode.
Note that the tensile resistance at foundation tip level is given by classical
178 In-place resistance

bearing capacity theory as “Ao su Nc – Ao q”, q being the surcharge. However,

the term “Ao q” is equal and opposite to the soil plug self-weight, Wplug.
Hence, a Wplug term does not appear in the expression for Vmax, the axial
resistance at pile head level. Hence, “plugged” Vmax and Vload (axial pull-out
load) are given by the equations:

Vmax Fo Qbase Fo N c,circlesu,av,L Abase (10.1)

Vload Fo Qbase Wsteel (10.2)

If reverse end-bearing (“passive suction”) cannot be relied upon, then either

the “coring” or “leaking” failure mode can occur. Both modes imply some
form of hydraulic leak and zero pressure under the top plate. Examples of
design situations that readily spring to mind include:

(a) A top plate seal cannot be guaranteed to function perfectly during

(say) 30 years sustained tension load, and, due to leakage, the pressure
under the top plate will eventually reduce to (almost) zero.
(b) A hydraulic leak occurs at (or just below) foundation tip level due to
a seam or layer of permeable material in the vicinity of the foundation
tip which behaves essentially hydrostatically (no excess water pres-
sure) and provides zero reverse end-bearing. Note that seams, between
20 mm and 76 mm thick according to ASTM 2487 (ASTM 2011) and
ASTM 2488 (ASTM 2009), may be undetected by CPTu. Moreover, in
the horizontal direction, a 20 mm thick fine sand seam (permeability
k = 10−4 m/s) has 40 times the transmissivity of a 5 m thick clay layer
(k = 10−8 m/s).
(c) A hydraulic leak occurs at foundation tip level due to a gap extending
from seafloor to tip level caused by excessive foundation rotation of
an anchor pile with a wrongly located pad-eye. Again, this results in
zero end-bearing resistance.
(d) Another possible scenario is due to anchor chain trenching, which
may adversely influence reverse end-bearing resistance, usually relied
upon for FPSO mooring anchors. To help ensure backwards rotation,
design typically includes a deeper anchor lug level than is ‘optimal’,
and this further reduces the distance between lug and tip. If the chain
trench extends to lug level, then there is an increased risk of either (i)
a reduced drainage path length H (H = tip to trench base, was H = tip
to seafloor), or (ii) a hydraulic leak occurring between lug and tip (i.e.
seawater breaking tip suction). Both situations could result in hydro-
static water pressure at foundation tip level during the design life-
time). Hence, as a consequence of trenching, reverse end-bearing was
cautiously omitted by Alderlieste et al. (2016).
In-place resistance 179

In the “coring” failure model, there is no reverse end-bearing under the tip
wall, and axial resistance (Vmax) equals outer friction (Fo) and inner friction
(Fi). The corresponding Vload (axial pull-out load) equals Fo + Fi + Wsteel. In
the “leaking” failure model, Vmax simply equals outer friction (Fo), and Vload
(axial pull-out load) equals Fo + Wsteel + Wplug.
Hence, “coring” Vmax and Vload are given by the equations:

Vmax Fo Fi (10.3)

Vload Fo Fi Wsteel (10.4)

and “leaking” or “plugged” by the equations:

Vmax = Fo (10.5)

Vload Fo Wsteel Wplug (10.6)

Design axial tensile resistance is usually taken to be one of the following:

if reverse end-bearing (“passive suction”) can be relied upon:

“plugged” Vmax resistance
if reverse end-bearing (“passive suction”) cannot be relied upon:
if there is no possibility of the “leaking” failure model occurring:
“coring” Vmax resistance
if the “leaking” failure model can occur:
minimum of the “coring” and “leaking” Vmax resistances.

Note that for a normally consolidated clay (su/σ'vo≈ 0.25) with clay-steel
inner adhesion factor αi ≈ 0.6, “leaking” is more critical than “coring”
failure when embedment ratio L/D ≈ 3. That is, if “leaking” failure can
occur, then “coring” resistance is probably governing for short, stubby
foundations, whereas “leaking” is more likely to be critical for longer
foundations.
Finally, for routine FEA, it is very difficult to model pull-out using either
a “plugged” failure model with a user-assigned Nc value (i.e. reduced REB;
Clukey and Morrison, 1993; Clukey et al., 1995), or a “leaking” failure
model. In the remaining “plugged” and “coring” cases, design practice
should be to obtain and interpret the sum of soil forces along the inner and
outer shaft areas (i.e. mobilised friction Fo and Fi) and over the base area (i.e.
mobilised bearing resistance Qbase). If “plugged” failure is occurring, then
one would expect Vload ≈ Fo + Qbase and Vload ≈ Fo + Fi if “coring”. For the
“plugged” condition, the back-figured Nc,circle value should be compared
with the Skempton (1951) value for circular foundations. In particular,
Nc,circle is expected to be in excess of 9 for L/D = 2.5, for the reasons given in
Section 10.5.2.
180 In-place resistance

10.5.1.2 Compressive V loads (support foundations) in

“undrained” soil
For compressive V loads, the “leaking” failure mode does not occur and the
number of failure modes reduces from three to two (i.e. “plugged” and “cor-
ing”). Wsteel acts in the opposite sense: it reduces the design compressive
load. In addition, Vwall (wall tip end-bearing resistance) is added to the “cor-
ing” failure mode. It is seen that these aspects are similar to compressive
offshore pipe pile design procedure.
However, design axial compressive resistance depends on whether or not
there is “top plate bearing” (i.e. soil plug or grout is in contact with the top
plate underside):
if “top plate bearing” can be relied upon:
“plugged” Vmax resistance
if “top plate bearing” cannot be relied upon:
minimum of the “plugged” and “coring” Vmax resistances.
This procedure is unlike conventional pipe pile design. The difference in
procedure derives from from the differing soil plug lengths Lplug (both in clay
and sand) for pipe piles and intermediate foundations. Pipe pile Lplug values
are usually somewhat less than L (pile embedded length); this is mainly due
to inertial effects during installation by impact driving (Smith and Chow,
1982; Chow, 1981). Typically, for both sands and clays, Lplug ≈ 0.9 L. Hence,
there is generally no “top plate bearing” for pipe piles. On the other hand,
intermediate foundations installed by suction in clay causes soil plug heave
to match part of the soil volume displaced by the foundation steel. Section
9.10.2 (on plug heave in clay) suggests that for clays, Lplug ≈ 1.02 L. Hence,
for suction-installed intermediate foundations in clay, there is generally “top
plate bearing”, and the “plugged” failure model is appropriate.
Table 10.1 summarises the Vmax and Vload equations.

Table 10.1 Maximum Axial Resistance and Load in Undrained Soil.

Failure Maximum axial Maximum axial

Axial load model resistance V max load V load
Tension “plugged” F o + N c s u,av,tip A base F o + N c s u,av,tip A base + W steel
(anchor “coring” Fo + Fi F o + F i + W steel
foundation) “leaking” Fi F i + W steel + W plug
Compression “plugged” F o + N c s u,av,tip A base F o + N c s u,av,tip A base − W steel
(support “coring” F o + F i + N c s u,tip A wall F o + F i + N c s u,tip A wall − W steel
foundation) “leaking” Not applicable Not applicable

Note(s): See Figure 10.1 for tensile failure models

In-place resistance 181

10.5.1.3 Example – maximum axial resistance

To illustrate the equations and to examine the sensitivity to the three failure
modes, embedded anchor lengths L have been obtained for a constant diam-
eter D = 5 m and 7500 kN factored tensile pull-out load. Results in Table
10.2 used the following data:

su = 2 + 1.5 z [kPa, m], γsub,soil = 6 kN/m3

D = 5 m, D/WTside = 200, D/WTtop = 100, Lplug = L, γsub,steel = 67 kN/m3
αi = αo= 0.65
αD,su = 0.5
λL,Bw = 1.0.

Table 10.2 shows that (for the data presented here):

• there is a progressive L increase – embedment ratios L/D are ≈ 3, 4 and

4.5 for the “plugged”, “coring” and “leaking” failure models
respectively
• “plugged” is the most efficient – soil and drainage conditions permitting
Bw,caisson for “plugged” is 0.75 of the corresponding “leaking” value
• all L/D ratios are < 5, and hence amenable to suction installation.

Table 10.2 F oundation Embedded Length L for 5 m Diameter Anchor, 7500 kN

Factored Tensile V load in Undrained Soil and Varying Failure Modes.

Failure Mode

Parameter “plugged” “coring” “leaking”

L [m] 15.64 20.01 22.33
L/D [–] 3.13 4.00 4.47
Resistances
Fo [kN] 2192 3474 4274
Fi [kN] Not applicable 3439 Not applicable
Q base [kN] 4830 Not applicable Not applicable
V max [kN] 7022 6914 4274
Loads
B w,caisson [kN] 475 589 651
B w,plug [kN] Not applicable Not applicable 2578
B w = B w,caisson [kN] 475 589 3229
+ B w,plug
resistances + loads
V load = V max [kN] 7500 7504 7503
+ Bw

Note(s): H = M = 0.
182 In-place resistance

Note that these L values are solely for pure V load. Placing the lug level at
optimum depth below seafloor for resisting combined load gives signifi-
cantly smaller L values – between 12 m and 14 m (see the Example in
Section 10.7).

10.5.2 Undrained (“clay”) soil response

Skin friction resistance and Pile-soil adhesion coefficient α:

• α values are generally lower than recommended by ISO 19901-4 (ISO

2016a), Section 8.1.3 (Skin friction and end-bearing in clay soils) for
driven pipe piles. In NC clay, α is around 15–20% lower.
• α self-weight penetration zone may be higher than in the suction instal-
lation zone.
• Outer skin friction may be higher than and inner skin friction (e.g.
Jeanjean, 2006; Jeanjean et al., 2006).
• Cautious values may be αo = αi = 0.65.
• Reduced αi (e.g. αi = 0.3) may be appropriate above internal stiffeners
in clay due to remoulding (αi = 0 if extruded above stiffener).

End-bearing resistance and Bearing capacity factor Nc:

For constant shear strength with depth, the equations are:

N c,strip 7.5 for wall end bearing strip (Skempton, 1951) (10.7)

N c,circle min 6 1 0.2 L / D , 9 (Skempton, 1951)) (10.8)

N c,circle min 6.2 1 0.34 arctan L / D , 9 (DNV, 2018) (10.9)

Linearly increasing shear strength with depth:

• For (reverse) end-bearing over the full foundation base area, a reason-
able approach is to take Nc = Nc,circle and su averaged over (say) 0.25 D
below tip level. This approach is consistent with that used for jack-up
spudcan leg penetration assessments, see. ISO 19905-1:2016 (ISO
2016b), Section A.9.3.2.2. Other organisations may have different
opinions regarding averaging depth etc.
• For (reverse) end-bearing over the foundation wall tip area, take Nc =
Nc,strip and su at tip level.
• Other methods are available for shallow foundations; see Section
3.2.1 on in-place resistance modes.
In-place resistance 183

Note that it is cautious to assume that bearing capacity factor Nc,circle has a
maximum value of 9 for closed-ended and plugged open-ended piles. See
Section 10.5.5 on axial myths.

10.5.3 Undrained (“sand”) soil response

Assessing tensile resistance of undrained sand under transient load is
extremely difficult. Considerations include:

• A key aspect is suction development under top plate.

• equivalent su is very sensitive to sand dilatancy parameter D (e.g.
DNV, 1992)
• suction is limited by pore fluid cavitation (e.g. McManus and Davis,
1997)
• Additional considerations include:
• pore water pressure build-up and dissipation
• reliability of soil parameters (e.g. k, cv and S-N curve)
• stress redistribution (especially outer skin friction) from skirt to
plug base
• sand plug loosening due to installation
• sand plug re-compaction (if any) during previous environmental
(storm) loading
• amount of compressive preload applied prior to (single) tensile
wave load
• V resistance reduction due to HM loads (coupling).

Some examples of evidence for tensile resistance in sand include:

• Field: Draupner jacket, New Year’s storm, 1 January 1995, piezometer

measurements indicated that the bucket foundation could resist short
duration loads without any problems (Svanø et al., 1997, Hansteen
et al., 2003).
• Laboratory: Pressure chamber (and onshore field) testing of suction
caissons under tensile loading in sand (Kelly et al., 2003, 2004, 2006a,
2006b, Houlsby et al., 2005).
• Numerical: Transient cyclic loading of suction caissons in sand
(Cerfontaine et al., 2016).

Commentary:

• These challenges tend to disappear with increasing embedment ratio

(think pipe piles) BUT suction installation when L/D > 1 is difficult.
• This is very different to the clay–sand separation challenge described
in Section 9.12.2 on the water pocket model.
184 In-place resistance

• If concerns exist about high tensile axial resistance, then permanent

passive suction (Section 12.5) could be considered.

10.5.4 Drained (“sand”) soil response

10.5.4.1 Compressive loads
If the failure mode is “plugged”, then Vmax is given by:

Vmax subL2 / 2o D subL N q Abase (10.10)

And, if it is a “coring” failure mode, then we have:

Vmax subL2 / 2 o i D subL N q A wall (10.11)

Table 9.6 gives values of ISO 19902 pile bearing capacity factor Nq.

10.5.4.2 Tensile loads
Assuming a “coring” failure mode, and ignoring reverse end-bearing below
the wall tip, then Vmax is given by:

Vmax subL2 / 2 o i D (10.12)

Note that friction resistance for intermediate foundations using ISO 19901-4
(2016) for driven piles is considered to be optimistic. The reasons for
decreased friction include:

(a) less densification (lower energy and soil volume disturbance)

(b) inner friction reduced due to sand plug loosening (suction
installation)
(c) outer friction reduced due to installation (vibratory installation only).

10.5.5 Axial myths
10.5.5.1 Bearing capacity factor Nc
It is a geotechnical myth that Nc,circle has a maximum value of 9 (Skempton,
1951) for closed-ended and plugged open-ended piles. This value is cau-
tious. The DNV (2005) equation is merely a “better” (non-linear) fit to
the original Skempton (1951) Figure 2 curve. Skempton’s work was based
on theoretical considerations, small-scale laboratory, and two field tests
on screw cylinders/screw piles. All these boundary conditions are not
In-place resistance 185

Table 10.3 Bearing Capacity Factor N c, circle – Numerical Analyses.

N c , circle Reference

Skempton Griffiths Salgado Edwards

Boundary (1951) (1982) et al. et al.
Condition Figure 2 Figure 4 (2004) (2005)
Feature Figure 11 Figure 3 Remarks
Open/closed Open Closed Closed Closed Griffiths
hole (1982)
Figure 5
(open
hole) ≈
same as
Skempton
(1951)
curve
Smooth/rough Not Smooth ? Smooth
sides applicable
Load/ (field/ Displ Pressure Displ
displacement laboratory/
control base theoretical)
Rough/smooth Smooth Rough Rough
base

necessarily the same as for (plugged) pile/intermediate foundation tip end-

bearing. Table 10.3 lists more recent numerical analyses using both Limit
Equilibrium and Finite Element Analysis. For a given tip penetration, the
Nc,circle value is influenced (increases) with increasing pile-soil adhesion coef-
ficient αo, tip roughness and plug rigidity. Figure 10.2 compares the Nc,circle
– L/D curves for these various conditions.
For comparison, 2D axisymmetric FEA of CPT penetration, a 60° cone,
gives Nk ≈ 12.5 (van den Berg, 1994). In addition, the more recent Figure 10.2
Nc,circle data are valid only for “closed-hole” situations (i.e. closed-ended and
plugged open-ended). Lower values should be used for “open-hole” situa-
tions. Incidentally, this explains the lower Nc,spudcan values (Houlsby and
Martin, 2003 and ISO 19905-1:2016, Annex E1): an “open hole” situation
exists above the spudcan to permit backflow.
Physical confirmation for Nc,circle exceeding 9 has also been obtained from
a number of centrifuge tests on “plugged” caissons. Fuglsang and Steensen-
Bach (1991) performed both 1 g laboratory tests and 40 g centrifuge tests
on L/D = 2 caissons to verify the presence of reverse end-bearing failure.
They obtained average Nc,circle values of 9.2 (weak 10 kPa clay) and 8.1
(stronger 20 kPa clay), i.e. both in good agreement with the Skempton
(1951) value of 8.5 for circular foundations. Clukey and Morrison (1993)
186 In-place resistance

Figure 10.2 N c,circle versus embedment ratio showing dependency on various

boundary conditions (see Table 10.3).

reported an Nc,circle factor of approximately 11 for caissons with L/D = 2 in

NC clay. This factor may be optimistic (see e.g. Randolph and House, 2002).
Randolph and House (2002) found Nc,circle to be between 14 and 15 for
short term pull-out load applied to L/D = 4 caissons in NC clay. Chen and
Randolph (2007b) studied caissons with prototype D = 4 m, L = 12 m (i.e.
an embedment ratio L/D = 3) in kaolin clay. For NC clay, inferred combina-
tions of average αo and Nc,circle values were (0.90, 11.0) or (0.76, 12), with αo
> 1 for Nc,circle of 10 or less. For LOC clay, the corresponding average αo and
Nc,circle combinations were (0.73, 12) or (0.95, 10.5), with αo = 1.1 for Nc,circle
= 9.5. Jeanjean et al. (2006) performed double-walled pull-out tests, allow-
ing separation of Fo, Fi and Qbase components, on D = 1.88 m, L = 11.25 m
(L/D = 6) sealed caissons in NC kaolin clay. Combinations of αo and Nc,circle
values of (0.85, 9.0) or (0.6, 12) were possible.
At the time of writing, no confirmation has been found from high-quality
instrumented large-scale axial pile load tests in clay. This probably stems
from the fact that piles in clay (usually L/D > 10) are friction piles, for which
parameter αo is far more important than Nc,circle. For intermediate “plugged”
foundations (0.5 ≤ L/D ≤ 10), both αo and Nc,circle are relevant.

10.5.5.2 Skin friction and end-bearing

Another geotechnical myth is that skin friction and end-bearing contribu-
tions to pile axial capacity are uncoupled (e.g. Section A.8.1.4.2.1, ISO
19901-4:2016). In fact, skin friction and end-bearing are weakly coupled.
In-place resistance 187

Load spread of skin friction increases surcharge (and hence end-bearing) at

foundation tip level. This coupling is more apparent in sands than clays,
especially in the other direction – end-bearing pressure increases normal
stress/skin friction at/near tip level. This fact is reflected in CPT qc-based
methods for assessing pile capacity in sand.

10.6 MAXIMUM LATERAL RESISTANCE – SUPPORT

FOUNDATIONS

VHM loads are at seafloor (see Figure 3.1a). Care should be taken in distin-
guishing between free and fixed head resistances H: “Fixed head” Hmax
occurs when the foundation cannot rotate (rotation θxz = 0) and V = 0. “Free
head” Ho is when V and M are both zero. Figure 10.3a shows their positions
in the MH plane. Normalised Hmax and Ho values versus foundation embed-
ment ratio for various undrained shear strength profiles (Figure 10.3b) are
given in Figure 10.3c and 10.3d. For embedment ratios L/D > 3, reasonable
Np values are 10.5 and around 3 for most clay profiles, and hence the ratio
Hmax/Ho ≈ 3.5.

(a)

Figure 10.3 M
aximum lateral resistance in clay. Normalised resistance factor N p
= H/(LDs u,av,L ) (a) MH plane, lateral resistances “fixed head” H max and
“free head” H o (Kay and Palix, 2010).
188 In-place resistance

(b)

Figure 10.3 C
ontinued: Maximum lateral resistance in clay. Normalised resis-
tance factor N p = H/(LDs u,av,L ) (b) Soil profiles.

Randolph et al. (1998) used AGSPANC (Advanced Geomechanics, 2002),

an upper bound solution software, which can give optimistic values at low
embedment ratios (Andersen et al., 2005). Kay (2015) used FEA and ignored
internal scooping (if any) at low embedment ratios, where it can be seen that
Hmax and Ho values tend to converge. This is a consequence of lateral resis-
tance being dominated by base shear, not lateral pressure. At higher embed-
ment ratios, Hmax values for uniform soil become constant, and both solutions
agree with one another and also with the Randolph and Houlsby (1984)
solution for a laterally loaded disc.
In-place resistance 189

(c)

(d)

Figure 10.3 C
ontinued: Maximum lateral resistance in clay. Normalised resis-
tance factor N p = H/(LDs u,av,L ) (c) Horizontal capacity of suction
caissons (Supachawarote, 2006), N p versus L/D (α o = 0.65) (Kay,
2015).
190 In-place resistance

10.7 MAXIMUM LATERAL RESISTANCE AND LUG

POSITION (SUCTION ANCHOR PILES)

These paragraphs are about suction anchor pile lateral capacity. These are
essentially rigid caissons. Analyses of laterally loaded piles are not discussed.
The terms anchor chain and anchor line are synonymous.
Lateral performance, particularly moment failure of the steel at design
loads, may be governing (rather than soil failure). Providing additional steel
stiffeners at or near the lug level should prevent this. Note that additional
steel increases installation resistance.
Unless load and chain inclination is provided at lug level, a separate anal-
ysis is required of the anchor chain below seafloor. This is to assess the
decreased load and increased uplift angle, θlug, at lug level (see Figure 3.1b).
Note that competent soils, such as sand and stiff clay, cause significant
changes to the chain profile, particularly a further θlug increase. In such cases,
the lug level is usually restricted to relatively shallow depths below seafloor
in order to avoid axial pull-out. Equilibrium equations for an embedded
anchor chain have been given by Degenkamp and Dutta (1989) and
Neubecker and Randolph (1995).
If there is no restriction about lug depth, then maximum horizontal resis-
tance (Hmax) is obtained such that, in the ultimate limit state, the caisson
behaves as a “fixed head” body, translating laterally with zero rotation as
shown on Figure 8.1b. In this case, optimum lug level (on the caisson shaft) is
determined by the intersection of the line of action of the lug chain load and the
caisson centreline. In practice, the lug level is just below optimal (say, around
0.5 m for a 3 m diameter caisson). This ensures backwards rotation at failure,
reducing potential for a tensile crack to open on the caisson trailing edge.
The optimum lug level is close to the weighted average of the shear
strength profile, and the general expression for any shear strength profile is:
L L

ez
0

su, z z dz / su, z dz
0
(10.13)

Values of ez are approximately 2/3 L for normally consolidated clay 0.5 L

for constant shear strength clays, and 0.75 L and 0.25 L for the extreme
“stepped” and “crust” shear strength profiles shown on Figure 10.3b. Small
corrections are necessary for small embedment ratios to account for base
shear (Kay and Palix, 2011; Kay, 2015).
“Free head” and “fixed head” lateral resistances Hmax and Ho are given by
the following:

H max = N p,fixed L Dsu,av,L

(10.14)
Ho = N p,freeL Dsu,av,L
In-place resistance 191

where
L = caisson embedded length
D = caisson outer diameter
su,av,L = average su,z between caisson head (seafloor) and caisson tip (L)

and

Np,fixed = “fixed head” lateral bearing capacity factor

Np,free = “free head” lateral bearing capacity factor

Lateral bearing capacity factors Np vary with foundation embedment

ratio L/D; see Figure 10.3. For L/D > 1.5, reasonable “standard” Np values
are 10.5 and around 3 for most clay profiles for fixed and free head respec-
tively. At shallower depths, they both increase and vary rapidly, due to the
increasing base shear contribution.
From the generic equations H = Np L D su,av,L, and Np,fixed/Np,free = 3.5, it is
easily shown that, for NC clay, if the lug is moved upwards from optimum
to (near) seafloor for a given diameter D, the embedded length L has to be
multiplied by a factor √3.5 ≈ 1.9 in order to obtain the same H value.
Similarly, if the embedment ratio (L/D) is kept constant, then the L multi-
plier is 3√3.5 ≈ 1.5. Neither is geotechnically efficient. This underlines the
necessity (if possible) to place the lug at/near optimum depth.
If the lug level is at seafloor, then lateral resistance is reduced by about a
factor 3.5 because the caisson is able to rotate. The corresponding lateral
resistance is the “free head” resistance (Ho).
If the anchor lug level is close to (just below) optimum, then a modest
resistance decrease is warranted; see Figure 10.4. If more precision is
required, then consider using one of the Table 3.1 analysis methods. Note
incidentally that Figure 10.4 recovers the aforementioned “standard” Np
values (i.e. 10.5 and 3).
If anchor lug level is close to seafloor but not at optimum, then this is an
intermediate case. Choices include

(a) assume “free” head resistance – slightly cautious

(b) use one of the Table 3.1 analysis tools.

Example – Fixed Head Pile. Use the same data as used previously in Vmax
example (Section 10.5.1), basically a 5 m diameter anchor pile to withstand
7500 kN factored load at lug level (Tlug) acting at 30° (θlug). Additional data
included:

   lug offset at pile outer radius (D/2, i.e. 2.5 m)

   “fixed” head pile – i.e. lug at optimum depth zlug.opt,fix
192 In-place resistance

Figure 10.4 Effect of anchor lug level on lateral resistance (Randolph and
Gourvenec, 2011).

Program CAISSON_VHM (Kay, 2015) results for the three “plugged”,

“coring” and “leaking” failure models are given in Table 10.4. This table
shows that (for the aforementioned data):

• there is a progressive L increase – L values are ≈ 12.13 m, 14.10 m and

14.14 m for the “plugged”, “coring” and “leaking” failure models
respectively
• “plugged” is the most efficient – soil and drainage conditions
permitting
• increasing L by around 2 m covers the possibility of “coring” and
“leaking”
• all L/D ratios are around 2.5, and hence amenable to suction
installation
• zlug,opt,fix/L (lug level depths : embedded length ratios) around 0.5–0.6
are less than the “2/3 L” value; this is due to the small embedment
depth ratios (L/D ≈ 2.5), where base shear plays a role.
• since they are “stubby” foundations (L/D ≈ 2.5), smaller D values may
well prove to be a more economical solution.

Note that these L values are for combined VHM load. Putting the lug level
at seafloor, under pure V load, gives significantly higher L values – between
15.6 m and 22.3 m (see the Example in Section 10.5).
Figure 10.5 shows the corresponding three Vlug-Hlug envelopes, together
with the actual Tlug load components, namely Vlug = Tlug sin(θlug) and
Hlug = Tlug cos(θlug). The three shapes are dissimilar to V-Hmax diagrams – see
In-place resistance 193

Table 10.4 F oundation Embedded Lengths L for 5 m Diameter Anchor. 7500 kN

Factored Tensile T lug acting at θ lug = 30° in Undrained Soil.

Parameter V max Failure Mode

“plugged” “coring” “leaking”

L opt,fix [m] 12.13 14.10 14.14
L opt,fix /D [–] 2.43 2.82 2.83
z lug,opt,fix [m] 6.76 8.11 8.14
z lug,opt,fix /L [–] 0.56 0.58 0.58
Resistances + loads
V max [kN] 5235 3601 1821
H max [kN] 6954 9442 9508
B w,caisson [kN] 383 435 436
B w,plug [kN] Not applicable Not applicable 1633
B w = B w,caisson + B w,plug [kN] 383 435 2069

Notes:
zlug,opt,fix: caisson optimum lug level depth (below seafloor)
Lopt,fix: caisson optimum embedded length (“fixed head” condition)

Figures 8.9 (ISO 19901-4) and 10.12 (Supachawarote et al., 2004). For
example, taking the “plugged” failure model, (a) Hlug first increases, and
then decreases, as θlug decreases from 30° to 0° and (b) Vlug first increases,
and then decreases, as θlug increases from 30° to 90°. The latter is beneficial:
Tlug exceeds 7500 kN should θlug exceed 30°..

Figure 10.5 V lug -H lug envelopes and design T lug load.

194 In-place resistance

The Vlug- Hlug envelope shape is the locus of points where Vlug and Hlug
loads (at constant rlug, zlug, varying θlug) intersect the “tongue”-shaped VHM
resistance envelope in 3D VHM space shown in Figure 2.2e. This does not
have the same shape as ellipsoidal V-Hmax diagrams, which are vertical slices
through the same envelope at Hmax.

10.8 MAXIMUM TORSIONAL RESISTANCE

In assessing maximum torsional resistance, due consideration should be

given during site investigation and interpretation to the possible occurrence
of discrete layers of low strength material along which displacements could
preferentially occur. Failure due to the formation of internal mechanisms
within the confined soil plug (above skirt tip level) should also be
considered.
The maximum available torsional resistance, Tmax is given by the lesser of
the “plugged” and “coring” failure models, namely:

Tmax min Tinner Touter ,Touter Tbase (10.15)

For a uniform circular foundation with no protuberances in clay, this

gives:

Tinner = torsion shear resistance of the inside of the foundation

   = αi su,av,L L π Di2/2
Touter = torsion shear resistance of the outside of the foundation
   = αo su,av,L L π D2/2
Tbase = torsion shear resistance over the full foundation base
   = su,tip π D3/12
αi = clay-steel inner adhesion factor
αo = clay-steel outer adhesion factor
su,av,L = average su between seafloor and foundation tip (L)
su,tip = soil undrained shear strength at foundation tip depth (L)
D = foundation outer diameter
Di = foundation inside diameter
L = foundation embedded length.
Exterior protuberances (if any) provide a (small) additional torsional
resistance ∆T, given by:
T N c, protsu, prot A prote x (10.16)

where
Nc,prot = bearing capacity factor
In-place resistance 195

su,prot = undrained shear strength at protuberance depth

Aprot = protuberance bearing area
ex = horizontal distance between neutral axis and lateral force resultant.

For “undrained” clay profiles, usually the “plugged” (not “coring” mode)
is critical, and hence Tmax = Touter + Tbase

10.9 TILT AND TWIST – ANCHOR PILES

These paragraphs are about intermediate anchor pile design for tilt (non-
vertical or out-of-plumb) and twist (misalignment or misorientation). Both
reduce holding capacity compared to a perfectly installed and orientated
anchor. For resistance calculations, the most unfavourable situation (i.e.
maximum tilt and maximum twist occurring at the same time) is often stud-
ied in detailed design. Typical specifications for installation tolerance are
±5.0° for pile tilt and ±7.5° for pile twist. Generally, maximum tilt (+5.0°)
reduces capacity more than maximum twist (+7.5°).
Tilt is accounted for by increasing or decreasing the chain load angle at
lug level (θlug) by the tilt angle. This approach is similar to the common
assumption for inclined piles, namely that the axial and lateral responses are
largely unaffected by small angles of inclination from the vertical (e.g.
Poulos and Davis, 1980). Adding (e.g.) 5° tilt to θlug increases the V load
component on the anchor, and subtracting 5° increases the H load compo-
nent. Usually, anchors are more sensitive to V (than H) loading, and hence
+5° tilt is more unfavourable than −5°. A non-horizontal seafloor, especially
if it is sloping down in the same direction as the mooring line(s), should be
considered. A reasonable procedure is to take the design tilt to be the greater
of slope angle and tilt tolerance angle. For example, if the seafloor slope
angle = 10°, and tilt tolerance = ±5.0°, then take design tilt = +10°.
Twist influences suction anchor foundation resistance less than tilt. This is
seen in typical tolerance values: namely < ±5.0° for tilt and < ±7.5° for twist.
Due to twist, part of the maximum available friction resistance will be used
in torsional shear when the chain pulls on the suction anchor. The basic
principles are given in Figure 10.6. Assuming the torsion load is fully applied
first, a simple calculation can be done: knowing the applied torsion load and
the maximum available resistance (on the suction anchor perimeter and
base), the amount of shear utilised can be calculated. Vector analysis then
gives the amount of shear available, and hence the reduction percentage, on
the maximum frictional resistance for axial loading. To simplify design cal-
culations, the pile-soil outer adhesion factor (αo) for axial loading can there-
fore be decreased by the reduction percentage. Typically, this αo decrease is
small – around 5% (Senders and Kay, 2002). In the worked example that
follows, the αo decrease is around 2.5% (was 0.65, now 0.635). Interested
196 In-place resistance

Figure 10.6 E ffects of twist on suction anchor holding capacity (Taiebat and
Carter, 2005).

engineers may care to repeat the example calculation for 15° (was 7.5°)
twist. This occurred at Na Kika, Gulf of Mexico, details of which were given
in Section 6.1.2. Large diameter foundations (D > 15 m, say) will probably
give αo decreases > 5%, and, if Tload/Tmax is large, Tload – VHM resistance
interaction effects (Section 10.10) may need to be considered.

10.9.1 Example – twist
To illustrate the difference twist (misalignment) makes to clay-steel outer
adhesion factor αo, consider a typical anchor pile in a NC soil profile. The
relevant data are:

Pile: D = 5 m, L = 30 m
Soil: su = 2 + 1.5 z [kPa, m], αo= 0.65
(su,av,L = 2 kPa + 1.5 kPa/m 15 m = 24.5 kPa, su,tip = 2 kPa + 1.5 kPa/m
30 m = 47 kPa)
Lug: Tlug = 3 MN, θlug = 30°
Twist: ψlug = 7.5°
The calculations to find αo* (i.e. adhesion αo allowing for torsion load Tload)
are straightforward. They are as follows:

Vlug = pile axial load due to Tlug

= Tlug sin(θlug) = 1500 kN
Hlug = pile lateral load due to Tlug
= Tlug cos(θlug) = 2598 kN
Tload = pile torsion load due to twist (see Figure 10.6a)
= Hlug D/2 sin(ψlug) = 2598 kN 2.5 m sin(7.5°) = 848 kNm
In-place resistance 197

Tmax = maximum available torsional resistance, assume “coring” failure

mode
= Touter + Tbase
= αo su,av,L L π D2/2 + su,tip π D3/12
= (0.65 × 24.5 kPa 30 m π (5 m)2/2) + (47 kPa π (5 m)3/12)
= 18,761 + 1,538 = 20,299 kNm
ηT = pile-soil shear mobilisation factor due to torsion load Tload
= Tload/Tmax = 848 kNm/20,299 kNm = 0.042
Fo = pile shaft friction resistance
= αo su,av,L L π D
= 0.65 × 24.5 kPa 30 m π 5 m = 7504 kN
ηV = pile-soil shear mobilisation factor due to axial load Vlug, ignore
REB
= Vlug/Fo = 1500 kN/7504 kN= 0.200
ηV* = pile-soil shear mobilisation factor due to combined Vlug and Tload;
see Figure 10.6b
= √ (ηV2 − ηT2) = √ (0.2002 − 0.0422) = √ (0.038) = 0.196
αo* = clay-steel outer adhesion factor allowing for torsion load Tload
= αo (ηV*/ηV) = 0.65 (0.196/0.200) = 0.635.

10.10 RESISTANCE UNDER COMBINED VHM(T) LOADS

Values of maximum vertical resistance Vmax (Section 10.5), maximum lateral

resistance Hmax (Section 10.6), and maximum torsional resistance Tmax
(Section 10.8) are not achievable under combined VHMT loads. This is
because, like shallow foundations and unlike deep (pile) foundations, cou-
pling occurs:

• Torsion loads decrease intermediate foundation VHM resistance.

• HM loads decrease intermediate foundation V and T resistance.

For routine design, these two factors are considered using numerical analy-
sis models, either based on limit equilibrium and plastic limit analysis
(Section 10.11) or a resistance envelope approach (Section 10.12). The 3D
FEA is generally not used for preliminary (and the majority of routine) geo-
technical design of intermediate foundations.
For routine design, it is considered reasonable to assume that Tload
affects only axial V (not lateral HM) resistance. Justification for this is
given by Taiebat and Carter (2005) – as shown in Figure 10.7 – Saviano
and Pisanò (2017) – shown in Figure 10.8 – and also Suroor and Hossain
(2015), not shown.
198 In-place resistance

(a)

(b)

Figure 10.7 E ffect of torsion (T) load on vertical (V) and horizontal (H) resis-
tance (Taiebat and Carter, 2005).

10.11 IN-PLACE RESISTANCE ANALYSIS METHODS

10.11.1 General
Besides VHM(T) coupling, base shear and possible tension cracking need
to be considered – this is as for shallow foundation design, but unlike pile
foundation design. Unlike shallow foundations in clay, there is not gener-
ally a need to consider gap formation below the foundation base. Analysis
methods include Limit Equilibrium, Plastic Limit and Resistance Envelope
(aka Yield Function). The latter is a comparatively recent development in
geotechnical engineering.

10.11.2 Undrained soil response (clay)

Support foundations for deepwater projects have many different mudmat and
caisson types, each with various geometries, resulting in substantial numbers
In-place resistance 199

(d)

Figure 10.8 E ffect of torsion (T) load on vertical (V) and horizontal (H) resis-
tance (Saviano and Pisanò, 2017): (a, b, c) effects of misalignment on
VH capacity envelope, (d) HT capacity diagram.

of preliminary optimisation analyses. Each analysis usually includes many

(100+) VHM load cases, from which the governing ones have to be found.
A key issue is to balance conflicting demands of speed, accuracy and cost.
Anchor foundations have fewer load cases and often less variation in
geometry/aspect ratio. Capacity is sensitive to lug level depth and chain load
angle, and key geotechnical issues are lug level and separate chain response
analyses if Tlug loads are applied below seafloor. Complications include (a)
iterating between separate caisson and chain models – which should use the
200 In-place resistance

same design soil profile – and (b) differing vertical and horizontal Factors of
Safety on lug load Tlug.
Table 10.5 compares intermediate foundation resistance software func-
tionality. Only CAISSON_VHM uses a yield function (details given in
Section 10.12), and additional functionality includes:

• caisson tilt inclusion by adding tilt angle to θlug – see Section 10.9
• axial resistance Vmax using either (a) “plugged”, “coring” or “leaking”’
failure models (Section 10.5.2) or (b) user-defined
• caisson buoyant weight Bw using either WT/D ratios or user-defined
• embedded chain profile using Degenkamp and Dutta (1989); these
chain element equations were first stated by Vivatrat et al. (1982), and
are identical to those given in Appendix A3 of DNV EP-R301 (2000)
• separate λT load factors on V and H components of lug load Tlug
• resistance modifications – see Section 10.12.7
• caisson twist (misalignment): reductions are usually small, less impor-
tant than tilt (Senders and Kay, 2002), and can be simply done by
decreasing αo – see Section 10.9.

Table 10.5 C
aisson In-place Resistance Software – Comparison Functionality
(Kay, 2015).

Software

Functionality A B C D E F
Solution method PL ULS ULS YF FE 2D FE 3D
Support foundations ⨯ ✓ ✓ ✓ ✓ ✓
Anchor foundations ✓ ⨯ ✓ ✓ ✓ ✓
Chain profile ⨯ ⨯ ✓ ✓ ⨯ ⨯
Rotational failure ✓ ✓ ✓ ✓ ✓ ✓
Zero rotation failure ✓ ⨯ ✓ ✓ ✓ ✓
Internal scoop failure ⨯ ✓ ✓ ⨯ ⨯ ✓
Tension crack ✓ ✓ ✓ ⨯ ✓ ✓
High axial load ⨯ ⨯ ✓ ✓ ✓ ✓
Reduced REB ✓ ⨯ ⨯ ✓ ⨯ ⨯
L, D optimisation ⨯ ⨯ ⨯ ✓ ⨯ ⨯
Multiple L,D ⨯ ⨯ ⨯ ✓ ⨯ ⨯
Multiple load cases ⨯ ⨯ ⨯ ✓ ⨯ ⨯
Multiple soil layers ⨯ ✓ ✓ ✓ ✓ ✓
Tilt ✓ ⨯ ✓ ✓ ✓ ✓
Twist ⨯ ⨯ ✓ ✓ ✓ ✓
Soil reactions ⨯ ⨯ ⨯ ⨯ ✓ ✓

Notes:
A: AGSPANC (AG, 2003); FALL16 (OTRC, 2008); B: CANCAP2 (Fugro, 2009); C: CAISSON (Kennedy
et al., 2013); D: CAISSON_VHM (Kay, 2015); E: BIFURC (Jostad, 1997); SPCalc (XG Geotools, 2014),
F: ABAQUS, PLAXIS 3D etc.
ULS: limit equilibrium (neither LB nor UB); PL: plastic limit analysis (usually UB);YF: yield function (based
on FE); FE: finite element analysis (slight UB), either 2D or 3D; LB: lower bound; UB: upper bound.
In-place resistance 201

10.11.3 Drained soil response (sand)

There are no VHM(T) drained resistance envelope methods for intermediate
foundations. In sand profiles, pile (rather than soil) failure occurs first. In
addition, OWT monopile in-place response (e.g. tilt) is usually of major
concern, and stubby support caissons are more likely to behave undrained
than drained under extreme storm loads. The only method is that proposed
by Broms (1964b); he used Plastic Limit Analysis (Upper Bound) for rigid
piles under lateral (HM) load. V load coupling, base shear and internal
scooping were not considered. Such situations require the use of 3D FEA.

10.12 VHM(T) RESISTANCE ENVELOPE METHODS

10.12.1 General
This section can be skipped by engineers uninterested in foundation opti-
misation. The majority of this section presents methods for undrained
soils. This is because, as noted, a generic envelope for intermediate (and
also for deeply embedded surface) foundations in drained sand under
VHM loads is not yet available.

10.12.2 Undrained soil response (“clay”)

Intermediate foundation capacity in clay may be quickly and reasonably
assessed using VHMT resistance envelope theory. If failure envelopes and
ultimate limit state (ULS) can be described by algebraic expressions, itera-
tion can be employed to identify minimum foundation area for given soil
strength and loading conditions.
For surface and shallow foundations (embedment ratio < 0.5), resistance
envelopes are presented in section ISO 19901-4:2016 Clause 7.3.5 and
Annex A.7.3.5. Because of their complexity and sensitivity to the soil su
profile, only a few results are currently available. These are for a limited
number of foundation shapes and undrained shear strength profiles, some of
which include or exclude base tension, and several of which are described by
approximating algebraic expressions enabling foundation optimisation
(Gourvenec et al., 2017).

10.12.2.1 Resistance envelopes for undrained soil response (“clay”)

Table 10.6 summarises resistance envelopes available for intermediate foun-
dations in clay. They are in chronological order, earliest first. A common
feature to all is that, since they are all FEA based, unlike ULS/plasticity
based methods (e.g. AGSPANC/FALL16/CANCAP2/HVMCAP), one is “not
getting back one’s own input”.
It can be seen that, for VHM loading, the most comprehensive work is
that given by Kay and Palix (2010, 2011). Their resistance envelope is
202 In-place resistance

Table 10.6 R
esistance Envelopes for Intermediate Foundations in Undrained
Soil (Clay).

Reference Envelope Soil s u profile L/D Remarks

Supachawarote V-H max NC 1.5, 3
et al. (2004) and 5
Taiebat and VHT Uniform 2 and 4
Carter (2005)
Kay et al. VHM NC, uniform, 0.5 - 6 CAISSON_VHM
(2010–2015) stepped and
crust
Van Dijk (2015) VHM NC, uniform 0.125 - 2 Internal scoop
and stepped
Sørlie (2013) VHMT NC 5 Project specific
Ahn et al. (2014) V-H max NC and uniform 2 - 10
Gerolymos et al. VHM Uniform 1, 2 and 3 Square caisson
(2015)
Suroor and VT NC and uniform 1, 3 and 7 V&T applied at
Hossain (2015) zlug/L = 0.7, α = 1

Note: NC, uniform, stepped and crust soil su profiles given on Figure 10.3b.

based on around 5500 quasi 3D (2D Fourier) finite element analyses for a
large number of caisson L/D values (0.5 ≤ L/D ≤ 6) and soil su profiles
(Figure 10.3b).
The resistance envelope is “tongue”-shaped (see Figure 10.9) in VHM
space, and consists of a rotated ellipse in M-H space (Figure 10.9b), and
recovers the familiar ellipsoidal shape of VH diagrams for shallow founda-
tions in V-Hmax space (Figure 10.9c). The complete envelope is described by
three simple equations, and all parameter values are smooth functions of
foundation embedment ratio (L/D) and non-dimensional soil shear strength
profile (ez,su/L). For a multi-layered soil profile, interpolation is used between
the su profiles shown in Figure 10.3b. Major assumptions included no ten-
sion crack and no internal soil scoop failure. Torsion loading, usually small,
can be accounted for by decreasing su (Senders and Kay, 2002; DNV, 2005).
Benchmarks (including Andersen et al., 2005; Randolph and Houlsby, 1984;
Taiebat and Carter, 2010), plus design optimisation examples for support
and anchor foundations, are given by Kay and Palix (2010, 2011). Van Dijk
(2015) gave resistance envelope equations for open-ended caissons, consid-
ering internal scooping.
The results can be applied to any type of rigid intermediate foundation –
either support or anchor pile. For anchor foundations, since the VHM load
reference point is on the centreline at seafloor (see Figure 3.1), and the foun-
dation is essentially rigid, chain loads and inclination angle at lug level (Tlug
and θlug) need to be transformed to seafloor VHM load using the three
equations
In-place resistance 203

(a)

(b)

(c)

Figure 10.9 Intermediate foundation resistance envelope: (a) “tongue”-shaped

VHM envelope, (b) rotated ellipse at base (the MH plane) and (c) an
ellipsoid in the V-H max plane. LRP at seafloor (Kay, 2013).
204 In-place resistance

V = Vlug (10.17)

H = H lug (10.18)

M Vlug r.lug H lug z lug (10.19)

where
Vlug = Tlug sin(θlug)
Hlug = Tlug cos(θlug)
rlug = lug radial offset from foundation centreline
zlug = lug level depth below seafloor

No anchor pile tilt (θtilt = 0) is assumed in the aforementioned equations, and

(restoring) M values are therefore slightly cautious. This is because in the MH
plane, the M load is reduced, meaning that the MH load data point lies further
away from (and inside) the resistance envelope. These second order effects are
small and are therefore usually neglected. If required, lug coordinates (rlug,zlug)
can be transformed for any θtilt (clockwise positive) using the equations:

rlug,tilt cos tilt rlug sin tilt z lug (10.20)

z lug,tilt sin tilt rlug cos tilt z lug (10.21)

10.12.2.2 Example seafloor VHM loads

Consider an anchor pile D = 5 m, with rlug = D/2 = 2.5 m and zlug = 16 m. Pile
is subjected to load Tlug = 10 MN acting at an angle θlug = 30°. Find seafloor
VHM loads for tilt angles θtilt = 0 and +5°.

For θtilt = 0, we have

V = Tlug sin(θlug) = 5000 kN

H = Tlug cos(θlug) = 8660 kN
M = −Vlug r.lug − Hlug zlug = −5000 kN × 2.5 m − 8660 × 16 m = −151,060 kNm

For θtilt = +5°, the seafloor V and H loads remain unchanged, cos(θtilt) = 0.996
and sin(θtilt) = 0.087, and the transformed lug coordinates are given by

rlug,θtilt = cos(θtilt) rlug − sin(θtilt) zlug = 0.996 × 2.5 − 0.087 × 16 = 1.096 m

(was 2.5 m)
zlug,θtilt = sin(θtilt) rlug + cos(θtilt) zlug= 0.087 × 2.5 + 0.996 × 16 = and 16.157 m
(was 16 m)
and

M = −Vlug rlug,θtilt − Hlug zlug,θtilt = −5000 kN × 1.096 m − 8660 × 16.157 m =

−145,400 kNm (was −151,060 kNm, a 4% reduction)
In-place resistance 205

10.12.3 MH ellipses
For intermediate foundations, it is convenient to work in terms of dimen-
sionless M and H values given by the equations (Figure 10.10)
H H / L Dsu,av,L (10.22)

M M / L2Dsu, av, L (10.23)

MH ellipses are defined by the following three parameters

(i) ΦMH – rotation angle

(ii) aMH – major semi-axis length
(iii) bMH – minor semi-axis length

Ad (i): ΦMH – rotation angle

M/H ≈ ez/L (1/2 for uniform su profile, 2/3 for NC su profile, 3/4 for
“stepped” su profile, see Figure 10.3
MH atan e z / L (10.24)

Ad (ii): since Hmax lies close to the semi-major axis (see Figure 10.11)

aMH cos MH / H max ; H max 10.5 (10.25)

Ad (iii): since we have an almost straight line between Ho and Mo (see

Figure 10.11)
bMH Ho sin MH ; Ho 3 (10.26)

Hence, one can obtain ΦMH, aHM and bHM values from

(a) soil su profile

(b) caisson “free head” (Ho) capacity
(c) caisson “fixed head” (Hmax) capacity

Figure 10.3c and 10.3d have compared normalised CAISSON_VHM free

and fixed head resistance values with those of Randolph et al. (1998)
obtained using AGSPANC (Advanced Geomechanics, 2002).

10.12.4 V-Hmax ellipsoids
Senders and Kay (2002), Supachawarote et al. (2004) and Kay and Palix
(2010) all found that V-Hmax resistance envelopes for caissons in clay (L/D
in the range 0.5–6.0) were ellipsoidal and could be fitted with equations of
the type
206 In-place resistance

Figure 10.10 M
H ellipses at V = 0. Summary final data points, caisson D = 5 m,
L/D = 1.5, 2, 3, 4, 5 and 6. Soil s u profiles: (a) and (b) Constant, e z,
s u /L = 1/2, (c) and (d) NC, e z,su /L = 2/3 and (e) and (f) Stepped, e z,su
/L = 3/4. (Kay and Palix, 2010)
In-place resistance 207

Figure 10.11 M
H ellipse resistance envelope. Key data points H 0 , M 0 and H max
(Kay and Palix, 2010).

aVH bVH
H max,V V
1 (10.27)
H max Vmax

Figure 10.12 shows dimensionless V-Hmax envelopes for different L/D val-
ues. As L/D increases, the less V-H interaction occurs, i.e. circles become
more square. Hence, ellipsoid curvature parameters aVH and bVH are again
functions of L/D and the su profile.
For example, for normally consolidated clay, Supachawarote et al. (2004)
gives

aVH L / D 0.5 (10.28)

bVH 4.5 L / D / 3 (10.29)

The high bVH value implies that vertical resistance is less affected by hori-
zontal load than vice versa.
208 In-place resistance

Figure 10.12 V
-H max resistance. Dimensionless envelopes for differing caisson
embedment ratios, L/D. Normally consolidated clay (Supachawarote
et al., 2004).

10.12.5 VHM envelope – equations

The complete VHM resistance envelope is given by rotated ellipses in the
MH plane (Section 10.12.3) plus ellipsoids in the V-Hmax plane
(Section 10.12.4). Factor Hmax,V/Hmax accounts for non-zero V load. The pair
of parametric equations for a rotated ellipse are:
H load L H
LDsu,av,L
max, V
H max
a MH cos t cos MH bMH sin t sin MH (10.30)

Mload L H
2
max,V aMH cos t sin MH bMH sin t cos MH (10.31)
L Dsu, av,L H max

and the ellipsoid equation is:

aVH bVH
H max, V L V
load L 1 (10.32)
H max Vmax

10.12.6 VHM envelope – yield function

The geometrical form of a rotated ellipse corresponding to the pair of para-
metric Equations (10. 30 and 10.31) is:

cos X sin Y 2 sin X cos Y 2

MH MH MH MH

a MH b MH

H (10.33)
max_ V
H max
In-place resistance 209

Eliminating (Hmax,V/Hmax) from Equations (10.30) and (10.31), the corre-

sponding yield function, equation is obtained, namely:

2
cos MH H sin MH M
fVHM
aMH
2

sin MH H cos MH M V aVH

2 bVH

1 (10.34)
bMH Vmax

The first two terms of the yield function, Equation (10.34), are independent
of V. They represent the geometrical form of a rotated ellipse in non-dimen-
sional (M*, H*) space when V = 0. The third term, which is independent of
H and M, has a non-dimensional axial load term, |V/Vmax|, and double expo-
nents. This term “shrinks” the yield surface (i.e. the M*, H*ellipses) for
non-zero axial load. For (almost) pure V load, especially when |V/Vmax| > 1,
numerical problems (imaginary numbers) with exponents are avoided by
using a simpler form of Equation (10.34), namely

fvhm V Vmax 1if V Vmax 1 (10.35)

10.12.7 Modifying lateral and Vmax resistance

Major CAISSON_VHM assumptions include (a) the caisson is fully sur-
rounded by soil (b) suction is possible at the soil-steel interface (c) there is
no failure within the soil plug (d) H and M load is co-planar.
In non-routine design, it may be necessary to consider exceptions such as:

(i) “gapping” in the active soil zone behind caisson for both support and
anchor foundations (e.g. AG, 2003; DNV, 2005)
(ii) “anchor chain trenching” – a tapered wedge-shaped trench occurring
in front of caisson due to anchor chain motions (e.g. Alderlieste et al.,
2016; Colliat et al., 2018); this cannot occur for support foundations,
as they are top loaded and have no chain
(iii) “vertical slot” above pad-eye for anchor foundations, which may
occur during installation
(iv) “reduced end-bearing” for tension loaded anchor foundations (Clukey
and Morrison, 1993; Clukey et al., 1995)
(v) “internal scoop” for support foundations at shallow L/D (Bransby
and Yun, 2009; van Dijk, 2015)
(vi) “wings” for anchor foundations (Dührkop and Grabe, 2009)
(vii) non-co-planar HM load.
210 In-place resistance

Apart from (vi), and possibly (vii), all decrease in-place resistance.
Since all geotechnical foundation models are inaccurate to a certain
degree, and the CAISSON_VHM resistance model is no exception, it is rea-
sonable to consider it to be also valid for these exceptions without a signifi-
cant decrease in model accuracy.
Hence, four reduction factors (η) are available. Figure 10.13 shows how
ηaMH, ηbMH and ηφMH modify the MH ellipse parameters aMH, bMH and φMH,
and any ηFo and ηREB combination modify Vmax. Table 10.7 gives η examples/
senses – values come from design specifications, literature or numerical
analyses. Seafloor trenching reduces Ho more than Hmax resistance, with a
smaller Vmax reduction which is mainly related to Fo (Alderlieste et al., 2016).
Because of this, and the MH ellipse shape, η order is ηaMH < ηbMH < ηφMH
< ηFo < 1. For “internal scoops” (possible only when L/D < 1.5), moment
reduction is more severe than Hmax (van Dijk, 2015). Hence η order is ηbMH
< ηaMH < 1 = ηFo = ηREB. Adding “wings” increases axial friction less than
lateral resistance; hence η order is 1 = ηREB > ηFo > ηaMH ≈ ηbMH. Note that,
since a seafloor trench increases optimum lug level (OLL) depth, whereas
wings have the opposite effect (OLL decreases), ηφMH > 1 and ηφMH < 1
respectively.
Lateral and axial resistances are modified by applying reduction factors to
auto-computed aMH, bMH and Vmax values. The shape of the yield surface (i.e.
MH ellipses and V-Hmax ellipsoids) is assumed unchanged, and the size of
the yield surface is defined by parameters Vmax and aMH, bMH.
Reduction factors ηaMH and ηbMH values < 1 may be used to decrease lat-
eral resistance due to the possibility of a “tapered wedge”-shaped soil gap
occurring in front of the caisson anchor foundation caused by anchor chain
motions. Because optimum lug level (OLL) depth also increases, an ηφMH
value > 1 is also required. To assess project-specific ηaMH, ηbMH and ηφMH
values, 3D FEA are needed (Alderlieste et al., 2016).
The same principle is also applied to “gapping”. Preliminary reduction
factor ηaMH sources possibly include COFS (2003) API/Deepstar project and
pairs of cases C1 and C3 and C2 and C4 from Andersen et al. (2005). From
these, indicative reductions in Vmax (giving ηFo and/or ηREB) and Hmax (ηaMH)
may be assessed. A simpler (very cautious) approach is to simply reduce su
by 50% over the crack depth and use η values = 1. Again, 3D FEA is neces-
sary for detailed/final design.
Lateral resistance. MH ellipses: (a) ηaMH < 1, (b) ηaMH < 1, (c) ηaMH and
ηaMH < 1, (d) ηφMH > 1, (e) ηφMH < 1
Ad (a) through (c):

• Use ηaMH and ηbMH to shrink major and/or minor axes, and hence lat-
eral resistance.
• Hmax resistance is proportional (directly related) to ηaMH.
• Ho resistance is a function of both ηaMH and ηbMH..
In-place resistance 211

Figure 10.13 In-place resistance reduction. Top two rows (a) through (e): MH
ellipses. Bottom row (f) through (h): V-H max ellipsoids.

Ad (d) and (e):

• Use ηφMH > 1 to rotate ellipse clockwise (and hence increase OLL
depth) and ηφMH < 1 for anti-clockwise rotation (decrease OLL depth)

Ellipse rotation (i.e. ηφMH /= 1) alters both Hmax and Ho resistances; ellipse
shape is unchanged.
Axial resistance. V-Hmax ellipsoids (f) ηFo < 1, (g) ηaMH < 1, (h) ηFo and ηaMH
<1
Use either ηFo and/or ηREB to reduce axial resistance Vmax
212 In-place resistance

Table 10.7 Resistance Reduction Examples.

Foundation type Lateral Axial

Feature Support Anchor η aMH η bMH η φMH η Fo η REB

Gapping ⨯ ✓ ✓< 1 ⨯ ✓< 1 ⨯
Seafloor trench ⨯ ✓ ✓< 1 ✓< 1 ✓> 1 ✓< 1 ⨯
Vertical slot ⨯ ✓ ⨯ ⨯ ⨯ ✓< 1 ⨯
Reduced REB ⨯ ✓ ⨯ ⨯ ⨯ ⨯ ✓< 1
Internal scoop ✓ ✓ ⨯ ✓< 1 ⨯ ⨯ ⨯
Wings ⨯ ✓ ✓> 1 ⨯ ✓< 1 ✓> 1 ⨯
HM + ✓ ⨯ ✓> 1 ✓> 1 ⨯ ⨯ ⨯
HM − ✓ ⨯ ✓< 1 ✓< 1 ⨯ ⨯ ⨯

Notes:
ηaMH MH ellipse resistance semi-major axis length reduction factor
ηbMH MH ellipse resistance semi-minor axis length reduction factor
ηφMH MH ellipse resistance rotation angle increase factor
ηFo caisson outer skin friction reduction factor
ηREB caisson tip reverse end-bearing resistance reduction factor
REB caisson tip reverse end-bearing resistance
HM non co-planar HM load
+ overturning M load
− restoring M load

10.12.8 Resistance comparisons
CAISSON_VHM was verified internally (Kay and Palix, 2010–2015). The
following paragraphs give details of external (third-party) comparisons
made with other programs.

10.12.8.1 Support foundations
Organisation A requested CAISSON_VHM analyses for three projects,
comprising two support foundations and one anchor foundation, during
their free trial period.
Organisation A made spot checks at L/D ≈ 0.5, 3 and 1, see L and zlug
italicised data in Table 10.8, and showed reasonable agreement. In addition,
less than 1 s computer time was needed for all CAISSON_VHM analyses (3
runs, 1 or 2 load cases, 24 caisson diameters).

10.12.8.2 Anchor pile (and chain) foundations

As part of their acceptance criteria, Organisation B performed CAISSON_
VHM–BIFURC–PLAXIS 3D comparisons for six anchor sites with Tlug up to
≈ 10 MN.
In-place resistance 213

Table 10.8 Support Foundations – Geometry Comparisons.

ORGANISATION A C AISSON_VHM
Parameter Organisation CASSION Remarks
A Analysis VHM
Analysis
Project 1 – Spreadsheet Support foundation, free head.
SUBSEA Single soil layer. Two VHM
TEMPLATE load cases. Optimum
Diameter D [m] 5.5 5.5 caisson embedded lengths
Embedded length 3 2.4 L opt found for seven caisson
L [m] diameters (2.5 m to 5.5 m).
Project 2 – Internal Support foundation, free head.
SUBSEA Software Two soil layers. Factored
STRUCTURE loads VHM(T) = 2.27, 2.0,
(MONOPILE) 20.6 and 15.9 (MN,m).
Diameter D [m] 8.5 8.5 Reduced clay-steel outer
Embedded length 27.5 27.5 friction adhesion factor
L [m] α o,twist used to account for T
load. Optimum caisson
embedded lengths L opt found
for ten caisson diameters
(4.0 m to 8.5 m).
Project 3 – PLAXIS 3D Anchor foundation, fixed
SUCTION head. Two soil layers.
ANCHOR Factored chain load 17 MN
(MOORING applied at seafloor with an
SYSTEM) (1) angle of 24°. Optimum
Diameter D [m] 12 12 caisson geometry
Embedded length 15 14 (embedded lengths L opt,fix
L [m] and optimum lug level
Lug level depth, 27.5 27.5 depths z lug,opt,fix ) found for
z lug [m] seven caisson diameters
(9 m to 15 m).
Note: PLAXIS 3D analysis was not optimised

CAISSON_VHM (Kay, 2015) uses a yield function based on ≈ 5500

quasi-3D FEA analyses. BIFURC (NGI, 1997) is a 2D FEA plane strain pro-
gram taking 3D effects into consideration using side shear roughness fac-
tors. It uses 8-noded isoparametric elements (pile and soil mass) and 6-noded
interfaces (pile-soil). PLAXIS 3D (PLAXIS, 2013) is a 3D FEA program.
Both BIFURC and PLAXIS 3D are non-linear. As might be expected, the cor-
responding CAISSON_VHM–BIFURC–PLAXIS 3D computer run times are
in the approximate ratio 1 : 100 : 1000.
Based on the results given in Figure 10.14, it is seen that BIFURC predicts
the lowest resistance, and the PLAXIS 3D results, where available, provide
the highest resistance. CAISSON_VHM provides the mid-prediction.
214 In-place resistance

10.12.9 MH ellipse – design examples

Simple calculations can be made without having to use the yield function/
full resistance envelope equations. As an example, find embedded lengths
L for a 5 m diameter caisson, when subjected to three load cases (over-
turning moment, zero moment and restoring moment). The soil is nor-
mally consolidated soft clay (su(kPa) = 2 + 1.5z(m)). Assume that V load is
small and take “standard” Np values of 3 and 10.5 for the “free head” and
“fixed head” situations respectively. Figure 10.15 shows the MathCAD
calculations and associated results (L = 7.5 m, 5.5 m and 2.5 m). Note that
the first calculation makes use of the fact that the resistance envelope is
essentially a straight line between points Ho and Mo (see Section 10.12.3
and Figure 10.11).
In practice, even for preliminary design, the soil profile is usually more
complicated: Np values vary with L/D, and slightly different embedded
lengths L would be obtained. In addition, for the “fixed head” caisson, the

Figure 10.14 Comparison anchor pile Tlug – BIFURC, CAISSON_VHM and PLAXIS 3D.
In-place resistance 215

extremely low embedment ratio L/D implies that cross plates may be neces-
sary in order to prevent internal scoop failure. These calculations assume
that the caisson is not heavily loaded axially – Figure 10.12 shows that
V-Hmax ellipsoids are almost vertical until around V/Vmax = 0.5.
Figure 10.15 shows the example embedded length L calculations. 5 m
diameter caisson in NC clay (su(kPa) = 2 + 1.5z) with factored lateral load
H = 500 kN.

Case AA: overturning moment M = 1500 kNm (L = 7.5 m).

Case BB: free head, M = 0 (L = 5.5 m).
Case CC: fixed head, restoring moment (L = 2.5 m)

Figure 10.15 E xample embedded length L calculations. 5 m diameter caisson in NC clay

(sukPa = 2 + 1.5z) with factored lateral load H = 500 kN. Case AA: overturn-
ing moment M = 1500 kNm (L = 7.5 m). Case BB: free head, M = 0 (L = 5.5
m). Case CC: fixed head, restoring moment (L = 2.5 m).
216 In-place resistance

10.12.10 V-Hmax ellipsoid – design example

An example for optimum lug level (OLL) for anchors is given in Figure 10.16.
This compares HARMONY (Griffiths, 1985, basis of CAISSON_VHM)
with the Andersen et al. (2005) L/D = 1 and L/D = 5 benchmarks.
Unlike in the previous section, the equations are considerably more com-
plex to implement in a simple MathCAD document for illustrative purposes.
These challenges are because:

• there are two equations to be solved – one for the resistance envelope and
another for the chain load at lug level – as summarised in Figure 10.17
• expressions for maximum caisson axial and lateral resistance, plus 3
MH ellipse constants (aMH, bMH and ΦMH) and two VH ellipsoid con-
stants (aVH and bVH) all have to be programmed in terms of caisson
geometry (and shear strength profile).

Figure 10.16 Vmax-Hmax ellipsoid example, optimum lug level (after Kay and Palix, 2010).
In-place resistance 217

Figure 10.17 Optimum lug level - minimisation function equations.

10.12.11 VHM envelope – support foundation design

example
Kay and Palix (2011) presented a typical design application for a 200 mT
manifold founded on a caisson in normally consolidated clay (su profile given
in Figure 10.18a) in the Mya North Field, Bay of Bengal, offshore Myanmar.
The water depth at the manifold location is approximately 190 m. Using
CAISSON_VHM, firstly, the 100+ load cases (operating, hydrotest, acciden-
tal and seismic) were first reduced in number by plotting factored HM load
case data points together with HM resistance envelopes for various caisson
geometries (Figure 10.18b). The most critical load cases are those lying clos-
est to a resistance envelope; if the loads plot within the resistance envelope,
then the design is satisfactory (loads < resistance) but not optimal (loads =
resistance). It is seen that the seismic load cases are more critical than either
operating or accidental loading. In addition, caissons with D = 4 m, L = 12
m will not provide sufficient resistance – they need to be slightly longer.
Similarly, 5 m diameter caissons can be shorter than 12.5 m. Finally, note
that this is a classic case of overturning H and M loading – all data points
lie in the first quadrant of the HM diagram.
Secondly, the factored loads for the most critical load case (factored VHM
loads = 2795, 1079, 7267 [kN, m]) were then used, together with the three
envelope equations, to find the unknown required optimum embedment
lengths Lopt for various caisson diameters D, as shown on Figure 10.18(c).
Results obtained using CANCAP2 limit equilibrium analysis (Fugro, 2009)
are also given for comparison. CANCAP2 automatically searches for a fail-
ure surface giving the lowest safety factor; analysis options include “internal
scoops” – i.e. surfaces passing within soil plug inside the caisson (an exam-
ple is given in Palix et al., 2010). It is seen that CANCAP2 and CAISSON_
VHM curves provide similar results for L/D > 0.5 where no internal scoop
is developed. Had a very shallow caisson option been selected, the use of
internal cross plates could have been considered for two reasons: (a) reduce
settlement magnitude, and (b) prevent “internal scoop” failure.
As expected, a 4 m diameter caisson needs to penetrate more than 12 m
(Lopt = 12.8 m), and a 5 m one less than 12.5 m (Lopt = 11.4 m). Note that
L/D values are compatible with theory (the VHM resistance envelope equa-
tions are valid only for L/D > 0.5), and suction-assisted installation is likely
218 In-place resistance

(a)

Figure 10.18 C
aisson support foundation design example using CAISSON_VHM
and CANCAP2: (a) soil s u profile, NC clay, s u ≈ 2z; (b) MH plot-load
cases and envelopes for given caisson geometries (V = 3000 kN);
(c) required optimum caisson embedded lengths, L opt , and embed-
ment ratios, L opt /D, for given caisson diameters D (factored VHM
loads = 2795, 1079, 7267 [kN, m]) (Kay and Palix, 2011).

to give an insufficient FOS against base failure when L/D > 6). Based on the
resulting Lopt-D curve, and the requirement to limit 30 year settlements to
less than 0.25 m, the final caisson dimensions were D = 5.0 m, embedded
length L = 11.4 m, total length = 12.4 m, side WT = 25.4 mm. The caisson
weight was 73 tonnes.
PLAXIS 3D (PLAXIS, 2013) was used to verify the final caisson geometry
for two of the most critical VHM load cases. The caisson was modelled as
totally rigid with elastic material (steel). The clay-steel outer adhesion factor,
αo, was nominally 0.65, but with an appropriate reduction for torsion
moment. The failure points (solid blue diamonds) obtained by PLAXIS 3D
for the two load cases were added to the CANCAP2 MH resistance enve-
lope and load case MH data (Figure 10.19). It can be seen that the resistance
from the PLAXIS 3D FEA model is very slightly higher than the CANCAP2
resistance envelope. The PLAXIS 3D analysis needed 24 hours to run (2010)
for each load case, whereas all the CAISSON_VHM resistance envelope
analyses took less than 10 seconds.
In-place resistance 219

(b)

(c)

Figure 10.18 C ontinued

220 In-place resistance

Figure 10.19 C
aisson support foundation design example using CAISSON_VHM:
PLAXIS 3D verification. Caisson D = 5 m, L = 11.4 m.

10.12.12 VHM envelope – anchor foundation design

example
Kay (2015) presented a project example for FPSO anchors, offshore Nigeria.
The objective was to find optimum foundation geometries for 0° and 5° tilt.
Data were as follows:

caisson: D = 3 m, 3.5 m and 4 m. Wall and top plate thicknesses are D/200
and D/100
lug: on caisson outer face, zlug at optimum depth
soil: su = 2 + z (0 ≤ z ≤ 5 m), su= 7 + 1.73 (z - 5) (5 m ≤ z ≤ 20 m) kPa, γsub
= 4 kN/m3
soil-pile: αo = 0.6 (including twist allowance), αi = 0.65.
pull-out Vmax: lesser of plugged and unplugged conditions, ηREB = 0.8
chain: 81 mm diameter, submerged weight 1.12 kN/m.
loads and FOS: see Table 10.9.

Table 10.10 compares detailed results for both tilt angles. Figure 10.20a
summarises the caisson geometry optimisation mode, which finds the
optimum “fixed head” zlug,opt,fix and Lopt,fix for given caisson diameter(s) D,
In-place resistance 221

Table 10.9 FPSO Anchor Design Example: Chain Load Data.

T z0 θ z0 λT
Load Case [kN] [deg] [–]
1: All Lines Intact 2045 7.5 2.00
2: One Line Damaged 3377 13.7 1.50
3: Transient 4211 15.5 1.25

Note: 3.5 m OD, LC3

Table 10.10 F PSO Anchor Design Example. CAISSON_VHM Optimisation Results.

Parameter 0° tilt 5° tilt

L opt,fix [m] 13.2 13.8
z lug,opt,fix [m bsf] 8.5 8.7
T lug [kN] 4077 4097
θ lug [deg] 19.1 23.6
Bw [kN] 192.2 199.3
V max [kN] 1934 2101
H max [kN] 5456 5981

Note: 3.5 m OD, Load Case 3

seafloor chain loads Tz0 and inclinations θz0 (i.e. Tlug and θlug at caisson lug
level zlug). A 5° tilt angle was used. Figure 10.20b shows that LC3 always
has the largest embedment length L and is therefore critical. The associ-
ated maximum embedment ratio L/D is just over 6, suggesting that base
heave installation may be an issue had D been less than 3 m. Figure 10.20c
and Figure 10.20d provide optimum buoyant weight Bw and geotechnical
efficiency (defined as η = Tlug/Bw) data used to finalise foundation geome-
try. LC3 efficiency η values decrease from ≈ 22 (D = 3 m) to 17 (D = 4 m),
showing that “deeper is better”. Final design was based on 3.5 m OD
caissons (5° tilt, LC3), see Table 10.12. Additional checks subsequently
made included capacity, lug level, tilt and twist. The sensitivity mode used
(vary θlug and/or zlug, find Tlug) is shown on Figure 10.21a. Figure 10.21b
and 10.21c show Tlug sensitivity to θlug and zlug respectively.
Less than 5 s computer time was needed for the two runs for Figures 10.20
and 10.21.
Alderlieste et al. (2016) give details of a more complicated anchor chain
trenching example. This used PLAXIS 3D to first assess lateral resistance
reduction factors, followed by CAISSON_VHM analyses for in-place resis-
tance. Details are given in Section 12.12.
222
In-place resistance
Figure 10.20 F PSO anchor design example using CAISSON_VHM, geometry optimisation mode (zero rotation). 5° tilt: (a) optimisa-
tion solution – (a) find L opt,fix and z lug,opt,fix for given D, T lug and θ lug (b) L opt,fix , and L/D versus D, (c) B w and L opt,fix versus D,
(d) efficiency (T lug /B w ) and L/D – D (Kay, 2015).
In-place resistance 223

(a)

(b) (c)

Figure 10.21 F PSO anchor design example, T lug sensitivity mode (non-zero rota-
tion). D = 3.5 m, LC3, λ T = 1.25; (a) optimisation solution – find T lug
for given L, D, z lug and θ lug , (b) T lug λ T and T lug versus θ lug at z lug,opt,fix ,
(c) T lug versus z lug at θ lug (Kay, 2015).

10.12.13 Drained soil response (“sand”)

A generic envelope for intermediate (and deeply embedded surface) founda-
tions under VHM loads is not yet available. Derivation is difficult: in the
HM plane results are extremely sensitive to V load (unlike undrained clay).
Also, there are more parameter variables (φ', Ψ', tan(δ), Ko, etc.) and gap-
ping to be considered, and at what caisson displacement to take failure resis-
tance in numerical analyses. Finally, sand profile characterisation is more
difficult than clay – it is rare to get “homogeneous” conditions (e.g. a con-
stant relative density Dr) in the field.
Some work has been done by extrapolating the “rugby ball”-shaped enve-
lope available for surface foundations (Butterfield and Gottardi, 1994) and
VHM equation (Gottardi et al., 1999). Senders (2005), in order to provide
tensile (and compressive) axial resistance, simply added an additional V
term equal to outer friction and foundation steel submerged weight, although
this severely underestimates the maximum lateral resistance due to extra
224 In-place resistance

embedment. Govoni et al. (2016) provide h0, m0 and a values for strips in
30° sand at L/D = 0 and 0.5.

10.13 RESISTANCE AT SHALLOW PENETRATION

Occasionally it may be necessary to assess intermediate foundation resis-

tance at shallow penetration (embedment ratio L/D < 0.5). Examples include:

• limited penetration due to competent soil

• partial penetration and summer storm.

In such cases, shallow foundation design principles (e.g. ISO 19901-4:2016

Section 7.3.5 and Annex A.7.3.5) should be used. Scour is not generally
considered.

10.14 RESISTANCE IN (WEAK) ROCK

10.14.1 General
The axial resistance and initial stiffness of weak rock will depend greatly on
the mode of installation. Driving in weak rock will lead to the creation of a
remoulded or crushed zone around the wall of the structure and potentially
will lead to radial cracks around the structure (Section 9.13). The axial and
lateral resistance of piles in weak rock are discussed in the following section
for both installation cases.

10.14.2 Axial resistance of driven piles in weak rock

After driving, the rock around the pile wall will be partially or totally
crushed. The material at the rock-pile interface will either be closer to gran-
ular material or cohesive material. In both cases, this crushing will result in
decreasing normal stresses that will significantly reduce the shaft friction
compared to a drilled and grouted interface. The axial friction in weak rock
will mainly depend on rock mineralogy, void ratio, rock strength, rock mass
properties and all factors affecting the in-situ stress (e.g. fractures and
joints).
Two main approaches can be considered for assessing the shaft friction of
driven piles in rock (Irvine et al., 2015):

• Effective stress methods (as for cohesionless soils):

Qs K.v. tan . As f1. As

(10.36)
where:
Qs = shaft capacity
In-place resistance 225

K = coefficient of lateral earth pressure

σ'v = overburden stress
δ = friction angle between the rock and pile wall
As = surface area of the pile shaft
fl = limiting skin friction.

The K factor should ideally be measured by pressuremeter testing that would

account for in-place conditions such as rock mass stiffness and anisotropy.

• Total stress methods where a pile-rock adhesion is calculated (as for

cohesive soils):

Qs .UCS. A s fl . As (10.37)
Where:
α = adhesion factor
UCS = Unconfined Compressive Strength.

This latter method appears to be better suited to weathered mudstone and

weak calcareous sedimentary rock. Terente et al. (2017) propose a lower
bound adhesion factor formulation based on published driven piles data
(see Figure 10.22).
The adhesion factor is given by the following equation:

0.11. UCS
0.5
(10.38)

For cemented carbonate formations, ARGEMA (1994) propose an adhesion

factor ranging between 0.1 and 0.2 based on Burt and Harris (1980) and
Puech et al. (1988). Mention is made of field tests showing evidence of shaft
friction up to 300 kPa for rock formation with UCS up to 5 MPa.
In chalk, driving will generate pore pressure in the remoulded material
(also referred to as putty chalk) that will further reduce the effective stresses
around the pile. CIRIA (2002) recommend considering a shaft friction just
after driving of 120 kPa for high-density Grade A chalk and only 20 kPa for
all other chalks. These recommendations do not account for effects of time
and cyclic loading, both having opposite and significant impacts on chalk.
Much effort is currently being made by the wind energy industry to better
characterise the shaft friction in general or for specific projects. Extensive
laboratory tests (see Houlston et al., 2017; Dührkop et al., 2017) or off-
shore or onshore pile tests campaigns (see Barbosa et al., 2015) are more
and more commonly conducted to justify using higher friction values. A JIP
project, ALPACA, was initiated in 2018 to propose new design methodolo-
gies for pile in chalk. This project includes a series of onshore pile tests
226 In-place resistance

Figure 10.22 A
dhesion factor (α) against rock material UCS (Terente et al.,
2017).

conducted in the UK (St Nicolas site), comprising axial and lateral load tests
(Jardine et al., 2019).

10.14.3 Axial resistance of drilled and grouted

piles in rock
Pile construction factors are playing a great role in the quality of the grout/
rock interface. For drilled and grouted piles, aspects that need to be consid-
ered are:

• transfer of steel/grout forces

• hole stability
• presence of granular material (risk of wash out)
• roughness of the hole
• cleaning of the hole (for both axial friction and end-bearing)
• fracturing breakdown (grout pressure)
• grout quality (strength, swelling/shrinkage, heat of hydration).

Hole stability is one of the first factors to consider for a drilled and grouted
foundation. It is noted that when sub-layers of granular material are encoun-
tered in the rock matrix, the drilling fluid circulation will most probably
wash out these materials. Wash out will produce large voids that could lead
to stress relief in the rock mass, and also require a significant amount of
additional grout to be injected to fill these voids. If pressurised, the injection
of grout can also lead to amplification or creation of fractures.
In-place resistance 227

Potential effects of drilling mud need to be considered when determining

the shaft friction. The sidewall of the hole can develop a layer of slaked
mud or clay, which will never gain the strength of the rock. Roughness of
the hole wall is also an important factor. Depending on the roughness, the
shearing will occur in a zone impacted by the drilling (slaked mud or
crushed material) or within the rock mass. To ensure a good shear transfer,
the roughness can be artificially increased by the use of an under-reamer at
regular depths. Figure 10.23 shows the drilling head equipped with an
under-reamer used for the Pluto project offshore Australia. Grooves were
made every 1.5 m to perform so-called “ribbed” rock socket foundations
for the Pluto jacket.
The cleanness of the drilled hole plays an important role on the end-
bearing response. In general, the end-bearing capacity should be factored or
neglected in the design, depending on pile construction aspects such as
removal of drill cuttings from the base of the hole and presence of disconti-
nuities within the rock mass.
Skin friction for socket piles can develop in one of three ways:

• through a shearing bond between the grout and the steel pile
• sliding friction between the grout shaft and the rock (when the drilled
hole is smooth and grout infiltration limited)
• dilation of an unbounded rock-grout interface until asperities shear off.

For stubby piles submitted to a large number of cycles of loading, the axial
resistance should not rely on the passive shear capacity between the steel
and cast grout (c.f. DNVGL RP 0419 recommendations for grouted connec-
tions). Pile movement during grout curing could jeopardise the steel/grout
bonding. In addition, cyclic lateral loading will tend to create a gap between
the grout and the steel. The shear capacity of the interface can be increased

Figure 10.23 P
luto – drilling head (3m diameter) with underreamer. (Courtesy of
Large Diameter Drilling (LDD))
228 In-place resistance

by using shear keys, i.e. protuberances on the steel surface, on the interface
that when subsequently cast into the grout provides a mechanical resistance
to relative sliding between the steel and grout. API RP 2A WSD (2000) and
ISO 19902 (2011) provide guidance on the number of shear keys to be
considered.
In non-carbonate weak to moderately strong rock, high friction values
have been recorded. Most published methodologies define the unit shaft
friction of drilled and grouted piles as a function of the Unconfined
Compressive Strength (UCS), independent of the type of friction developed
(i.e. sliding or dilation). Table 10.11 lists the main methodologies proposed
for shaft friction, fs, where

fs . UCS (10.39)
m

where:
UCS: Unconfined Compressive Strength
m = constant usually taken as 0.5
α: adhesion factor comprised between 0.2 and 0.45.

The large variability obtained by these empirical methods highlights the

limitation by the use of the UCS as the only reference. Figure 10.24 shows
the spread of adhesion factor for different rock types (Pells et al., 1980). In
addition, these methods do not account for drilled hole dimensions (i.e.
diameter and roughness), natural fractures, rock mineralogy and porosity.
Seidel and Collingwood (2001) proposed another approach incorporat-
ing the complex mechanisms of shear transfer at the interface between the
socketed piles and the surrounding rock. They define a shaft resistance coef-
ficient (SRC) incorporating the contribution of the rock strength (drained
intact and residual strength parameters are used), socket roughness, rock
mass modulus, Poisson’s ratio and socket diameter. The formulation of the
SRC is proposed as

Table 10.11 U
ltimate Friction Formulation for Drilled and Grouted Piles in
Weak to Moderately Strong Rock.

Drilled and grouted methods Ultimate friction formulation

1. Rosenberg and Journeaux (1976) 0.375(UCS) 0.515
2. Horvath (1978) 0.33(UCS) 0.5
3. Horvath and Kenney (1979) 0.2–0.25(UCS) 0.5
4. Meigh and Wolski (1979) 0.22(UCS) 0.5
5. Williams and Pells (1981) α∙β∙UCS (where β is a constant)
6. Rowe and Armitage (1987) 0,45.(UCS) 0.5
In-place resistance 229

Figure 10.24 Shaft resistance correlations following Seidel and Collingwood

(2001) for roughness classes R1-R3 of Pells et al. (1980).

n r (10.40)
SRC c
1 ds

where:
∆r is the mean roughness height
ds is the socket diameter
ηc is the construction method reduction factor
n is the ratio Em/UCS of rock mass modulus over unconfined compressive
strength.

The effect of the SRC on the adhesion factor is provided in Figure 10.25.
For slightly cemented carbonate rock (UCS < 5MPa), ARGEMA (1994) sug-
gest considering the ultimate shaft friction at the carbonate rock–grout
interface proposed by Abbs and Needham (1985; see Figure 10.26). For
chalk, CIRIA (2002) differentiate the axial resistance depending on the den-
sity. For low to medium density, the friction is a function of the average
effective stress, fs = 0.8 σv'. For high-density chalk, the axial resistance is
proportional to the UCS: fs = 0.1 UCS.
230 In-place resistance

The methods just outlined rely on the measurement of the unconfined

compressive strength of the rock. More recent designs rely on an analogy
between the displacement of the rock socket and displacement of small scale
rock/grout interface loaded under constant normal stiffness (CNS) condi-
tions. CNS conditions are believed to be more representative of the

Figure 10.25 E ffect of shaft resistance coefficient (SRC) on socket adhesion fac-
tor (Seidel and Collingwood, 2001).

Figure 10.26 U
ltimate shaft friction at the carbonate rock–grout interface (from
Abbs and Needham, 1985).
In-place resistance 231

evolution of normal stresses applied on the interface during rock socket pile
displacement (Figure 10.27 and CFMS, 2019).
The imposed stiffness should be representative of the actual rigidity of
the rock mass. The rock–grout interface is represented by a regular saw-
tooth with a chord length and an asperity height corresponding to the
roughness of the in situ hole. Guidance for characterising rock masses for
designing drilled and grouted offshore pile foundations can be found in
Puech and Quiterio-Mendoza (2019), and testing conditions are described
by Stavropolou et al. (2019).
In the case of weak rock (i.e. grout being much stronger than the rock),
the interface response will be governed by the rock behaviour. The failure
surface will pass through the weak rock. For stronger rocks (i.e. strength
similar to the grout), it is more the “geometrical” dilatancy that will govern
the interface response.
The main challenge of this experimental methodology (based on CNS
tests) is to get access to proper design parameters and installation conditions
at an early stage in the design process. The measured shear force is very

Figure 10.27 Idealised displacement of a rock socket.

232 In-place resistance

sensitive to the shape of the asperities. It remains, however, the best way to
understand the degradation of asperities under cyclic loading without
requiring full-scale tests.

10.14.4 Lateral resistance in rock

It should be highlighted that experience of foundations installed in weak
rock is relatively limited, particularly for driven piles. Rock is often encoun-
tered below upper soils and therefore usually mobilised far from its failure
point.
The most commonly considered approaches to derive the lateral resis-
tance of a pile in weak rock are listed in Table 10.12.
The first two p-y methodologies (Abbs, 1983; Fragio et al., 1985) are
composed of an initial intact rock resistance followed by residual resistance
at large strains, shown in Figure 10.28. In the Reese (1997) framework, the
lateral pressure increases until reaching an ultimate value.
An alternative model, CHIPPER (Erbrich, 2004), takes cyclic loading
into account explicitly. CHIPPER aims to capture the brittle stress-strain
behaviour through an algorithm modelling the development of fractures
(“chipping”) near the surface of the rock (Figure 10.29).
Other methodologies for calcareous sand have been proposed (e.g.
Novello, 1999; Dyson and Randolph, 2001) but are applicable only to
weakly cemented calcareous sediments in which CPTs can penetrate.

Table 10.12 Methods to Determine Lateral Resistance in Weak Rock.

Reference UCS (MPa) Rock type Failure mode Based on

Abbs 0.5–5 Carbonated Brittle near the Theoretical

(1993) surface and basis
ductile deeper
Fragio et al. 9–36 Calcareous Brittle near the Bored piles, D
(1985) Claystone surface and = 0.405 m
ductile deeper
Reese 3–16 Various Ductile Bored piles,
(1997) L/D = 6 to 11
Dyson and CPT can Cemented Ductile Centrifuge
Randolph penetrate carbonated tests
(2001) soil
Erbrich No range Cemented Brittle near the Theoretical +
(2004) provided carbonated surface and centrifuge
soil ductile deeper tests + 3D
FE

Notes: L: embedment length, D: pile diameter, UCS: unconfined compressive strength

In-place resistance 233

(a)

(b)

Figure 10.28 C
haracteristic shape of p-y curves for weak rock according to (a)
Fragio et al. (1985) and (b) Reese (1997) methods. (With permis-
sion from ASCE)
234 In-place resistance

Figure 10.29 C
oncept of “chipping” and characteristic shape of p-y curves (after
Erbrich, 2004).

The main limitations of these existing methodologies are as follows:

• Only a few field tests were considered.

• Most of the field tests included bored piles. The effect of pile installa-
tion by driving is therefore not considered explicitly (it could be by
introducing a damage factor but with no guidance on how to define
this damage factor).
• Methodologies were developed for slender piles (apart from the
CHIPPER method also used for large diameter anchor piles, D = 5 m)
and do not provide any recommendations for extrapolation to large
diameter piles.
• The effect of cyclic loading is not discussed in these methods apart
from the CHIPPER method proposing a cycle-by-cycle approach.
• These methods use UCS (or cone resistance for Dyson and Randolph,
2001) and Rock Quality Designation (RQD). They do not consider the
porosity, carbonate content or fracture orientations.
In-place resistance 235

10.15 GROUP RESISTANCE

10.15.1 General
Groups of suction foundations may be installed. Examples are given in Figure
10.30. They include two or more suction anchors for FPSO mooring points,
and braced foundation clusters to ensure a near-vertical subsea template.

10.15.2 Considerations
Considerations include:

• centre-to-centre spacing – interaction potential

• braced or unbraced – kinematic constraints
• MH load – directionality

Like pile groups, the primary consideration is the spacing between individ-
ual foundations. At close spacings, when foundations interact, bracing (if
any) imposes kinematic restraints – an unbraced self-installing platform
behaves differently than a braced tripod.
Group load (capacity and response) is normally considered when the centre-
to-centre spacing is < 8 D. Generally, for foundations in undrained “clay”,
the group capacity may be less than the single suction foundation capacity
multiplied by the number of foundations in the group. Axial and lateral
group deflections are normally larger than that of a single suction founda-
tion subjected to the average load of the group.

10.15.3 Design procedures
There are no established procedures for designing groups of shallow or
intermediate foundation systems.
Therefore, engineers resort to ad hoc (case-by-case) approaches.

10.15.4 Design tools
Design tools (listed in order simplest first) include:

• equivalent area foundation

• VHgroup = sum(single); Mgroup= sum(V*lever arms)
• plastic limit analysis/resistance envelope
• 3D FEA.

10.15.5 Braced support groups

10.15.5.1 Shallow foundation groups
Various offshore structures are founded on multiple shallow foundations
such that the kinematic constraint of the structure leads to the foundation
236 In-place resistance

(a)

(b)

Figure 10.30 E xamples of suction foundation groups: (a) four-legged structure

(Courtesy Bo B. Randulff, Equinor), (b) cluster foundation (Courtesy
SPT Offshore).
In-place resistance 237

(c)

(d)

Figure 10.30 C
ontinued: Examples of suction foundation groups: (c) cluster pile
for Ceiba project (Courtesy SPT Offshore), (d) L6-B platform dur-
ing sail-out (from Alderlieste and Van Blaaderen, 2015).
238 In-place resistance

elements to act in concert. Examples include jackets permanently supported

on suction cans, piled jackets temporarily supported by mudmats, and a
variety of subsea infrastructure.
The vertical and horizontal capacity of structurally connected shallow
foundation systems is relatively unaffected compared to a single foundation
of equivalent bearing area and embedment ratio. However, the structural
rigidity of a multi-footing foundation system enhances overturning capacity
due to the structural connection between adjacent foundations (Murff,
1994; Fisher and Cathie, 2003; Gourvenec and Steinepreis, 2007; Gourvenec
and Jensen, 2009). There is no established procedure for stability calcula-
tions of a shallow foundation system comprising co-joined footings, and
ad-hoc approaches may involve simply summing the ultimate limit states of
the individual footings under vertical and horizontal load and considering a
push-pull mechanism for moment resistance, using a method of linear
springs or carrying out project-specific finite element analysis (Randolph
and Gourvenec, 2011).

10.15.5.2 Anchor pile groups

Cox et al. (1984) reported results of model laterally loaded pile groups in
soft clay. Their data, summarised in Figure 10.31, suggests that groups of 3
and 5 side-by-side piles with centre-to-centre spacings of 3D to 4D (i.e.
2D-to-3D soil gap) develop lateral resistances approaching that of single
isolated piles. The braced boundary condition is dissimilar to that usually
encountered for pile anchor clusters (Figure 10.30).

Figure 10.31 Laterally loaded anchor. Braced pile groups (Cox et al., 1984).
In-place resistance 239

10.16 CLOSURE

This chapter has covered in-place resistance, both for support and anchor
pile type intermediate foundations, and the following sections subdivide
comments into those for “clay” and “sand”.

10.16.1 Clay
As for installation, models for undrained “clay” are more developed than
for drained “sand”. Reasonably accurate models have been given to assess
maximum available Vmax, Hmax and Ho (“fixed” and “free head”) and Tmax
resistances. For anchor foundations, pull-out Vmax has three possible failure
models – “plugged”, “coring” and “leaking” (see Figure 10.1), for which
both applicability criteria and equations are provided.
VHM(T) interaction effects are important: HM load affects V resistance,
and, to a lesser extent, T load affects V resistance. Like shallow foundations,
VHM(T) yield envelopes for undrained soil are a promising development
for preliminary/routine design. Equations and software are available. PLET
and FPSO anchor design examples are also given.
Both FEA and centrifuge tests suggest that the original (Skempton, 1951)
bearing capacity factor Nc,circle is cautious. Hence Qtip is probably underesti-
mated. This has implications for both reverse end-bearing (in-place, this
chapter) and plug heave (installation, Chapter 9) failure.
For anchor piles, Vmax-Hmax and Vlug-Hlug, envelope shapes differ; compare
Figure 10.16 with Figure 10.5. This should be considered during detailed
design when making sensitivity analyses.

10.16.2 Sand
Models for “sand” are less mature than for “clay”. This is as expected: effec-
tive stress governs strength, and pore water build-up/dissipation is an addi-
tional complicating factor.
For sustained load, sands behave drained, and calculations can be easily
made to assess both the compressive resistance Vmax (which is typically very
large) and the tensile pull-out resistance (which is very small). There is also
a good understanding of near-surface foundations to combined VHM loads,
largely based on experimental data. Yield surfaces defining possible VHM
load combinations have been published, but these are for either shallow
foundations or jack-up spudcans. Hence, as a start, a nice research project
would be to derive a reasonably cautious model for assessing intermediate
foundation resistance in drained sand under static VHM loading. Besides
foundation geometry, which must take account of embedment, key soil
parameters include friction angle φ' and dilatancy parameter ψ'. This
research would hopefully lead to a drained resistance envelope and yield
240 In-place resistance

function similar to, but far more complex than, that developed for interme-
diate foundations in undrained soil.
For short-term loading (e.g. a single wave load or boat impact), resistance
estimates are challenging because the equivalent undrained sand strength is
very difficult to assess. Undrained strengths are principally determined by
dilatancy parameter D (DNV, 1992) and the onset of pore fluid cavitation.
Hence both vertical compressive and tensile resistances are typically very
large, but difficult to assess. The resistance to combined moment and hori-
zontal loads is also considered to be large. Like clays, VHM interaction
effects must be considered. Resistance assessments become even more com-
plex for cyclic load. An even more challenging research project would be to
initially assess resistance in undrained sand under static VHM loading.
Chapter 11

In-place response

11.1 INTRODUCTION

This chapter deals with intermediate foundation response under static

(Section 11.2), dynamic and non-dynamic cyclic loads (Section 11.3). It is
seen that the response topics considered are basically similar to those for
shallow foundation design.
The definition of dynamic and non-dynamic cyclic loads is as follows: a
cyclic load can be termed dynamic if the rate of loading is such that inertial
forces are significant compared to the static loads (e.g. boat impact,
earthquake or ice shock forces). If inertial forces are negligible (e.g. wave
forces), then the loading is said to be non-dynamic cyclic (Chan and Hannah,
1980).

11.2 DISPLACEMENT UNDER STATIC LOADS

11.2.1 General
Calculation of foundation displacements (usually axial settlements) should
consider the following components:

• immediate displacements
• primary consolidation settlement
• secondary compression (creep) settlement
• settlements induced by cyclic loads
• regional settlements.

The first three components are the same as for shallow foundations (ISO
19901-4:2016, section 7.5.2.1). Differential settlements (induced by spatial
soil variability, moments, torque and eccentricity) are not assessed for indi-
vidual intermediate support foundations, as they are essentially rigid.

241
242 In-place response

For routine design, soil type considerations include:

(a) Sands (and stiff clays and rock): settlement magnitudes are small and not
problematic. Hence, settlements are not considered for routine design.
(b) Clays (fine-grained lithogenous sediments) are the most abundant of
all marine sediments (about 70% by volume). These clays are usually
normally consolidated (or at best lightly overconsolidated) and give
the largest settlements.
(c) Organic soils (peats) are rarely encountered offshore.

Table 11.1 summarises the relative importance of each of these three settle-
ment components for different soil types.
Settlement considerations include:

(a) There is no experience with long-term settlements.

(b) The largest settlement component is judged to be consolidation; how-
ever, cyclic and/or creep components could be significant. For example,
extreme wave loading on a stubby caisson in sand may degrade outer
friction, causing untoward settlement of a non-underbase grouted sup-
port foundation. Similarly, OWT monopiles have strict tilt criteria for
accumulated displacement due to omni-directional cyclic lateral load.
(c) Consolidation of a clay soil plug may be important; the plug is incom-
pressible during immediate loading but undergoes subsequent primary
consolidation.
(d) Settlement assessments should generally aim to maximise the largest
component (usually soil plug compression). Hence, the most impor-
tant soil parameters are magnitude of outer friction and soil compres-
sion indices.

Section 11.2.3 discusses primary consolidation models.

Table 11.1 R
elative Importance of Settlement Components for Different Soil Types.
Intermediate Support Foundations – Routine Design

Soil Type

Settlement Component Sand Clay Organic Clay (Weak) Rock

immediate yes possibly possibly yes
primary consolidation no yes yes no
secondary compression no possibly yes no
(creep)
cyclic possibly possibly possibly Possibly for
weak rock
regional possibly possibly possibly possibly

Note: Based on Table 5.1 of Holtz (1991)

In-place response 243

11.2.2 Immediate displacement
References for methods to assess immediate, elastic displacements that
account for non-uniform soil profiles (e.g. linearly increasing soil strength)
and foundation embedment are provided in this section. The results are
generally used to provide a range of foundation stiffnesses for use by
structural engineers or to check the initial portion of “load-settlement”
response from geotechnical finite difference/element analyses.
The complete 4 × 4 seafloor foundation stiffness matrix (Section 8.3 and
Figure 8.6c) should be assessed. Note that the lateral stiffness sub-matrix
should not be derived using either p-y curves for deep (pile) foundations or
elastic solutions for flexible piles. Reasons include the fact that intermediate
foundations are essentially rigid and generally larger in diameter. Instead,
either elastic solutions or numerical analysis should be used (DNV OS-J101,
2014a).
In order to evaluate the terms appearing in the stiffness matrix, sources
for useful elastic solutions (all continuum based) for statically loaded
circular intermediate foundations are listed in Table 11.2). It is seen that
only Doherty and Deeks (2003) provide all six stiffnesses – and in non-
homogeneous soil – but only for embedment ratios L/D ≤ 2. Of these
solutions, only Baguelin et al. (1977) include a weaker zone. Their equation
12 can be used to demonstrate that a disturbed soil annulus has only a
limited effect on lateral stiffness kxx. For example, a disk with diameter D =
5 m, outside boundary radius R = 15D, soil Young’s Modulus E' = 100 MPa,
and Poisson’s ratio ν' = 0.3 has an intact stiffness kxx = 121 MN/m/m length.
For a 0.5 m wide disturbed zone with E' = 40 MPa, kxx reduces to 113
MN/m/m length, i.e. a small 7% decrease. This has implications for laterally
loaded OWT monopiles installed by vibratory methods, where both the
annulus width and soil modulus are likely to be greater than for driven
foundations. That is, since lateral stiffness is largely due to the far field
displacement regime, and most lateral displacements at depth remain in the
elastic zone, driven and vibrated monopile lateral responses under HM load
will not differ significantly.
Until Gupta and Basu (2016) and Arany et al. (2017), there were no elastic
continuum solutions for laterally loaded OWT type rigid caissons
(embedment ratio L/D ≤ 10) in non-homogeneous soil. Carter and Kulhawy
(1992) were valid for L/D ≤ 10 but are only for uniform soil and Doherty
et al. (2005) included non-homogeneous soil but for L/D ≤ 2.
There are no elastic solutions for deeply embedded rigid circular founda-
tions under MH loading – this situation occurs at the base of a 1-D Winkler
support model for OWT monopiles: Poulos and Davis (1991) give solutions
for V load only and for MH load on an embedded vertical plate (their sec-
tion 15.4), and Doherty and Deeks’ (2003) caisson geometry (b) is valid
only for L/D = 2.
244 In-place response

Table 11.2 E lastic Continuum Solutions for Intermediate Rigid Foundations

– Static Load

Foundation Stiffness Matrix

Term

Axial Lateral Torsional

Shallow or Intermediate kxθ

Foundation Type kzz kxx kθx kθθ kψψ Reference Remarks
Surface ✓ ✓ ⨯ ✓ ✓/⨯ DNV Table G-1
VHMT load (2014a) Circular
footing on
stratum over
bedrock or on
stratum over
half space
Embedded footing ✓ ✓ ⨯ ✓ ✓/⨯ DNV Table G-2
VHM load (2014a) Circular
footing
embedded in
stratum over
bedrock
L/D ≤ 1
Embedded footing ✓ ✓ ⨯ ✓ ✓ Gazetas homogeneous
arbitrarily shaped (1991) half-space
VHMT load
Four geometries: ✓ ✓ ✓ ✓ ✓ Doherty Rigid foundation
(a) footing: at base of and Non-
hole; (b) embedded Deeks homogeneous
footing; (c) solid (2003) half-space
caisson; (d) caisson L/D ≤ 2
VHMT load
Solid caisson ⨯ ✓ ✓ ✓ ⨯ Carter and Corresponds to
HM load Kulhawy Doherty and
(1992) Deeks (2003)
geometry (c) in
homogeneous
half-space.
L/D ≤ 10
Solid caisson ⨯ ✓ ⨯ ✓ ⨯ Gupta and Multi-layered
HM load Basu soil
(2016)
Solid caisson ⨯ ✓ ✓ ✓ ⨯ Arany et al. Table 3
HM load (2017)
Pile ✓ ⨯ ⨯ ⨯ ⨯ Randolph Equation (30)
V load and
Wroth
(1978)
Disk ⨯ ✓ ⨯ ⨯ ⨯ Baguelin et Equation (8)
H load al. (1977)
Pile ⨯ ⨯ ⨯ ⨯ ✓ Randolph Equation (9)
T load (1981)
In-place response 245

11.2.3 Primary consolidation settlement

According to Hernandez-Martinez et al. (2009), “No literature source has
reported in situ long term settlement measurements for skirted foundations.”
Experience with monitoring gravity platforms whose foundation comprises
skirted walls is available; Lunne et al. (1982) presented a comprehensive
study showing settlement records obtained from skirted foundations of plat-
forms in the North Sea for continuous monitoring periods that varies from
approximately 7–8 months to 40 months. However, this experience could not
be transferred directly to skirted foundation analyses because these platform
measurements are restricted to small depth to width ratio. More recently,
Svanø et al. (1997) reported settlement records from the Troll A platform
foundation in the North Sea after 10 years of continuous monitoring.
Centrifuge tests provide an alternative to further investigate skirted founda-
tion long-term settlement evaluation. So far, a number of centrifuge tests have
been performed worldwide to better understand issues such as installation and
capacity, but no emphasis has yet been placed on long-term settlement.
Chen and Randolph (2007a) reported results from a series of centrifuge
tests where instrumented suction anchors (simulating a suction anchor
prototype of 14.4 m length, 3.6 m diameter, 0.06 m wall thickness and 230
kN self-weight) were modelled in normally consolidated reconstituted
kaolin. Although their main objective was to investigate the external radial
stress changes around suction anchors only, they reported that after the
suction anchor was installed to target penetration depth, followed by one
hour of consolidation (equivalent to 1.7 years at prototype scale), the
settlement of the suction anchor model was 0.10 mm (equivalent to 2 cm at
prototype scale). This settlement was due, however, only to suction anchor
self-weight. Larger settlement is anticipated once additional loads are
considered as the case of manifold and pump stations.
The primary consolidation settlement model given in the following para-
graphs is considered reasonable for routine intermediate support foundation
design in clay: both data and model accuracy (Lambe, 1973) are similar.
Load transfer of seafloor load V is via outer skin friction, and the rest is
taken by base load. The base load compresses the soil plug upwards and the
soil below foundation tip level downwards. The corresponding foundation
consolidation settlement model is shown in Figure 11.1 and consists of
compressing two non-linear springs in series (Figure 11.1a). The top spring
represents the soil plug and the bottom one the soil below tip level
(embedment depth L).
Considering the foundation outer steel and soil plug base (Figure 11.1b),
seafloor (permanent) load QSLS is resisted by mobilised outer skin friction Fo,
and the remainder by base load (Qbase = QSLS − Fo), acting over the complete
foundation area Abase. Figure 11.1c and 11.1d show the forces on the
foundation steel and soil plug respectively for the usual case of no top plate
bearing. Wall base load, Qwall, is essentially zero (Figure 11.1c). Hence,
246 In-place response

(a) (b)

(c) (d)

(e)

Figure 11.1 P
rimary Consolidation Settlement (support foundation): (a) simple
two-spring model, (b) through (e) corresponding (SLS, permanent)
loads and resistances; Soil Plug submerged weight not shown for
clarity.
In-place response 247

mobilised inner skin friction Fi, which acts upwards on the steel and down-
wards on the soil plug, has a value equal to Qbase. Considering loads on the
clay soil plug (Figure 11.1d), the top spring is compressed upwards from tip
level by a vertical effective stress increase ∆σv' = Qbase/Abase. Due to inner
friction resistance Fi, ∆σv' decreases with distance above tip level, and the
corresponding uplug (soil plug spring compression) final primary consolida-
tion settlement value can be assessed using textbook equations (e.g. Holtz,
1991) for 1-D oedometer compression and soil parameters Cc, Cs (compres-
sion indices) and p'c (preconsolidation pressure).
Additional settlement will occur due to increased vertical soil stresses in
the vicinity of the foundation. Therefore, settlement assessment also requires
calculation of the stress distribution in the soil below the foundation. Hence,
utip (soil settlement below tip level) can be computed using loads at tip level
and conventional load spread theory or similar. Distributions of ∆σv' (which
may be assumed to be uniform) due to two loads which need to be superim-
posed. Figure 11.1e shows that they are (i) base load Qbase acting over a
circle diameter D and (ii) mobilised outer skin friction Fo spread out with a
load spread factor ns to act over an annulus with inner diameter D and outer
diameter D + 2 L/ns.
For a fixed geometry, consolidation settlement magnitudes are sensitive
to Fo, Cc, p'c and Cs (in that order). If uncertainties exist, then analyses
should consider a high estimate of uplug. This is achieved by using a low
estimate of Fo. The model presented here is useful for foundation geome-
tries where plug final settlement may be significant, generally embedment
ratios L/D less than around 3. Longer intermediate foundations behave
more like pipe piles – tip settlement component utip dominates but is small
at working loads.
For axially loaded support foundations in clay, settlement rates are a
product of 1D vertical drainage within soil plug plus 2D drainage (com-
bined vertical + radial outwards) below foundation tip level. The major
unknown is the mass coefficient of consolidation value.

11.2.4 Secondary compression (creep) settlement

For fine-grained soils, in the absence of project-specific data, a reasonable
value for secondary compression index Cα (slope of voids ratio e versus log
time) is given by Cα/Cc ≈ 0.04, where Cc is the compression index (Terzaghi
et al., 1996).

11.2.5 Settlement induced by cyclic loads

Janbu (1985) provides a (simple neat) cumulative strain model, while
O’Riordan (1991) provide a useful review of the topic.
248 In-place response

11.2.6 Regional settlement
Causes of regional settlement include (most common first):

• oil, gas and water extraction

• incomplete primary consolidation (caused by rapid sedimentation)
• ongoing tectonic movements.

11.3 DISPLACEMENT UNDER DYNAMIC AND CYCLIC

LOADS

Do not use either pile cyclic p-y data (ISO 19901-4:2016a Section 8.5, soil
reaction for piles under lateral loads) or elastic solutions for flexible piles for
intermediate foundation design.
For OWT intermediate foundations (aka monopiles) the main geotechni-
cal design challenge is that tilt should not exceed the allowable threshold.
Typical SLS criterion is that accumulated tilt at seafloor < 0.5°. A secondary
challenge is eigen frequency assessment of complete OWT system (i.e. foun-
dation, substructure, mast and rotor nacelle).
Monopile design requires a good knowledge of the soil structure interaction
at small displacements. Offshore wind turbines are commonly designed so
that their first natural frequency avoids frequencies of operational and envi-
ronmental loads, see Figure 11.2 (CFMS, 2019). Design also needs to ensure
that the permanent inclination of the foundation does not exceed a fixed limit
at the turbine end of life (the inclination typically should remain < 0.5°).
From back analysis of monitored data, Kallehave et al. (2012) found cur-
rent design models fail to reproduce the natural frequency although it is a
key parameter for the design. Most soil-pile stiffness models derived for
flexible piles tend to underestimate the natural frequency.
The method normally used to design piles under lateral and overturning
loads is based on a Winkler modelling approach, commonly termed the p-y
method. P-y curves proposed in standards API RP2A (API, 2000) or
DNV-GL RP C212, (DNV-GL, 2017, 2019) are generally regarded as being
satisfactory for jacket pile design but fail to provide reliable response for
large diameter piles at small displacements. Several semi-empirical
approaches have been proposed to stiffen the initial lateral response for
“large diameter” piles, notably:

• Kallehave et al. (2012), Sorensen et al. (2010) and Kirsch et al. (2014)
for sand
• Stevens and Audibert (1979) and Kirsch et al. (2014) for clay.

The initial response could also be calibrated on the basis of in situ measure-
ments such as High Pressure Dilatometer (HPD) tests, or using finite
In-place response 249

Figure 11.2 S chematic of environmental conditions to consider in modelling an

offshore wind turbine (CFMS, 2019).

element analyses. DNV-GL RP C212 (DNV-GL, 2017, 2019) propose to use

Stevens and Audibert (1979) to account for the “diameter effect” for piles
up to approximately 2.5 m but highlight that alternative methods such as
finite element (FE) analysis are needed for larger piles.
The PISA (Pile Soil Analysis) project (Byrne et al., 2020) proposed a new
approach based on the calibration of soil reaction functions (1D model)
using 3D FE analyses. The four soil reaction components included in the 1D
model are (see also Figure 11.3):

• lateral resistance represented by p-y curves

• shear strength at the base of the pile called “base shear”
• axial shear stress along the shaft called “distributed moment”
• a rotational resistance called “base moment”.

The contribution of the three last additional components start to be non-

negligible when the L/D ratio decreases and becomes less than 3.
Within the framework of the PISA project, the four elements of the ground
response have been quantified in dense sands (Burd et al., 2020) and stiff
clays (Byrne et al., 2020). The current modelling procedure is limited to
monotonic loading. Ongoing work is being carried out to extend the method
to cyclic loading (unpublished as of this writing).
250 In-place response

Figure 11.3 P
ISA model: (a) soil reaction components incorporated in the PISA
design model; (b) 1D FE model employed in the PISA analysis.

Various methods have been proposed to derive the accumulated displace-

ment (or rotation) under cyclic loading in sand (e.g. Lin and Liao, 1999;
LeBlanc et al., 2010; Abadie et al., 2017). The SOLCYP JIP proposed a
design approach for both sand and clay (SOLCYP 2017). Although, while
initially focused on slender piles, the global approach accounts for the pile
rigidity through a coefficient of rigidity, CR.

11.4 RESPONSE IN (WEAK) ROCK

Pile response in rock is most commonly determined through a local approach

where the pile is modelled using a Timoshenko beam element (the
Timoshenko beam model being more adequate for rigid piles than the Euler
beam model) and the soil is modelled as a discrete set of independent springs
along the length of the pile (i.e. springs operate in the horizontal plane, with
no interaction with the springs above or below. The PISA project (Burd
In-place response 251

Figure 11.4 Illustration of chalk-pile load deflection concept (Muir Wood et al.,
2015).

et al., 2020), highlighted that the so-called p-y method was originally devel-
oped for relatively slender piles and should not be sufficient to model the
soil structure interaction of rigid piles with L/D < 6. However, in the case of
weak rock, field experiences tend to show that properly calibrated p-y curves
can predict reasonable pile rotation.
In the case of weak carbonate rock, pile installation (see Section 9.13) can
lead to the creation of a crushed or remoulded zone around the pile. In that
case, the subsequent lateral response will be softer than if the pile was sur-
rounded by intact rock. To account for the presence of this remoulded zone,
Muir Wood et al. (2015) proposed to consider two springs in series
(Figure 11.4). The overall deflection of the pile (y) under a given load is
assumed to be the sum of the deflection of the remoulded chalk (yr) and the
deflection of the intact chalk (ys).
A similar concept was applied for driven and drilled and grouted piles in
calcarenite (Lovera, 2019) where the grout annulus, crushed zone and intact
rock were modelled as three springs in series (Figure 11.5).
The equivalent initial local stiffness can be deduced using

p Eoed
Eequiv (11.1)
y a grout Eoed crushed Eoed
Dsteel Egrout Dcrushed Grock

where:
For the grout,
252 In-place response

Figure 11.5 M
odel for pile response in weak rock with grout anulus, crushed
zone and intact rock modelled as three springs in series (Lovera
2019).

a parameter equals zero for a driven pile and 1 for drilled and grouted piles;

Dsteel is steel tubular diameter;

δgrout is grout annulus thickness;
Egrout represents shear modulus of the grout;

λ parameter is smaller than one, taking into account that the pressure does
not apply on the all circumference.

For the crushed/remoulded zone,

Dcrushed = Dsteel for driven pile and Dgrout drilled hole diameter for drilled and
grouted piles;
δcrushed is the crushed zone thickness;
Eoed is the oedometer modulus of the crushed zone material at low stresses.

For the rock,

Grock is the shear modulus of the rock.

Cyclic degradation under lateral loading is generally marginal for non-car-

bonate competent rocks. For weak carbonate rocks, experience shows that
significant accumulated displacements can be observed. These accumulated
displacements are due to both the compaction of the crushed zone and the
cyclic degradation of the rock mass. Based on observations made during
field tests (Palix and Lovera, 2020), Lovera et al. (2020) proposed a rheo-
logical model to predict the accumulation of cyclic displacements based on
an analogy with the creep phenomenon, illustrated in Figure 11.6.
At the element scale, deformations under cyclic loading and creep are
shown to share some similarities. Under both creep and cyclic loading, the
sample is weakened and can fail at a lower level of loading than under
In-place response 253

Figure 11.6 S ketch of the rheological model considered for piles driven in soft
rock (Lovera et al., 2020).

monotonic conditions. For a high maintained load (similarly for a high

maximum cyclic load), the strain accumulates faster and faster, leading
quickly to failure. For a low maintained load (similarly for a low maximum
cyclic load), the strain accumulates at the beginning and then stabilises.
The Lovera et al. (2020) model combines the non-linear static p-y curves
with Kelvin-Voigt models in series. A Kelvin-Voigt model involves two
parameters: a stiffness (E1) that controls the maximum accumulated dis-
placement when the stabilised state is reached and a viscosity η1 that con-
trols the rate at which the stabilised state is reached (the higher η1, the
slower the stabilised state is reached). For driven piles, two springs are con-
sidered: one for the crushed rock for which the accumulation of displace-
ment (y1) stabilises quickly (low value of η1) and one for the weak rock for
which the accumulation of displacement (y2) is more progressive (high value
of η2).

11.5 CLOSURE

This chapter has discussed in-place response and has shown that different
intermediate foundation types have their own challenges. Anchor founda-
tion movements are usually non-critical and are rarely assessed.
Support foundations may be conveniently subdivided into those for oil
and gas and renewable projects. The challenges are listed below in frequency
order (most commonly encountered first).
Support platform projects require 4 × 4 foundation stiffness matrices to
pass to the structural engineer. Simple (elastic) solutions have been listed to
evaluate the stiffness terms kzz (axial), kxx, kxθ, kθx, kθθ (lateral, coupled) and
kΨΨ (torsional). ISO/API type “p-y” curves are unsuitable for intermediate
foundation lateral response assessments due to the rigid body rotational
behaviour of intermediate foundations.
254 In-place response

Renewable support foundations, especially OWT monopiles, require

analysis of both tilt and Eigen frequencies.
Support platform projects with high mobilised axial V/Vmax values in
weak clay profiles may require analysis, particularly of primary consolida-
tion settlement magnitude. To this end, a reasonable two-spring model has
been presented.
The Doherty and Deeks (2003) continuum solutions are comprehensive,
covering a wide range of geometries and non-homogeneous elastic half-
space profiles routinely encountered in design. However, they are valid only
for embedment ratios L/D ≤ 2. It would be nice if their work could be
extended to cover higher embedment ratios as OWT monopiles have L/D
values up to 10.
Chapter 12

Miscellaneous design considerations

12.1 INTRODUCTION

An overview of the miscellaneous design considerations for intermediate

foundations discussed in this final chapter are shown in Figure 12.1. Some
are non-geotechnical, reflecting the necessary interaction between other off-
shore engineering disciplines (structural, installation, hydraulic, etc.).

Figure 12.1 Miscellaneous design considerations for intermediate foundations.

255
256 Miscellaneous design considerations

12.2 SCOUR PROTECTION

The in-place resistance of intermediate foundations with low embedment

ratios (less than around 2) in sand is generally sensitive to seafloor scour. If
scour is a potential problem, then scour protection is generally recom-
mended as an alternative to regular seafloor/foundation inspections.
Protection types include frond mats, rock fill, sandbags, or fences and diver-
sion berms.
If there is a significant time lag between foundation and scour protection
installation, additional in-place analyses are normally performed of the
unprotected foundation to verify that factored design loads do not exceed
the factored soil resistance. Such analyses use a specified scour geometry, for
which soil strength values outside and below the foundation may be reduced,
plus environmental VHMT loads and load factors γL appropriate to the time
period considered.
In lieu of project-specific data, the general and local scour default recom-
mendations given for pipe pile foundations in sand (ISO 19901-4:2016a,
Section A.8.5.2), namely local scour depth equal to 1.5 D and 6 D overbur-
den reduction depth, are inappropriate for intermediate foundations.
Whitehouse et al. (2011) review scour development around monopiles in
sand for nine offshore sites and show that scour protection slightly decreases
scour depth and puts it further away from the pile (compare their Figures 2a
and 2b). DNV (2014a) gives general comments on scour around monopiles
and equations to calculate local scour in their Appendix J. At the time of
writing there are no publications on scour around large diameter founda-
tions (e.g. D > 7.5 m).

12.3 INTERACTION WITH ADJACENT INFRASTRUCTURE

12.3.1 Introduction
Intermediate foundations may interact with the following adjacent infra-
structure (the most important are listed first):

• jack-up spudcans
• adjacent intermediate foundations
• adjacent shallow foundations and pipelines
• conductor wells.

The remainder of this section gives information about interaction between

intermediate support foundations and other adjacent infrastructure, located
either on or beneath the seafloor. At the time of writing, only one publica-
tion in the public domain covers interaction of intermediate foundations
with jack-up spudcans (Alderlieste and Dekker, 2018). Hence, most infor-
mation is extrapolated from other studies.
Miscellaneous design considerations 257

12.3.2 Jack-up spudcans
Minimal-facilities hydrocarbons platforms normally require subsequent loca-
tion of a jack-up rig to install well conductors. Similarly, offshore wind tur-
bine monopiles or tripod substructures may be installed using a jack-up rig
fitted with spudcans. Figure 12.2 shows examples. Hence there may be a risk
of overstressing platform foundations. Such situations require consideration
of interaction between the jack-up spudcan(s) and platform foundations, par-
ticularly in weak soil profiles. Geotechnical analysis involves assessment of
soil movements induced by spudcans and the effects of these movements on

(a)

(b)

Figure 12.2 E xamples of jack-up spudcan – Intermediate foundation interaction:

(a) Work-over jack-up and Ophir well head platform (Alderlieste and
Dekker, 2018) and (b): installation jack-up vessel (Courtesy Fred.
Olsen Windcarrier).
258 Miscellaneous design considerations

the foundation, both lateral and axial (Mirza et al., 1988; Siciliano et al.,
1990). Normal practice for pipe pile foundations in normally consolidated
clay profiles is to ignore interaction effects provided that the closest soil gap
distance is at least one spudcan diameter (ISO 19905-1:2016b).
However, for intermediate foundations, this gap distance (1 spudcan
diameter D) should be critically reviewed. This is because the foundations
are essentially rigid: any load due to lateral soil displacement is transferred
directly into the platform leg, and not partially absorbed by pile bending.
A key issue is predicting lateral soil movements induced by the spudcan at
the intermediate foundation. These movements have both horizontal and
rotational components. Due to smaller foundation embedment depths (than
piles), rotational effects are usually more critical than lateral effects –
­foundation rotations could induce additional BM/overstress the platform
leg. This situation is similar to kicking one leg of a three-legged stool.
Jack-up spudcan–foundation interaction analysis is difficult. This is
because the accuracy of any prediction is a function of the accuracy of both
the data and the model (Lambe, 1973), and larger-than-usual inaccuracies
generally exist both for data (e.g. soil shear strength profile, spudcan geom-
etry and preload, foundation geometry and gap distance) and the model (i.e.
spudcan penetration tip depth and resulting lateral soil movements around
the spudcan).
Spudcan installation induces vertical foundation movements. However,
industry perception (for pile foundations) is that axial (downdrag/upthrust) is
not so critical as lateral. The majority of axial pile resistance is derived from
soil well below the spudcan and is rarely (possibly never) studied in detail.
Hence, instead of complex 3D large strain Finite Element Analyses, a rea-
sonable design approach may be to:

• assign cautious parameter values for data

• assess Low, Best and High Estimate foundation “p-y” data, either from
analytical equations (Section.11.2) or from 3D Finite Element Analyses
• assess cautious (high) spudcan penetration and corresponding lateral
soil movement profile at intermediate foundation centreline
• offset foundation p-y data using lateral soil movement profile
• use offset intermediate foundation stiffness matrices in platform struc-
tural analyses.

Note that care has to be taken with orientating foundation p-y data. This is
because the offsets need to be applied along a line connecting the spudcan
and the jacket leg(s), whereas the platform structural model automatically
rotates these to coincide with the storm/wind direction. However, since
spudcan penetration is unlikely to occur during a “small” (10-year summer
storm) event, then orientation may not be critical.
Further information on this approach is given by Xie (2009) and Xie et al.
(2017) on lateral soil movement data, by Tho et al. (2013 and 2015) on
Miscellaneous design considerations 259

decoupled y-shift method, by Dekker (2014) on foundation stiffness matrix

using two lateral springs and by DNV-OS-J101 (DNV, 2014b), which dep-
recates pipe pile p-y curves and suggests using FEA instead.
Note that, unlike piles, additional caisson wall thickness increases are
unnecessary – the caisson is almost rigid (and hence capable of withstanding
high BM) and lateral soil pressure increases are small (e.g. < 10 kPa).
Nevertheless, one should verify that foundations can withstand additional
soil pressures.
Like spudcan-pile interaction, it is likely that spudcan-intermediate foun-
dation interaction in competent soils is not usually a major concern. However,
in weak soil, where deep spudcan penetration occurs (≥ 0.5 D, say), then
detailed jack-up spudcan – foundation interaction analysis may lead to
considering:

• increased platform leg stiffening/braces

• increased foundation wall thickness
• minimizing spudcan diameter and penetration
• maximizing spudcan – foundation clearance
• using a mat supported (instead of an independent leg) MODU (mobile
offshore drilling unit)
• using the jack-up rig to install the platform (i.e. install foundations
into pre-shifted soil)
• pile (instead of intermediate) foundations.

Another issue is possible long-term intermediate foundation lateral resistance

decrease due to the presence of remoulded soil/footprint. This was studied by
Stewart (2005) for spudcan–pile (not intermediate) foundation interaction.
Alderlieste and Dekker (2018) have described detailed jack-up–suction
pile interaction design studies for the Ophir Well Head Platform (WHP),
offshore Malaysia. Due to the weak soil profile (Figure 12.3), and the high
spudcan preload, penetration depths between 7 m and 15 m were possible.
For the design scenario (Figure 12.4), these resulted in “free-field” soil lat-
eral displacements of 0.12 m and 0.19 m at seafloor and suction pile tip level
respectively. Even though the soil gap spacing (1.1 D) exceeded the ISO
19905-1:2016 recommendation, these displacements were sufficiently large
to necessitate detailed studies. Instead of using shifted p-y curves, a more
rational (less complicated) approach was used to obtain the suction pile loads
and displacements. Figure 12.5 shows WHP displacements: after installing
the starboard spudcan, only the closest suction pile undergoes “free-field”
soil lateral displacements (the other two suction piles are too far away), and
jacket rotation in plan occurs. As expected, symmetrical pile displacements
occur after installing the second (port) spudcan. The back suction pile fur-
thest away is the most highly loaded, since it restrains jacket movements. No
additional displacements are induced by the bow spudcan. Based on their
results, they concluded that 1 D minimum soil gap spacing is adequate.
260 Miscellaneous design considerations

Figure 12.3 O
phir well head platform, soil properties (Alderlieste and Dekker,
2018).

Figure 12.4 T
op view (left) and side view (right) of work-over jack-up rig at
Ophir well head platform. Suction pile diameter D = 6 m, target
penetration L = 13.7 m. Spudcan diameter D = 17 m, spudcan design
penetration h = 12.75 m, minimum spudcan-suction pile clear soil
gap S = 19.4 m (Alderlieste and Dekker, 2018).

Figure 12.5 S uction pile displacements due to spudcan penetration (not to scale)
at Ophir well head platform. Left: first spudcan installation. Right:
due to second spudcan installation (Alderlieste and Dekker, 2018).
Miscellaneous design considerations 261

12.3.3 Adjacent intermediate foundations

In-place resistance and response both should consider the possibility of
interaction with any adjacent intermediate foundations.
Considering block failure in clay, design failure zones should not intersect.
Hence, a general rule of thumb for undrained soil response using 45 degree
wedges is a clearance of at least 1.5 times the foundation embedded length.
Because of curvature of the passive failure surface, this clearance value
should be increased for sands.
Finite element analysis is advisable to capture the effect of complexities in
stratigraphy or geometry.

12.3.4 Adjacent shallow foundations and pipelines

In-place resistance and response of intermediate foundations should both
consider the possibility of interaction with any adjacent shallow founda-
tions and pipelines.
Interactions are discussed in Engin et al. (2015) for jack-up–mudmat (sea-
floor template), Versteele et al. (2015) for mat-supported MODU–pipeline,
and Kellezi et al. (2005) for spudcan-pipeline. Since an intermediate founda-
tion installation displaces less soil (in the order 10m3) than a jack-up spud-
can (in the order 150 m3), interaction effects on a pipeline are expected to be
less from an intermediate foundation than from a spudcan (at the same soil
gap spacing).

12.3.5 Conductor wells
Conductor installation techniques should consider the possibility of soil
removal (wash-out) occurring in sand and appropriate measures (e.g. using
driven isolation casings) taken. This is to prevent loss of resistance at nearby
intermediate foundations.
Intermediate foundations also interact with the conductor wells. Generally,
the only potential problem is that, when laterally loaded, intermediate foun-
dations may induce additional displacements in the conductors. However,
due to their flexibility, they will probably accommodate the extra lateral soil
movement without significant bending moment increases.

12.4 SHALLOW GAS

The possibility of shallow/dissolved gas adversely affecting in-place resis-

tance should be considered.
Geotechnical design of intermediate suction foundations where dissolved
gas is present in clay includes the following considerations:

• Site investigation: gassy soils are difficult to sample and test in the
laboratory. Therefore, site investigations should include in-situ tests.
262 Miscellaneous design considerations

• Installation: since such soils are “softer”, and the vertical stress reduc-
tion may cause gas to come out of solution, more heave than usual
may be anticipated. However, this may be offset by the fact that instal-
lation underpressures (i.e. tensile forces on the soil plug) are likely to
be low in normally consolidated soft clays.
• In-place response: in time, gas migration could create a gas “bell” under
the foundation top plate, causing an upwards force. This potential
problem is removed by providing open vents in the top plate. Assuming
vents are present, foundation compressive resistance can still accom-
modate bearing pressures under the foundation top plate. Maximum
axial tensile pull-out resistance, Vmax, cannot rely upon “passive suc-
tion” (since the top plate vent is permanently open). Instead, pull-out
should be based on outer friction plus the minimum of (inner friction
or soil plug weight) plus foundation submerged weight only, i.e. the
“coring” and “leaking” failure models shown in Figure 10.1.

12.5 PERMANENT PASSIVE SUCTION

The use of permanent underpressure may be considered in order to increase

foundation axial tensile resistance. This is particularly for sand profiles and
support foundation platforms (underpressure can be applied by reducing
water level within legs). Permanent underpressure (“passive suction”)
removes some of the uncertainties associated with temporary passive suc-
tion generated under the top plate discussed in Section 10.5.3 (on undrained
“sand” soil response). The top plate vent should be open to allow the pump
to remove water entering the soil plug. Pump design and seafloor outflow
rate are discussed in Section 12.7. The downside is that underpressure has
to be available, either permanently or during storm loading, throughout the
lifetime of the structure.

12.5.1 Example – permanent passive suction

An example assessment of the additional compressive force (∆V) and flow
rate Q (assuming maximum underpressure) has been made for the following
situation: caisson D = 10 m, L = 7.5 m, high estimate of sand permeability
k = 2e-4 m/s, with a low estimate of submerged unit weight, γsub= 9 kN/m3.
From Figure 9.16 (on comparison critical suction pressures), a low estimate
of ∆ucrit/(γsubD) value is 1.0, giving ∆ucrit = 90 kPa, corresponding to ∆V =
(πD2/4)* ∆ucrit = 7 MN. The flow rate Q can be found using Figure 9.5b and
the equation Q = (k D ∆u/γsub). Dimensionless factor F is a function of (L/D)
and kr, the soil plug: soil mass permeability ratio. A high estimate F value is
2.5, giving a (high estimate) Q value of 45 litres/s. For comparison, filling a
bath requires approximately 0.6 litres/s, and a 100% redundant suction
Miscellaneous design considerations 263

pump system can provide around 80 litres/s. Both ∆V and Q are considered
reasonable. Better (and hopefully lower) Q estimates may be possible by
careful examination of suction pump instrumentation data (penetration rate
and pump outflow rate) during installation, especially at/near final penetra-
tion, where steady state conditions are most likely to have been achieved,
and the water plug volume decrease is small: subtracting the water plug
volume decrease from the pump outflow gives Q, the amount of water enter-
ing the soil plug.
Laboratory model tests on “vacuum anchors” have been reported by sev-
eral researchers during the 1980s, including Goodman et al. (1961), Brown
and Nacci (1971), Wang et al. (1975, 1977, 1978) and Helfrich et al. (1976).
Wang et al. (1977) give an equation for (vertical) breakout capacity.
Permanent underpressure is currently not a relied-on solution.

12.6 TOP PLATE VENT DESIGN

Vent(s) in the top plate are required to expel the “water plug” (the volume
of water trapped between the top plate and seafloor) during the self-weight
penetration phase. The vent outlet pipes should be of sufficient diameter to
prevent high overpressures occurring in the “water plug” for the specified
range of penetration velocities. If water pressures are high, they may cause
piping along the foundation perimeter (sand profile) or base failure (clay soil
profile). If the foundation design relies on in-place “passive suction”, or
extraction using overpressure is envisaged, then vent design should incorpo-
rate a valve. The valve has to be open during installation, and subsequently
closed. This is to ensure a permanent seal.
For intermediate foundations installed by crane, an additional vent design
consideration may be while the foundation is being lowered through the
wave splash zone: vents should also be sufficiently large to allow essentially
free air egress/ingress and hence avoid “snatching” (large variations in lift-
ing wire loads).
The basic equation for incompressible fluid flow through a single vent
(nozzle) is given by the equation:

v n Cd 2 g H (12.1)

where

vn = vent (nozzle) velocity [m/s]

Cd = vent (nozzle) discharge coefficient [–]
g = acceleration due to gravity [m/s2]
∆H = vent (nozzle) differential head loss [m]
264 Miscellaneous design considerations

Hence, for a single vent, the overpressure increase, ∆u, and corresponding
vent (nozzle) resistance Rn, are given by:

u H water kPa (12.2)

R n u Ai – An kN (12.3)

where
∆H = vent (nozzle) differential head loss [m]
= [1/(2 g)] (vn2/Cd2)
γwater = fluid (sea water) density (= 10.05) [kN/m3]
Ai = pile internal area (π Di2/4) [m2]
An = vent (nozzle) area [m2]
g = acceleration due to gravity [m/s2]
vn = vent (nozzle) velocity [m/s]
= vpile Ai/An
vpile = foundation velocity [m/s]
Cd = vent (nozzle) discharge coefficient [–]

Equation (12.2) can also be used to assess the vent (nozzle) area An
required to prevent the overpressure ∆u causing base end-bearing (“push-
out”) failure. In addition to the foundation velocity Vpile, top plate vent area
An results are sensitive to the coefficient of discharge, Cd. The minimum Cd
value should be taken to maximise vent (nozzle) resistance Rn. If the vent is
sharp-edged and circular, the “ideal” Cd value is 0.611 (= π/(π + 2)). Hence a
cautious value is usually taken to be 0.61. Since entry into the seafloor is
usually the most critical stage, then it is usually sufficient to analyse this situ-
ation only. For controlled lowering, a reasonable value for foundation
touch-down velocity, vpile, is 0.2 m/s (see Section 9.9).

12.6.1 Example – top plate vent design

As an example of the use of Equations (12.2) and (12.3), verify that the vent
design (2 number 1.0 m diameter, each with area 0.8 m2) will not cause end-
bearing failure at seafloor. The relevant data, which are similar to those used
in the Section 9.9.3 (free-fall penetration) Case A example, are as follows:

D = 5.0 m, WT = 25 mm (Ai = 19.2 m2)

Nc,circle = 6
su,0 = 2 kPa
FOS = 1.5
Miscellaneous design considerations 265

seafloor velocity v0 = 0.2 m/s

two vents, diameter = 1.0 m
An = 0.8 m2
nnozzles = 2
vent (nozzle) discharge coefficient Cd = 0.61

Actual and allowable overpressures ∆u and ∆uall are evaluated as follows:

vent (nozzle) velocity vn = 2.41 m/s (= vpile Ai/(Av nnozzles))

vent differential head loss ∆H = 0.79 m (= 1/(2 g)][vn2/Cd2])
actual overpressure ∆u = 7.8 kPa (= ∆H γwater)
maximum allowable overpressure ∆uall = 8 kPa (= Nc,circle su,0/FOS).

Based on these results, the vent design is satisfactory (∆u < ∆uall).
As stated earlier, the seafloor situation is assumed to be the most oner-
ous. More complicated calculations are necessary should velocity increase
with depth below seafloor, for example due to free-fall caused by uncon-
trolled lowering (Section 9.9.3, free-fall penetration, and Figure 9.8, free
fall velocity – depth curve). In such cases, the Rn, the corresponding vent
(nozzle) resistance, has to be computed, and interested engineers may care
to check that Rn is 143 kN when vpile = 0.2 m/s.

12.7 SUCTION PUMP DESIGN

During installation, the suction pump maximum achievable flow rate should
be sufficient to discharge both the “water plug” volume plus the anticipated
outflow rate at seafloor. This is especially important in sands and gravels of
high permeability and large diameter foundations (e.g. > 10 m).
Furthermore, the suction pump used during installation should be capa-
ble of generating adequate underpressure/overpressure.
To calculate the “water plug” flow rate component, a penetration velocity
of 1 mm/s is generally used. This is a factor 20 slower than a CPT. Houlsby
and Byrne (2005) give an equation for assessing seafloor outflow rate due to
underpressure. A high estimate of soil permeability should be used.

12.7.1 Example – suction pump flow rate

A flow rate Q assessment is made for the example previously studied in
Section 12.5 (permanent passive suction). Relevant Q data are: caisson D =
10 m, WT = D/200 = 50 mm, L = 7.5 m, sand permeability k = 2e-4 m/s,
with a high estimate of seafloor outflow Qseafloor = 45 litres/s.
266 Miscellaneous design considerations

Inner diameter Di = D − 2 WT = 9.9 m

Water plug area Ai = π Di2/4 = 77.0 m2
Assume penetration velocity v = 0.001 m/s
Pump flow rate to discharge water plug Qplug = v Ai = 0.077 m3/s = 77 litres/s
Pump flow rate to discharge seafloor flow Qseafloor = 45 litres/s
Required minimum Q = 77 litres/s + 45 litres/s = 122 litres/s (440 m3/hour)

12.8 FOUNDATION INSTRUMENTATION

Provided that they are of sufficiently high quality, back-analyses of suction

foundation installation records assist in model refinement/calibration/
extension. Examples include Colliard and Wallerand (2008) and Chatziva­
sileiou (2014). In general, particularly for installations in sand, underpres-
sure data should be reliable: there should be no concerns about possible
friction losses in the exhaust system between the top plate and the point
where pressures were measured. In addition, if the preload value is large,
then it has to be accurately known. In this respect, a difficulty for braced
jackets (where all foundations are installed at the same time) is to assess the
preload taken by each foundation: usually only crane hook load data are
available.
Installation foundation instrumentation usually comprises measurements
of:

• penetration
• ambient water pressures acting externally on the top plate and inter-
nally (either under the top plate or at the suction outlet pipe, rather
than at the pump, in order to avoid errors due to pipe friction and
Venturi effects)
• soil plug heave
• inclination and north-seeking gyroscope (intermediate anchor
foundations)
• position (x, y, z co-ordinates)
• preload.

Preload measurements should be made both in air and under water using a
calibrated load cell mounted between the crane wire and the hook. The
foundation steel surface condition (both inside and outside) should also be
documented.
Long-term foundation instrumentation is a growing area for offshore
wind farms, where extensive monitoring systems, both short and long term,
are being used to optimize foundation design (e.g. Devriendt et al., 2013;
Byrne et al., 2015a, 2015b; Shonberg et al., 2017). In general, instrumenta-
tion should also be considered to improve geotechnical knowledge about
foundation response during storm loading (Svanø et al., 1997).
Miscellaneous design considerations 267

12.9 STEEL DESIGN

Structural steel designers should also check the foundation geometry. Critical
items normally include stress concentrations and fatigue in the vicinity of
the load application point, and buckling of the steel cylinder and top plate
under applied underpressure/overpressure.
In addition, foundation steel tip integrity assessments (e.g. Aldridge et al.,
2005; Erbrich et al., 2010) should be considered if there is a risk of the foun-
dation tip locally encountering an isolated gravel (or cobble or boulder-
sized) fragment (see Section 9.10.5), and this fragment cannot be fractured
(i.e. split longitudinally) or pushed to one side (i.e. bearing capacity failure
of the soil surrounding the fragment).
If required, like soil reaction stresses (Section 12.10), steel stresses may be
assessed by 3D non-linear FEA. However, the steel response is non-linear,
whereas it is generally sufficient to assume that the soil remains elastic. Usually
structural analysis software (SACS, Nastran, SESAM, ANSYS, etc.) is used
instead of geotechnical analysis software (e.g. PLAXIS or OptumG2/G3) to
determine steel stresses. Soil continuum elements (not a bed of Winkler springs)
should be placed both inside and outside the foundation. If the complete inter-
mediate foundation is being modelled, then a finite element mesh extending 3L
radially and 2L vertically should be sufficiently accurate (Kuhlemeyer, 1979). If
infinite elements are used, then the finite mesh radial extent can be reduced to
2L (Kay and Palix, 2010). Table 12.1 provides reasonable soil E and ν values.
For model verification, mesh nodes have 6 DOF (loads Fx, Fy, Fz, Mx, My
and Mz and displacements δx, δy, δz, φx, θy and Ψz), Figure 3.1a gives their
sign convention, and the RP is at seafloor on the foundation centreline.
Assuming lateral loading is dominant, a simple way to verify the model is to
run two load cases. The first applies a “fixed head” unit lateral displacement
(δx = 1 mm, remaining 5 DOF = 0) to the RP to find pile head lateral seafloor
stiffness matrix terms kxx (= Fx/ δx) and kxθ (= My/δx). The second load case
finds the missing two terms kθx (= Fx/θy) and kθθ (= My/θy) by applying “fixed
head” unit rotation (θy = 1 milliradian, remaining 5 DOF = 0). Off-diagonal
terms kxθ and kθx should be essentially identical. All four lateral stiffness
matrix terms can then be compared with solutions in Table 11.2 (Elastic
Continuum Solutions – Static Load).

12.10 SOIL REACTIONS

Soil reaction stresses are needed by structural steel designers as input to

structural design of suction foundations. Usually, the following three phases
need to be considered:

• installation and retrieval

• operation (in-place)
• removal.
268 Miscellaneous design considerations

Table 12.1 Soil Parameter Values for 3D FEA of Foundation Steel

Soil Type Typical Values Reference(s)

Clay E = 200 s u Poulos and Randolph
ν = 0.49 (1983) a
Sand E = 4 q c (NC) and 10 q c Lunne et al. (1997)
(OC)
ν = 0.25
Rock E = 350 σ c Attewell and Farmer (1976)
ν = 0.1

Notation:
E = Young’s Modulus
ν = Poisson’s Ratio
s u = soil undrained shear strength
q c = CPT cone tip resistance
σ c = unconfined compressive strength
Note:
a For laterally loaded piles.

However, during installation, retrieval and removal phases, general experi-

ence is that suction/excess water pressures govern steel stresses, and hence
soil reaction stresses are not critical. In such cases, they are derived only for
the operational phase.
If required, soil reaction stresses are normally assessed by 3D non-linear
FEA. This technique has largely replaced the use of simplified analytical
solutions, considering separate stress distributions due to different load
components. The separate components are then combined to give design soil
reaction stresses. Examples are given by NGI (1997).

12.11 UNDERBASE GROUTING

At the end of installation, a water-filled void may exist beneath the top plate.
This void generally occurs in sand profiles. The presence of a void is gener-
ally due to using a high estimate stickup value, primarily to ensure that the
target depth could be successfully reached. It is usually not practical to
embed the foundation beyond its target depth. Reasons include (a) the void
may be non-uniform and the (sand) plug may be in contact with the top
plate underside and/or internal stiffeners. In such cases, additional penetra-
tion may induce high underpressures and steel stresses, and (b) the jacket
will be (slightly) lower than intended, which may cause issues with pipelines
and other connections.
Because of the limited foundation embedment ratio (usually less than
one), there is a perceived risk of additional axial settlement occurring
during cyclic storm loads. Hence, underbase grouting is generally applied.
The objective is to replace the water in the void with cement grout.
Miscellaneous design considerations 269

A consequence is that such foundations cannot be removed by the appli-

cation of overpressure (unless a hole is drilled through the cement grout
into the sand plug prior to applying overpressure).
After curing, the cement grout should ideally be of similar stiffness and
strength to the in-situ seafloor soil. If compartments (or inverted T-beam
stiffeners under the top plate) exist, and there is only single grout entry and
exit vents, then “rat-holes” or similar must be provided in order to minimise
the risk of not completely filling the water-filled void. These considerations
are similar to those for large gravity base foundations (i.e. shallow founda-
tions) given in ISO 19903 (ISO 2019).

12.12 ANCHOR CHAIN TRENCHING

Anchor chain trenching of taut and semi-taut mooring systems is a rela-

tively recent observation. Bhattacharjee et al. (2014) described wire rope
and chain chafing integrity issues for the Sepentina FPSO taut anchoring
mooring system, offshore Equatorial Guinea in 475 m water depth. More
importantly, anchor chain trenches of varying dimensions were subse-
quently revealed in front of all nine suction piles; see Figure 6.6. Based on
soil loss, the piles were considered unfit for service. The replacement was a
(temporary) catenary mooring system comprising OMNI-Max free-fall
anchors.
Bhattacharjee et al. (2014) provided the first mention in the public domain
of anchor chain trenching. It is understood that similar trenches, some of
which reached lug level, have also been observed in weak NC clays in other
parts of the world, including the Gulf of Mexico and South-East Asia, for
taut and semi-taut mooring systems.
Anchor chain trenching 3D FEA by Hernandez-Martinez et al. (2015)
considered a typical Gulf of Guinea (GoG) soil profile, namely su,C = 2.4 +
1.57 z [kPa, m] and γsub = 3.5 kN/m3. The anchor geometry was D = 5 m, L
= 20 m, with zlug at 13.3 m bsf (i.e. 2/3 L) on the anchor outer surface. In
plan, this was simplified to an equivalent rectangular area (3.92 m by 5 m)
having the same cross-sectional plan area (i.e. 19.6 m2) as the actual anchor.
The chain trench geometry was “brick”-shaped – 5 m (1 D) wide down to
lug level in front of the anchor and with a flat trench base. For typical taut
anchoring systems, θlug typically varies between 40° and 20°, for which they
obtained anchor holding capacity reductions of 20% and 45% respectively,
as shown in Figure 12.6.
Arslan et al. (2015) also considered a rectangular trench with the same
width as the foundation, but in front of a 5 m diameter circular foundation,
and two embedded lengths, L = 14 m and 16 m. The generic soil profile was
su,z = 2 + 1.5 z [kPa, m]. The corresponding soil γsub value was not reported.
Their limited parametric study indicated that similar capacity reductions
(20% to 40%) are likely.
270 Miscellaneous design considerations

Figure 12.6 A
nchor chain trench 1D wide. V lug -H lug envelopes (Hernandez-
Martinez et al., 2015).

Alderlieste et al. (2016) adopted an improved approach, using 3D FEA of

a 2 m wide sloping trench in front of a circular 6 m diameter, 19.35 m embed-
ded length suction pile, for detailed design of a taut mooring anchor system
in Gulf of Guinea clay. Figure 12.7 summarizes their Vlug-Hlug capacity enve-
lopes, together with factored ALS loads. It is seen that there is a 4% reduc-
tion for ultimate vertical capacity, and 20% reduction for ultimate horizontal
capacity. The latter reduction is significant, but decreases to approximately
7% for the design load angles, which were at least 30o.
Figure 12.8 shows PLAXIS 3D incremental shear strain results for the
“fixed head” condition, and the corresponding interpreted failure mecha-
nism. The main feature is a cylindrical rotational failure mechanism, resem-
bling a pair of curved, partly closed circular elevator doors, extending from
seafloor down to the sloping trench base. Below this, there is a passive wedge
failure of limited extent, starting at pile tip level and exiting on the sloping
trench base.
Sassi et al. (2017) conducted 1 g and centrifuge tests of chain-trench-soil
interaction in both kaolin and remoulded Gulf of Guinea clay in order to
provide a conceptual framework for anchor chain trenching. In a compan-
ion paper, Versteele et al. (2017) give details of proprietary CASCI software,
an extended 2D chain-soil model with erosion, trench and tunnel stability
mechanisms.
New field evidence was presented in a pair of OTC papers by Colliat et al.
(2018) and Sassi et al. (2018). Both papers are Gulf of Guinea soil and met-
ocean specific. Colliat et al. (2018) present and discuss anchor trench
bathymetry data (width, depth and anchor-to-trench distances) for three
Miscellaneous design considerations 271

Figure 12.7 A
nchor chain trench D/3 wide. V lug -H lug envelopes and governing ALS
loads (Alderlieste et al., 2016).

TOTAL projects, each with an FPSO and off-loading terminal (OLT) buoy
taut line mooring system. Instead of vertical sided trenches modelled previ-
ously, due to erosion and slope instability, the trenches are wider at the top
than the base, with slope angles increasing to near-vertical for ≈ 3 m above
the base. Figure 12.9a shows a cross-section 20 m away from an anchor pile
(D ≈ 4.5 m, L ≈ 17 m).
Sassi et al. (2018) performed preliminary small scale (75 g–80 g) centri-
fuge tests on D = 4.5 m, L = 15 m, zlug = 9.5 m suction piles jacked into either
blocks extracted from undisturbed core samples (su = 1 + 1.4 z), or reconsti-
tuted clay (slightly higher su profile, gradients between 1.4 kPa/m and 2.0
kPa/m. A spatula, base width = D = 4.5 m (Figure 12.9b) excavated large
trenches reaching the suction anchor and lug level depth. From the first series
of tests (pure vertical pull-out; to model no reverse end-bearing (REB) the
anchor tip was in contact with a sand layer), the presence of a trench had
only a minor reduction (as expected). The “with REB” Qpl (= Fo + Qbase) and
“without REB” Qun (= Fo + Fi) resistances were ≈ 3.0 MN and ≈1.6 MN
respectively. Back-figured average α and Nc,circle values were low (0.3, 8).
272 Miscellaneous design considerations

(a)

(b)

Figure 12.8 A
nchor chain trench – “Fixed Head” incremental plastic strains and
failure mechanism (Alderlieste et al., 2016).

For the remaining two test series, soil pore pressure data were obtained at ≈
0.2 D, 1.2 D and 1.4 D below the tip, and these data confirmed that passive
suction (i.e. REB) was developed even when trenches were present. Anchor
holding capacity reductions of 11% were obtained for θlug between 30o and
40o. Surprisingly, this value is less than FEA by Hernandez-Martinez et al.
(2015) and Arslan et al. (2015) (20% and 45%) who had 1D wide trenches
and included full REB, but similar to that of Alderlieste et al. (2016) (≈ 7%)
who had a smaller trench width (D/3) and excluded REB (i.e. they had a
”plugged” soil response). 3D FEA predictions type C1 (Lambe, 1973) were
also made.
Miscellaneous design considerations 273

(a)

(b)

Figure 12.9 O
bserved anchor chain trench geometry in Gulf of Guinea clay
(Colliat et al., 2018; Sassi et al., 2018).

Feng et al. (2019) presented results of a parametric 3D finite-element

analyses investigation examining the geotechnical capacity of suction cais-
sons in a trenching seabed with soil properties and caisson geometry reflect-
ing the Serpentina field case. The results show that the reduction in the
geotechnical capacity becomes more significant with increasing trench width
due to the loss of soil support and a change in failure mechanism as the cais-
son rotates into the trench. For a given trench width, the reduction in capac-
ity becomes more significant as the load inclination angle to the horizontal
decreases. They show that a strategy to design for inevitable trenching by
moving the lug shallower to reduce the depth of trench formation is not
straightforward. The gain from a shallower trench may often be outweighed
by the reduction in capacity from rotation of the caisson at failure for load-
ing angles typical of taut moorings.
274 Miscellaneous design considerations

12.12.1 Commentary
Reverse end-bearing capacity accounts for ≈ 50% of anchor axial pull-out
resistance. Excluding reverse end-bearing capacity adds something of the
order of 2 m length for a 5 m diameter anchor pile (see the Fixed Head Pile
Example in Section 10.7).
To date, most of the aforementioned work has been confined to the Gulf
of Guinea (GoG). It is hoped that this will eventually extend to other areas
of the world where anchor chain trenching has been observed.
One possible explanation for vertical trenches is that the Gulf of Guinea
clays may be similar to those at Bothkennar UK. Both have strong in-situ
structure, higher than usual sensitivity, and are lightly overconsolidated.
Bothkennar soils have been subject to bioturbation by worms and exhibit
rapid strain-softening due to destructuring (Atkinson et al., 1992; ICE,
1992). It is possible that vane shear test, T-bar and CPT testing all underes-
timate Gulf of Guinea clay peak undrained shear strength, su,peak. This may
explain why vertical trench heights have been observed exceeding those
given by classical wedge theory. (Mini) pressuremeter testing, plus a strain-
softening material model (Kay and Van Woensel, 2002) may be useful.
Mitigation measures for anchor chain trenching include:

• replacing taut wire by a semi-taut or catenary anchoring system

• making the pile diameter as large as practicable to minimise capacity
decrease
• putting the lug further below optimum lug level (say 1 m instead of 0.5 m)
to decrease the risk of forward rotation
• designing to exclude reverse end-bearing in order to account for the
possible risk of water ingress (i.e. adopt the Figure 10.1 “leaking” fail-
ure mode).

12.13 CLOSURE

This ultimate chapter has covered miscellaneous geotechnical design consid-

erations. Purely structural considerations have been omitted, for example
pile tip integrity and design of offshore wind turbine monopile transition
piece connections.
The breadth of the topics, which is greater than for either shallow or deep
foundations, emphasizes the point made in Section 8.1.1 that “intermediate
foundation design usually requires close cooperation between geotechnical
and other offshore disciplines, particularly structural engineers”.
Acronyms

AG Fugro Advanced Geomechanics P/L, Australia

aka Also Known As
ALS Accidental Limit State
API American Petroleum Institute, USA
ASTM American Society for Testing and Materials, USA
BE Best Estimate
bsf Below Sea Floor
CD Consolidated Drained
CEL Coupled Eulerian Lagrangian
CEN Comité Européen de Normalisation
COFS Centre for Offshore Foundation Systems, University of Western
Australia (UWA)
COV Coefficient of Variation
CPT Cone Penetration Test
CPTu Cone Penetration Test with pore pressure measurement
CU Consolidated Undrained
DNV-GL Det Norske Veritas – Germanischer Lloyd A/S, Norway
DOF Degree of Freedom
DSS Direct Simple Shear
EDF-EN Electricité de France – Energies Nouvelles, France
FEA Finite Element Analysis
FEBV Fugro Engineers BV, the Netherlands
FLS Fatigue Limit State
FOS Factor of Safety
FPSO Floating Production Storage and Off-loading
GoG Gulf of Guinea
GoM Gulf of Mexico
HE High Estimate
IEC International Electrical Commission
ISFOG International Symposium on Frontiers in Offshore Geotechnics
ISO International Organization for Standardization
ISOPE International Society of Offshore and Polar Engineers
JIP Joint Industry Project

275
276 Acronyms

LAT Lowest Astronomical Tide

LC Load Case
LDFE Large Deformation Finite Element
LE Low Estimate
LOC Lightly Over Consolidated
LRFD Load and Resistance Factor Design
MODU Mobile Offshore Drilling Unit
NC Normally Consolidated
NGI Norwegian Geotechnical Institute, Norway
OC Over Consolidated
OLL Optimum Lug Level
OLT Offloading Terminal (i.e. offshore buoy)
OMAE Offshore Mechanics and Arctic Engineering
OTC Offshore Technology Conference, USA
OTRC Offshore Technology Research Center, USA
OWT Offshore Wind Turbine
PFD Partial Factor Design
PLEM Pipe Line End Manifold
PLET Pipe Line End Termination
PMT Pressuremeter Test
PSD Particle Size Distribution
REB Reverse End-Bearing
RP Reference Point (loads and displacements)
RQD Rock Quality Designation
RTA Riser Tower Anchor
SDOF Single Degree of Freedom
SLS Serviceability Limit State
SPT SPT Offshore BV, The Netherlands
SRD Soil Resistance to Driving
SUT Society for Underwater Technology, UK
SWP Self-Weight Penetration
TPD Target Penetration Depth
UCS Unconfined Compressive Strength
ULS Ultimate Limit State
UU Unconsolidated Undrained
WD Water Depth
WHP Well Head Platform
WT Wall Thickness
Notation

The terms caisson and foundation, lug and padeye are synonymous.

αD,su = su averaging parameter (αD,su ≈ 0.25)

  = clay undrained strength averaged from L to L + D αD,su
αi = clay-steel inner adhesion factor (unit friction fi = αi su)
αo = clay-steel outer adhesion factor (unit friction fo = αosu)
αu = excess pore pressure ratio at tip level = ∆u(tip)/∆u
βi = sand-steel inner adhesion factor (unit friction fi = βi σ’vo)
βo = sand-steel outer adhesion factor (unit friction fo = βo σ’vo)
γsub = soil submerged unit weight
γwater = (sea) water unit weight
δx = caisson head lateral displacement
δz = caisson head vertical displacement
ρsoil = soil density
ρsteel = steel density
ρwater = (sea) water density
ΔH = vent (nozzle) differential head loss
Δu = under/overpressure
Δuall = allowable under/overpressure
θxz = caisson rotation
θlug = anchor chain inclination from horizontal at lug level depth zlug
θz0 = anchor chain inclination from horizontal at seafloor
λL = partial load factor (API PFD/LRFD)
λL,Bw = partial load factor on Bw,steel and Bw,plug
φL = partial load factor (ISO 19902)
φm = partial resistance factor (API PFD/LRFD)
γm = partial resistance factor (ISO 19901)
σc = rock unconfined compressive strength
σ'v = soil effective vertical stress
σ'vo = soil effective in-situ vertical stress
ΦMH = ellipse rotation angle
  = atan(ez,su/L) − ∆ΦMH
ΔΦMH = ellipse rotation angle correction

277
278 Notation

ν = soil drained Poisson’s ratio

νu = soil undrained Poisson’s ratio = 0.5
Abase = caisson base area
  = π D2/4
Ai = caisson internal (suction) area
  = π Di2/4
Ap, Awall = caisson wall tip area
  = π (D2 − Di2)/4
Asi = caisson inner perimeter
  = π Di
Aso = caisson outer perimeter
  =πD
An = vent (nozzle) area
aMH = ellipse major semi-axis length
bMH = ellipse minor semi-axis length
ΔbMH = ellipse minor semi-axis length correction
aVH = ellipsoid parameter
bVH = ellipsoid parameter
Cd = vent (nozzle) discharge coefficient
D = caisson outer diameter
Di = caisson inner diameter
  = D − 2 WT
Dr = (cohesionless) soil relative density [−]
E = soil drained Young’s Modulus
Eu = soil undrained Young’s Modulus
ez =H load vertical eccentricity with respect to caisson head
(seafloor) = M/H L L

ez,su ∫ s
= analytical ez based on su,z = u,z
0
z dz
∫
/
0
su,z dz

ez,Hmax = ez based on M/H at Hmax

Fi = caisson inner skin friction resistance
Fo = caisson outer skin friction resistance
g = acceleration due to gravity
FOS = overall factor of safety
H = horizontal load at caisson head (seafloor)
H* = non-dimensional H value = H/(D L su,av,L)
H** = non-dimensional H value = H/(D2 su,av,L)
Hmax = caisson maximum “fixed head” horizontal resistance (V = 0)
  = Np,fixed L D su,av,L
Hmax,V = as Hmax , but with V ≠ 0
Ho = “free head” horizontal resistance (V and M = 0) = Np,free L D su,av,L
iavg = soil plug average upwards seepage gradient (= head loss/L)
icrit = soil plug critical upwards seepage gradient (= γsub/γw)
kf =empirical coefficient relating qc (or qt) to unit skin friction
resistance
Notation 279

kp = empirical coefficient relating qc (or qt) to unit end bearing

resistance
kr = sand plug : soil mass permeability ratio
L = caisson embedded length
Lplug = soil plug length (usually > L)
(L/D)crit = caisson critical embedded length : diameter ratio
M = moment load at caisson head (seafloor)
M* = non-dimensional M value = M/(D L2 su,av,L)
M** = non-dimensional M value = M/(D2 L su,av,L)
Mmax = caisson maximum moment resistance (V = 0)
Mo = caisson moment resistance (V and H = 0)
Nc,strip = strip foundation bearing capacity factor (clay) ≈ min[ 5 (1 + 0.2
L/D), 7.5 ]
Nc,circle = circular foundation bearing capacity factor (clay) ≈ min[ 6 (1 +
0.2 L/D), 9 ]
Np,fixed = “fixed head” lateral bearing capacity factor (clay) = Hmax/(D L
su,av,L)
Np,free = “free head” lateral bearing capacity factor (clay) = Ho/(D L su,av,L)
Nq = dimensionless bearing capacity factor (sand)
qc (qT) = CPT cone (total) tip resistance
qc,z = CPT cone tip resistance at depth z below seafloor
qc,L = CPT cone tip resistance at caisson embedded/tip depth L
Rn = vent (nozzle) resistance
rlug = lug radial offset from centreline
St = soil sensitivity = su/sur
su,z = soil undrained shear strength at depth z
sur = soil remoulded (residual) undrained shear strength
su,0 = soil undrained shear strength at seafloor (z = 0)
su,av,L = average su,z between caisson head (seafloor) and caisson tip (L)
su,tip = soil undrained shear strength at foundation tip depth (L)
su,av,tip = average su,z between caisson tip (L) and depth (L + D αD,su)
su,C = soil undrained shear strength in triaxial compression
su,DSS = soil undrained shear strength in direct simple shear
su,E = soil undrained shear strength in triaxial extension
t = MH resistance ellipse parameter (0 < t < 2π)
T = torsion load at caisson head (seafloor)
Tmax = caisson maximum torsional resistance (V, H and M = 0)
Tlug = anchor chain load at caisson lug level (zlug)
Tz0 = anchor chain load at seafloor
v0 = foundation (free-fall) velocity at seafloor
vpile = foundation (free-fall) velocity
vn = vent (nozzle) velocity
V = vertical load at caisson head (seafloor)
Vmax = caisson axial resistance (T, H and M = 0)
Wplug = soil plug submerged weight
280 Notation

Wsteel = caisson submerged weight

WT = caisson wall thickness
WTtop = caisson top plate thickness
z = depth below seafloor
zlug = lug level depth below seafloor
zlug,cL = lug level depth below seafloor on caisson centreline
References

Abadie, C.N., Byrne, B.W. and Houlsby, G.T. (2017) Modelling of monopile response
to cyclic lateral loading in sand. Proc. of the 8th Int. Conf. Offshore Site
Investigation and Geotechnics. OSIG 2017, London, Society of Underwater
Technology, pp. 1046–1053.
Abbs, A.F. (1983) Lateral pile analysis in weak carbonate rocks. Proc. Conference on
Geotechnical Practice in Offshore Engineering: University of Texas at Austin,
Austin, Texas, USA, 27–29 April, pp. 546–556.
Abbs, A.F. and Needham, A.D. (1985) Grouted piles in weak carbonate rocks. Proc.
Offshore Technology Conference, Houston, Texas, USA, Paper OTC-4852-MS.
Achmus, M., Kuo, Y-S. and Abdel-Rahman, K. (2009) Behavior of monopile
foundations under cyclic lateral load. Computers and Geotechnics, Vol. 36,
pp. 725–735.
Advanced Geomechanics (2002) Suction Pile Analysis Code, AGSpanc Version 4.2,
User’s Manual. Doc. No. AGR-1176, Rev 5.
AG (2003) AGSPANC: Suction Pile Analysis Code, User Manual Version 4.2
Advanced Geomechanics, Australia.
AGS (2000) Guidelines for Combined Geoenvironmental and Geotechnical
Investigations, Association of Geotechnical and Geoenvironmental Specialists,
Beckenham, UK.
Ahn, J., Lee, H. and Kim, Y-T. (2014) Holding capacity of suction caisson anchors
embedded in cohesive soils based on finite element analysis. International
Journal for Numerical and Analytical Methods in Geomechanics, Vol. 38, no.
15, pp. 1541–1555.
Alderlieste, E.A. and Dekker, M.J. (2018) Suction pile design and installation chal-
lenges for the Ophir WHP. Proc. 1st Vietnam Symposium on Advances in Offshore
Engineering (VSOE 2018), Hanoi, Vietnam, Springer, pp. 329–335.
Alderlieste, E.A., Romp, R.H., Kay, S. and Lofterød, A. (2016) Assessment of sea-
floor trench for suction pile moorings: a field case. Proc. Offshore Technology
Conference, Houston, Texas, USA, Paper OTC-27035-MS.
Alderlieste, E.A. and van Blaaderen, E.A. (2015) Installation of suction caissons for
an asymmetrical support structure in sandy soil. Proc. 3rd International
Symposium on Frontiers in Offshore Geotechnics, Taylor & Francis, Oslo,
Norway, pp. 215–220.

281
282 References

Aldridge, T.R., Carrington, T.M. and Kee, N.R. (2005) Propagation of pile tip dam-
age during installation. Proc. 1st International Symposium on Frontiers in
Offshore Geotechnics, Taylor & Francis, Perth, Australia, pp. 823–827.
Andersen, K.H. (2015) Cyclic soil parameters for offshore foundation design. Proc.
3rd International Symposium Frontiers in Offshore Geotechnics, June, Oslo,
Norway, pp. 5–82.
Andersen, K.H. and Jostad, H.P. (1999) Foundation design of skirted foundations
and anchors in clay. Proc. Offshore Technology Conference, Houston, Texas,
USA, Paper OTC-10824-MS.
Andersen, K.H. and Jostad, H.P. (2004) Shear strength along inside of suction anchor
skirt wall in clay. Proc. Offshore Technology Conference, Houston, Texas, USA,
Paper OTC-16844-MS.
Andersen, K.H., Jostad, H.P. and Dyvik, R. (2008) Penetration resistance of offshore
skirted foundations and anchors in dense sand. Journal of Geotechical and
Geoenvironmental Engineering, Vol. 134, no. 1, pp. 106–116.
Andersen, K.H., Murff, J.D., Randolph, M.F., Clukey, E.C., Erbrich, C.T., Jostad,
H.P., Jansen, B., Aubeny, C., Sharma, P. and Supachawarote, C. (2005) Suction
anchors for deepwater applications. Proc. 1st International Symposium on
Frontiers in Offshore Geotechnics, Taylor & Francis, Perth, Australia, pp. 3–30.
API (1997) Recommended practice for planning, designing and constructing Tension
Leg Platforms RP 2T, Ed 2 American Petroleum Institute, Washington, DC. (Ed 3
2015).
API (2000) Recommended practice for planning, designing and constructing fixed
offshore platforms - working stress design RP-2A, American Petroleum Institute,
Washington, USA.
API (2005) Recommended practice for design and analysis of station keeping sys-
tems for floating structures, API RP-2SK, American Petroleum Institute,
Washington, USA.
API (2008) Recommended practice for design and analysis of station keeping sys-
tems for floating structures, API RP-2SK, American Petroleum Institute,
Washington, USA.
API (2011) Recommended practice for geotechnical and foundation design consider-
ations RP 2GEO, Ed 1 American Petroleum Institute, Washington, DC. Errata (2014).
Arany, L., Bhattacharya, S., Macdonald, L. and Hogan, S.J. (2017) Design of mono-
piles for offshore wind turbines in 10 steps. Soil Dynamics and Earthquake
Engineering, Vol. 92, pp. 126–152.
ARGEMA (1994) Foundations in Carbonate Soils, English edn, Technip, Paris.
Arslan, H., Peterman, B.R., Wong, P.C. and Bhattacharjee, S. (2015) Remaining
capacity of the suction pile due to seabed trenching. Proc. 25th International
Symposium on Ocean and Polar Engineering Conference, Hawaii, pp. 924–931.
ASTM (2009) Standard practice for description and identification of soils (Visual-
Manual Procedure), ASTM D2488-09a. USA.
ASTM (2011) Standard practice for classification of soils for engineering purposes
(Unified Soil Classification System), ASTM D2487-11. USA.
Atkinson, J.H., Allman, M.A. and Böese, R.J. (1992) Influence of laboratory sample
preparation procedures on the strength and stiffness of intact Bothkennar soil
recovered using the Laval sampler. Géotechnique, Vol. 42, no. 2, pp. 349–354.
Attewell, P.B. and Farmer, I.W. (1976) Principles of engineering geology, Chapman
and Hall, London.
References 283

Baguelin, F., Frank, R. and Saïd, Y.H. (1977) Theoretical study of lateral reaction
mechanism of piles. Géotechnique, Vol. 27, no. 3, pp. 405–434.
Balder, T., de Lange, D.A., Elkadi, A.S.K., Egberts, P.J.P., Beuckelaers, W.J.A.P.,
Coronel, M., van Dijk, J., Atkinson, R. and Luger, H.J. (2020) Hydraulic pile
extraction scale tests (HyPE-ST): experimental design and preliminary results.
Proc. 4th International Symposium on Frontiers in Offshore Geotechnics, article
#3484, publication #1069 (IC-ISFOG21), 8–11 November 2021 (original dates
16–19 August 2020), Austin, Texas, USA, ASCE Geo-Institute and DFI (Proc pub-
lished in August 2020).
Ballard, J-C., Delvosal, O. and Yonatan, P.H. (2010) Simplified VH equations for
foundation punch-through sand into clay. Proc. 2nd International Symposium on
Frontiers in Offshore Geotechnics, Perth, Australia, pp. 650–655.
Barbosa, P. Geduhn, M. Jardine, R.J. Schroeder, F.C. and Horn, M. (2015) Offshore
pile load tests in chalk. Proc. 16th Eur. Conf. Soil Mech. Geotech. Eng, Edinburgh,
Scotland, pp. 2885–2890.
Bhattacharjee, S., Majhi, S.M., Smith, D. and Garrity, R. (2014) Serpentina FPSO
mooring integrity issues and system replacement: unique fast track approach.
Proc. Offshore Technology Conference, Houston, Texas, USA, Paper
OTC-25449-MS.
Bransby, M.F. and Yun, G-J. (2009) Undrained capacity of skirted strip foundations
under combined loading. Géotechnique, Vol. 59, no. 2, pp. 115–125.
Broms, B.B. (1964a) Lateral resistance of piles in cohesive soils. Journal of the Soil
Mechanics and Foundations Division, Proc. American Society of Civil Engineers,
Vol. 90, no. SM2, pp. 27–63.
Broms, B.B. (1964b) Lateral resistance of piles in cohesionless soils. Journal of the
Soil Mechanics and Foundations Division, Proc. American Society of Civil
Engineers, Vol. 90, no. SM3, pp. 123–156.
Broughton, P., Davies, R.L., Aldridge, T. and Carrington, T. (2002) Foundation
design for the refloat of the Maureen steel gravity platform. Proc. ICE, Geotechnical
Engineering Journal, Vol. 155, no. 2, pp. 111–118.
Brown, G.A. and Nacci, V.A. (1971) Performance of hydrostatic anchors in granular
soils. Proc. Offshore Technology Conference, Houston, Texas, USA, Paper
OTC-1472-MS.
BSI (1999) Code of practice for site investigations, British Standard BS 5930:1999.
British Standards Institution, UK.
Burd, H., Byrne, B., McAdam, R., Housby, G., Martin, C., Beuckelaers, W.,
Zdravkovic, L., Taborda, D., Potts, D., Jardine, J., Gavin, K., Doherty, P., Igoe, D.,
Skov Gretlund, J., Pacheco Andrade, M. and Muir Wood, A. (2017) Design aspects
for monopile foundations. Proc. Workshop TC209 19th ICSMGE, Seoul, 2017,
pp. 35–44.
Burd, H., David, M.G., Taborda, D., Zdravkovic, L., Abadie, C., Byrne, B., Housby,
G., Gavins, K.G., Igoe, D.J.P., Jardine, R.J., Martin, C., McAdam, R., Pedro, A.M.G.
and Potts, D.M. (2020) PISA design model for monopiles for offshore wind tur-
bines: application to a marine sand. Géotechnique. doi:10.1680/jgeot.18.P.277.
Burt, N.J. and Harris, R.P. (1980) Design, installation and testing of belled pile foun-
dations. Proc. Offshore Technology Conference, Houston, Texas, USA, Paper
OTC-3872-MS.
Butterfield, R. and Gottardi, G. (1994) A complete three-dimensional failure enve-
lope for shallow footings on sand. Géotechnique, Vol. 44, no. 1, pp. 181–184.
284 References

Bye, A., Erbrich, C., Rognlien, B. and Tjelta, T.I. (1995) Geotechnical design of
bucket foundations. Proc. Offshore Technology Conference, Houston, Texas,
USA, Paper OTC-7793-MS.
Byrne, B., Housby, G., Burd, H., Gavin, K.G., Igoe, D.J.P., Jardine, R.J., Martin, C.,
McAdam, R., Potts, D.M., David, M.G., Taborda, D., Zdravkovic, L. (2020) PISA
design model for monopiles for offshore wind turbines: application to a stiff gla-
cial clay till. Géotechnique. Published Online doi:10.1680/jgeot.18.P.255.
Byrne, B.W. (2017) New design methods for offshore wind turbine monopiles. Proc
8th International Conference on Offshore Site Investigation and Geotechnics,
held 12–14 September 2017, Royal Geographical Society, London.
Byrne, B.W., McAdam, R., Burd, H.J., Houlsby, G.T., Martin, C.M., Zdravkovic, L.,
Taborda, D.M.G., Potts, D.M., Jardine, R.J., Sideri, M., Schroeder, F.C., Gavin, K.,
Doherty, P., Igoe, D., Muir Wood, A., Kallehave, D. and Skov Gretlund, J. (2015a)
New design methods for large diameter piles under lateral loading for offshore
wind applications. Proc. 3rd International Symposium Frontiers in Offshore
Geotechnics, June, Oslo, Norway, pp. 705–710.
Byrne, B.W., McAdam, R., Burd, H.J., Houlsby, G.T., Martin, C.M., Gavin, K.,
Doherty, P., Igoe, D., Zdravkovic, L., Taborda, D.M.G., Potts, D.M., Jardine, R.J.,
Sideri, M., Schroeder, F.C., Muir Wood, A., Kallehave, D. and Skov Gretlund, J.
(2015b) Field testing of large diameter piles under lateral loading for offshore
wind applications. Proc. XVI ECSMGE Geotechnical Engineering for
Infrastructure and Development, Edinburgh, Scotland, pp. 1255–1260.
Carbon Trust (2019) Suction installed caisson foundations for offshore wind: design
guidelines. Carbon Trust. https://prod-drupal-files.storage.googleapis.com/docu-
ments/resource/public/owa-suction-caisson-design-guidelines-report.pdf.
Carter, J.P. and Kulhawy, F.H. (1992) Analysis of laterally loaded shafts in rock.
ASCE Journal of Geotechnical Engineering, Vol. 118, no. 6, pp. 839–855.
Cathie Associates (2014) OPILE instruction manual, Cathie Associates SA/NV,
Belgium.
CEN (2004) Eurocode 7: geotechnical design - part 1: general rules, European
Standard EN 1997-1:2004. (With Corrigendum EN 1997-1:2004/AC, February
2009, European Committee for Standardization.
Cerfontaine, B., Collin, F. and Charlier, R. (2016) Numerical modelling of transient
cyclic vertical loading of suction caissons in sand, Géotechnique, Vol. 66, no. 2,
pp. 121–136.
CFMS (2019) Recommendations for planning and designing foundations of offshore
wind turbines, French workgroup ‘Foundations of Offshore Wind Turbines,
Version 2.1. http://www.cfms-sols.org/sites/default/files/Rapport-cfms-eoliennes-
offshore-EN.pdf.
Chan, S.F. and Hannah, T.H. (1980) Repeated loading on single piles in sand. ASCE
Journal of Geotechnical and Geoenvironmental Engineering, Vol. 106, no. GT2,
pp. 171–188.
Chatzivasileiou, I. (2014) Installation of suction caissons in layered sand, MSc.
Thesis, Delft University of Technology, The Netherlands.
Chen, W. and Randolph, M.F. (2007a) External radial stress changes and axial capac-
ity for suction caisson in soft clay. Géotechnique, Vol. 57, no. 6, pp. 499–511.
Chen, W. and Randolph, M.F. (2007b) Uplift capacity of suction caissons under
sustained and cyclic loading in soft clay. ASCE Journal of Geotechnical and
Geoenvironmental Engineering, Vol. 133, no. 11, pp. 1352–1363.
References 285

Chow, Y.K. (1981) Dynamic behaviour of piles. Ph.D. Thesis, University of

Manchester, UK.
CIRIA (1984) Report 103 Design of laterally-loaded piles, CIRIA, London, UK,
CIRIA Report 103. Elson, W.K., Construction Industry Research and Information
Association.
CIRIA (2002) C574 Engineering in chalk, Lord, J. A. Clayton, C. R. I. and Mortimore,
R. N., Construction Industry Research and Information Association, UK.
Clausen, C.J.F. and Tjelta, T.I. (1996) Offshore platforms supported by bucket foun-
dations. 15th Congress, Copenhagen, June 16–20: Congress report, international
association for bridge and structural engineering IABSE, Zürich, pp. 819–829.
Clukey, E.C. and Morrison, M.J. (1993) A centrifuge and analytical study to evaluate
suction caissons for TLP applications in the Gulf of Mexico. Design and
Performance of Deep Foundations, Piles and Piers in Soil and Soft Rock, ASCE
Geotechnical Special Publication GSP No. 38, pp. 141–156.
Clukey, E.C., Morrison, M.J., Garnier, J. and Corté, J.F. (1995) The response of suc-
tion caissons in normally consolidated clays to cyclic TLP loading conditions. Proc.
Offshore Technology Conference, Houston, Texas, USA, Paper OTC-7796-MS.
COFS (2003) Deepwater anchor design practice. Phase II Report to API/Deepstar.
Vol. II-2 Appendix B. COFS 3D FE Capacity Analyses of Suction Caisson Anchors.
Australia.
Colliard, D. and Wallerand, R. (2008) Design and installation of suction piles in
West Africa deepwaters. Proc. 2nd British Geotechnical Association International
Conference on Foundations, 24–27 June, University of Dundee, Scotland, UK, IHS
BRE Press, pp. 825–836.
Colliat, J.-L. and Dendani, H. (2002) Girassol: geotechnical design analyses and
installation of the suction anchors. Proc. Society for Underwater Technology Conf.
Offshore Site Investigation and Geotechnics, London, pp. 107–122.
Colliat, J-L., Safinus, S., Boylan, N. and Schroeder, K. (2018) Formation and devel-
opment of seabed Trenching from subsea inspection data of deepwater Gulf of
Guinea moorings. Proc. Offshore Technology Conference, Houston, Texas, USA,
Paper OTC-29034-MS.
Cotter, O. (2009) The installation of suction caisson foundations for offshore renew-
able energy structures, PhD. Thesis, Oxford, UK.
Cox, W.R., Dixon, D.A. and Murphy, B.S. (1984) Lateral load tests on 25.4-mm
(1-in.) diameter piles in very soft clay in side-by-side and in-line groups. Laterally
loaded deep foundations: analysis and performance, ASTM STP 835, American
Society for Testing and Materials, Kansas City, Missouri, USA, pp. 122–139.
De Araujo, J.B., Machado, R.D. and Medeiros, C.J. (2004) High holding power torpedo
pile—results for the first long term application. Proc., 23rd Int. Conf. on Offshore
Mechanics and Arctic Engineering. ASME, Paper No. OMAE2004-51201.
Degenkamp, G. and Dutta, A. (1989) Soil resistances to embedded anchor chain in
soft clay. Journal of Geotechnical Engineering, Vol. 115. doi:10.1061/
(ASCE)0733-9410(1989)115:10(1420).
Dekker, M.J. (2014) The modelling of suction caisson foundations for multi-footed
structures, MSc. Thesis, Delft University of Technology, The Netherlands &
Norwegian University of Science and Technology (NTNU), Trondheim, Norway.
Dendani, H. (2003) Suction anchors: some critical aspects for their design and instal-
lation in clayey soils. Proc. Offshore Technology Conference, Houston, Texas,
USA, Paper OTC-15376-MS.
286 References

Devriendt, C., Van Ingelgem, Y., De Sitter, G., Jordaens, P.J. and Guillaume, P. (2013)
Monitoring of resonant frequencies and damping values of an offshore wind tur-
bine on a monopile foundation, Offshore Wind Infrastructure Application Lab,
Belgium, presentation to EWEA 2013.
Dijkhuizen, C., Coppens, T, and van der Graaf, P. (2003) Installation of the Horn
Mountain Spar using the enhanced DCV Balder. Proc. Offshore Technology
Conference, Houston, Texas, USA, Paper OTC-16367-MS.
DNV (1992) Foundations, Classification Notes No. 30.4. Det Norske Veritas.
DNV (2008) Design of Offshore Steel Structures, General (LRFD Method), Offshore
Standard DNV-OS-C101. Det Norske Veritas.
DNV (2014a) Design of Offshore Wind Turbine Structures, Offshore Standard
DNV-OS-J101. Det Norske Veritas.
DNV (2014b) Modelling and Analysis of Marine Operations, Recommended
Practice DNV-RP-H103. Det Norske Veritas.
DNV (2018) Geotechnical Design and Installation of Suction Anchors in Clay,
Recommended Practice DNV-RP-E303. Det Norske Veritas.
DNV-GL (2016) Support Structures for Wind Turbines, Standard DNVGL-ST-0126.
DNV-GL (2017, 2019) Offshore Soil Mechanics and Geotechnical Engineering,
Recommended Practice DNVGL-RP-C212.
Doherty, J.P. and Deeks, A.J. (2003) Elastic response of circular footings embedded
in a nonhomogeneous half-space. Géotechnique, Vol. 53, no. 8, pp. 703–714.
Doherty, J.P., Houlsby, G.T. and Deeks, A.J. (2005) Stiffness of flexible caisson foun-
dations embedded in nonhomogeneous elastic soil. ASCE Journal of Geotechnical
and Geoenvironmental Engineering, Vol. 131, no. 12, pp. 1498–1508.
Dührkop, J. and Grabe, J. (2009) Design of laterally loaded piles with bulge. Proc.
28th International Conference on Ocean, Offshore and Arctic Engineering,
ASME, Honolulu, Hawaii, USA. OMAE2009‑79087.
Dührkop, J., Maretzki, S. and Rieser, J. (2017) Re-evaluation of pile driveability in
chalk. Proc. of the 8th Int. Conf. Offshore Site Investigation and Geotechnics.
OSIG 2017, Society of Underwater Technology, London, Vol. 2, pp. 666–673.
Dyson, G.J. and Randolph, M. (2001) Monotonic lateral loading of piles in calcare-
ous sediments. ASCE Journal of Geotechnical and Geoenvironmental Engineering,
Vol. 127, no. 24, pp. 346–352.
Edwards, D.H., Zdravkovic, L. and Potts, D.M. (2005) Depth factors for undrained
bearing capacity. Géotechnique, Vol. 55, no. 10, pp. 755–758.
Engin, H.K., Khoa, H.D.V., Jostad, H.P. and Alm, T. (2015) Finite element analyses
of spudcan – subsea template interaction during jack-up rig installation. Proc. 3rd
International Symposium Frontiers in Offshore Geotechnics, June, Oslo, Norway,
705, pp. 1275–1280.
Erbrich, C.T. (2004) A new method for the design of laterally loaded anchor piles in
soft rock. Proc. Offshore Technology Conference, Houston, Texas, USA, Paper
OTC-16441-MS.
Erbrich, C.T., Barbosa-Cruz, E. and Barbour, R. (2010) Soil-pile interaction during
extrusion of an initially Deformed Pile. Proc. 2nd International Symposium on
Frontiers in Offshore Geotechnics, Perth, Australia, pp. 489–494.
Erbrich, C.T. and Hefer, P. (2002) Installation of the Laminaria suction piles – A case
history. Proc. Offshore Technology Conference, Houston, Texas, USA, Paper
OTC-14220-MS.
References 287

Erbrich, C.T., Lam, S.Y., Zhu, H., Derache, A., Sato, A. and Al-Showaiter, A. (2017)
Geotechnical design of anchor piles for Ichthys CPF and FPSO. Proc. of the 8th
Int. Conf. Offshore Site Investigation and Geotechnics. OSIG 2017, Society of
Underwater Technology, London, pp. 1186–1197.
Erbrich, C.T. and Tjelta, T.I. (1999) Installation of bucket foundations and suction
caissons in sand – geotechnical performance. Proc. Offshore Technology
Conference, Houston, Texas, USA, Paper OTC-10990-MS.
Feld, T. (2001) Suction buckets: a new innovation foundation concept, applied to
offshore wind turbines, PhD. Thesis, Aalborg University, Denmark.
Feng, X., Gourvenec, S. and White, D.J. (2019) Load capacity of caisson anchors
exposed to seabed trenching. Ocean Engineering, Vol. 171, no. 1, pp. 181–192.
Feng, X., Randolph, M.F., Gourvenec, S. and Wallerand, R. (2014) Design approach
for rectangular mudmats under fully three dimensional loading. Géotechnique,
Vol. 64, no. 1, pp. 51–63.
Finnie, I.M.S. and Randolph, M.F. (1994) Punch-through and liquefaction induced
failure of shallow foundations on calcareous sediments. Proc. 7th International
Conference on the Behaviour of Offshore Structures, BOSS ’94, Vol. 1, Pergamon,
Oxford, pp. 217–230.
Fisher, R. and Cathie, D. (2003) Optimisation of gravity based design for subsea appli-
cations. Proc. 2nd British Geotechnical Association International Conference on
Foundations, 2–5 September, University of Dundee, Scotland, UK, IHS BRE Press.
Fleshman, M.S. and Rice, J.D. (2014) Laboratory modeling of the mechanisms of
piping erosion initiation. ASCE Journal of Geotechnical and Geoenvironmental
Engineering, Vol. 140, no. 6. doi:10.1061/(ASCE)GT.1943-5606.0001106.
Fragio, A.G., Santiago, J.L. and Sutton, V.J.R. (1985) Load tests on grouted piles in
rock. Proc. Annual Offshore Technology Conference, Houston, Texas, USAs,
Paper OTC-4851-MS.
Frankenmolen, S.F., Erbrich, C.T. and Fearon, R.E. (2017) Successful installation of
large suction caissons and driven piles in carbonate soils. Proc. of the 8th Int.
Conf. Offshore Site Investigation and Geotechnics. OSIG 2017, London, Society
of Underwater Technology, pp. 539–548.
Fuglsang, L.D. and Steensen-Bach, J.O. (1991) Breakout resistance of suction piles in
clay. Centrifuge 91: Proc. International conference centrifuge, Boulder, Colorado,
13–14 June 1991, A.A. Balkema, Rotterdam, pp. 153–159.
Fugro (2009) CANCAP2 User’s Manual, Fugro Engineers BV Internal Document
No. FEBV/CDE/MAN/083.
Gazetas, G. (1991) Foundation vibrations. In: Foundation engineering handbook,
Fang, H-Y. (Ed.), 2nd edn, Van Nostrand Reinhold Company, New York, pp.
553–593.
Germanischer Lloyd (2013) Rules for the certification and construction, IV Industrial
services, Part 7 Offshore substations, Chapter 2 Structural Design, Germanischer
Lloyd, Hamburg.
Germanischer Lloyd Wind Energie GmbH (2005) Rules and guidelines: IV –
Industrial services, Part 2 – Guideline for the certification of offshore wind tur-
bines, Germanischer Lloyd, Hamburg.
Gerolymos, N., Zafeirakos, A. and Karapiperis, K. (2015) Generalized failure enve-
lope for caisson foundations in cohesive soil: static and dynamic loading. Soil
Dynamics and Earthquake Engineering, Vol. 78, pp. 154–174.
288 References

Gilbert, R.B. and Murff, J.D. (2001) Identifying uncertainties in the design of suction
caisson foundations. Proc. International Conference on Geotechnical, Geological
and Geophysical Properties of Deepwater Sediments honoring Wayne A. Dunlap,
OTRC, Houston, Texas, pp. 231–242.
Goodman, L.J., Lee, C.N. and Walker, F.J. (1961) The feasibility of vacuum anchor-
age in soil. Géotechnique, Vol. 1, no. 3, pp. 356–359.
Gottardi, G., Houlsby, G.T. and Butterfield, R. (1999) Plastic response of circular
footings on sand under general planar loading. Géotechnique, Vol. 49, no. 4, pp.
453–469.
Gourvenec, S. and Jensen, K. (2009) Effect of embedment and spacing on co-joined
skirted foundation systems on undrained limit states under general loading.
International Journal of Geomechanics, Vol. 9, no. 6, pp. 267–279.
Gourvenec, S. and Steinepreis, M. (2007) Undrained limit states of shallow founda-
tions acting in consort. ASCE, International Journal of Geomechanics, Vol. 7, no.
3, pp. 194–205.
Gourvenec, S.M., Feng, X., Randolph, M.F. and White, D.J. (2017) A toolbox for
optimizing geotechnical design of subsea foundations. Proc. Annual Offshore
Technology Conference, Houston, Texas, USA Paper OTC-27703-MS.
Govoni, L, Gottardi, G., Antoninia, A., Archettia, R. and Schweizer, J. (2016) Caisson
foundations for competitive offshore wind farms in Italy. Procedia Engineering,
Vol. 158, pp. 392–397.
Griffiths, D.V. (1982) Elasto-plastic analyses of deep foundations in cohesive soil.
International Journal for Numerical and Analytical Methods in Geomechanics,
Vol. 6, pp. 211–218.
Griffiths, D.V. (1985) HARMONY – A program for predicting the elasto-plastic
response of axisymmetric Bodies subjected to non-axisymmetric loading, Report
to Fugro-McClelland Engineers B.V., University of Manchester, UK.
Guo, W., Chu, J. and Kou, H. (2016) Model tests of soil heave plug formation in suc-
tion caisson. Proc. Institution of Civil Engineers, Geotechnical Engineering, Vol.
169, no. GE2, pp. 214–223.
Gupta, B.K. and Basu, D. (2016) Design charts for laterally loaded rigid piles in elas-
tic soil. Proc. ASCE Geo-Chicago 2016 GSP 270, Chicago, Illinois, USA, pp.
393–406.
Hansteen, O.E., Jostad, J.P. and Tjelta, T.I. (2003) Observed platform response to a
“monster” wave. Proc. 6th Int. Symposium on Field Measurements in Geomechanics,
Swets and Zeitlinger, Lisse, The Netherlands.
Heins, E. and Grabe, J. (2017) Class-A-prediction of lateral pile deformation with
respect to vibratory and impact pile driving. Computers and Geotechnics, Vol. 86,
pp. 108–119.
Helfrich, S.C., Brazill, R.L. and Richards, A.F. (1976) Pullout characteristics of a suc-
tion anchor in sand. Proc. of Offshore Technology Conference, Houston, Texas,
USA, Paper OTC-2469-MS.
Hernandez-Martinez, F.G., Rahim, A., Strandvik, S., Jostad, H.P. and Andersen, K.H.
(2009) Consolidation settlement of skirted foundations for subsea structures in
soft clay. Proc. 17th International Conference on Soil Mechanics and Geotechnical
Engineering, pp. 1167–1172.
Hernandez-Martinez, F.G., Saue, M., Schroder, K. and Jostad, H.P. (2015) Trenching
effects on holding capacity for In-service suction anchors in high plasticity clays.
References 289

Proc. 20th Offshore Symposium, Texas Section of the Society of Naval Architects
and Marine Engineers, Houston, Texas, USA.
Hogervorst, J.R. (1980) Field trials with large diameter suction piles. Proc. of
Offshore Technology Conference, Houston, Texas, USA, Paper OTC-3817-MS.
Holeyman, A., Peralta, P. and Charue, N. (2015) Boulder-soil-pile dynamic interac-
tion. Proc. 3rd International Symposium on Frontiers in Offshore Geotechnics,
Taylor & Francis, Oslo, Norway, pp. 563–568.
Hölscher, P., van Tol, A.F. and Huy, N.Q. (2009) Influence of rate effect and pore
water pressure during rapid load test of piles in sand. Proc. Rapid Load Testing on
Piles, van Tol, F. and Hölscher, P. (Eds.), CRC Press, pp. 59–72.
Holtz, R.D. (1991) Stress distribution and settlement of shallow foundations. In:
Foundation Engineering Handbook, Winterkorn, H.F. and Fang, H.Y. (Eds.), 2nd
edn, Van Nostrand Reinhold Company, New York, pp. 166–222.
Horvath, R.G. (1978) Field load test data on concrete to rock bond strength,
University of Toronto, Canada, Publication No. 78-07. 23.
Horvath, R.G. and Kenney, T.C. (1979) Shaft resistance of rock-socketed drilled
piers. Symposium on Deep Foundations, ASCE National Convention, Atlanta,
Georgia, USA, pp. 182–214.
Hossain, M., Lehane, B., Hu, Y. and Gao, Y. (2012) Soil flow mechanisms around
and between stiffeners of caissons during installation in clay. Canadian
Geotechnical Journal, Vol. 49, no. 4, pp. 442–459.
Houlsby, G.T. and Byrne, B.W. (2005) Design procedures for installation of suction
caissons in sand. Proc. Institution of Civil Engineers, Geotechnical Engineering,
Vol. 158, no. GE3, Paper 13818, pp. 135–144.
Houlsby, G.T., Kelly, R.B. and Byrne, B.W. (2005) The tensile capacity of suction
caissons under rapid loading. Proc. 1st International Symposium on Frontiers in
Offshore Geotechnics, Taylor & Francis, Perth, Australia, pp. 405–410.
Houlsby, G.T., Kelly, R.B., Huxtable, J. and Byrne, B.W. (2006) Field trials of suc-
tion caissons in sand for offshore wind turbine. Géotechnique, Vol. 56, no. 1,
pp. 3–10.
Houlsby, G.T. and Martin, C.T. (2003) Undrained bearing capacity factors for coni-
cal footings on clay. Géotechnique, Vol. 53, no. 5, pp. 513–520.
Houlston, P., Manceau, S., Benson, A. Alobaidi, I., Evans, R., Pacheco Andrade, M.
and Meincke, J.H. (2017) Laboratory testing based design method for assessing
cyclic response of driven piles in chalk. Proc. of the 8th Int. Conf. Offshore Site
Investigation and Geotechnics. OSIG 2017, Society of Underwater Technology,
pp. 449–456.
Huy, N.Q., van Tol, A.F. and Hölscher, P. (2009) Rapid model pile load tests in the
geotechnical centrifuge. Proc. Rapid Load Testing on Piles, van Tol, F. and
Hölscher, P. (Eds.), CRC Press, pp. 103–127.
Ibsen, C.B. and Thilsted, C.L. (2010) Numerical study of piping limits for suction
installation of offshore skirted foundations and anchors in layered sand. Proc. 2nd
International Symposium on Frontiers in Offshore Geotechnics, Perth, Australia,
pp. 421–426.
ICE (1992) Bothkennar soft clay test site: characterisation and lessons learned, 8th
Géotechnique Symposium in Print. Géotechnique, Vol. 42, no. 2, pp. 163–375.
IHC (2010) IHC Hydrohammer. Pile Driving Equipment, IHC Merwede, The
Netherlands, 13 pp.
290 References

Irvine, J., Terente, V., Lee, L.T. and Comrie, R. (2015) Driven pile design in weak
rock. Proc. 3rd International Symposium on Frontiers in Offshore Geotechnics,
Taylor & Francis, Oslo, Norway, pp. 569–574.
ISO (2006) Petroleum and natural gas industries – fixed concrete offshore structures,
International Organization for Standardization, ISO 19903-1:2019.
ISO (2007) Petroleum and natural gas industries – fixed steel offshore structures,
International Organization for Standardization, ISO 19902:2007.
ISO (2011) Petroleum and natural gas industries – specific requirements for offshore
structures – part 7: stationkeeping systems for floating offshore structures and
mobile offshore units, International Organization for Standardization, ISO/DIS
19901-7.
ISO (2014a) Petroleum and natural gas industries – specific requirements for off-
shore structures – part 7: stationkeeping systems for floating offshore structures
and mobile offshore units, ISO 19901-7:2013.
ISO (2014b) Petroleum and natural gas industries – specific requirements for off-
shore structures – part 8: marine soil investigations International Organization for
Standardization, ISO 19901-8:2014.
ISO (2016a) Petroleum and natural gas industries – specific requirements for off-
shore structures – part 4: geotechnical and foundation design considerations,
International Organization for Standardization, ISO 19901-4:2016.
ISO (2016b) Petroleum and natural gas industries – site specific assessment of mobile
offshore units - part 1: jack-ups, International Organization for Standardization,
ISO 19905-1:2016(E).
Janbu, N. (1985) Soil models in offshore engineering. Géotechnique, Vol. 35, no. 3,
pp. 241–281.
Jardine, R., Chow, F.C., Overy, R.F. and Standing, J.R. (2005) ICP design methods
for driven piles in sands and clays, Thomas Telford Ltd., London.
Jardine, R.J., Buckley, R.M., Byrne, B.W., Kontoe, S., McAdam, R.A., Andolfsson, T.,
Liu, T., Schranz, F. and Vinck, K. (2019) Improving the design of piles driven in
chalk through the ALPACA research project. Revue Française de Géotechnique,
Vol. 158, no. 2. doi:10.1051/geotech/2019008.
Jardine, R.J., Standing, J.R. and Chow, F.C. (2006) Some observations of the effects
of time on the capacity of piles driven in sand. Géotechnique, Vol. 56, no. 4, pp.
227–244.
Jeanjean, P. (2006) Set-up characteristics of suction anchors for soft Gulf of Mexico
clays: experience from field installation and retrieval. Proc. Offshore Technology
Conference, Houston, Texas, USA, Paper OTC-18005-MS.
Jeanjean, P., Znidarcic, D., Phillips, R., Ko, H-Y., Pfister, S., Cinicioglu, O. and
Schroeder, K. (2006) Centrifuge testing on suction anchors: Double-wall, over-
consolidated clay, and layered soil profile. Proc. Offshore Technology Conference,
Houston, Texas, USA, Paper OTC-18007-MS.
Jostad, H.P. (1997) BIFURC - Version 3 Undrained Capacity Analyses of Clay,
Technical Report, Norwegian Geotechnical Institute.
Kallehave, D., LeBlanc, C., Liingaard, M.A. (2012) Modification of the API P-Y
formulation of initial stiffness of sand, Integrated Geotechnologies - Present and
Future. Proc. of the 7th Int. Conf. Offshore Site Investigation and Geotechnics.
OSIG 2017, Society of Underwater Technology, London, pp. 465–472.
Karlsrud, K., Jensen, T.G., Lied, E.K.W., Nowacki, F. and Simonsen, A.S. (2014)
Significant ageing effects for axially loaded piles in sand and clay verified by new
References 291

field load tests. Proc. Offshore Technology Conference, Houston, Texas, USA,
Paper OTC-25197-MS.
Kay, S. (2013) Torpedo piles – VH capacity in clay using resistance envelope equa-
tions. Proc. 32nd International Conference on Ocean, Offshore and Arctic
Engineering, OMAE2013, 9–14 June, Nantes, France.
Kay, S. (2015) CAISSON_VHM: a program for caisson capacity under VHM Load
in undrained soils. Proc. 3rd International Symposium on Frontiers in Offshore
Geotechnics, Taylor & Francis, Oslo, Norway, pp. 277–282.
Kay, S. (2018) User’s Manual CAISSON_VHM Version 00.12, Computer Program
for Caisson Foundation Resistance in Undrained Soils during VHM Load, S.Kay
Consultant Report SKA/MAN/003, March, 850 pp.
Kay, S. and Palix, E. (2010) Caisson capacity in clay: VHM resistance envelope Part
2: VHM envelope equation and design procedures. Proc. 2nd International
Symposium on Frontiers in Offshore Geotechnics, Perth, Australia.
Kay, S. and Palix, E. (2011) Caisson capacity in clay: VHM Resistance Envelope –
Part 3: extension to shallow foundations. Proc. 30th International Conference on
Ocean, Offshore and Arctic Engineering, OMAE2011, 19–24 June, Rotterdam,
the Netherlands.
Kay, S. and Van Woensel, F. (2002) Cone pressuremeter tests in clay. Proc. 5th
European Conference Numerical Methods in Geotechnical Engineering, Paris,
4–6 September 2002, Laboratoire Central des Ponts et Chaussées (LCPC) and
Presses de l’École Nationale des Ponts et Chaussées, Paris, pp. 99–105.
Keaveny, J.V., Hansen, S.B., Madshus, C. and Dyvik, R. (1994) Horizontal capacity
of large-scale model anchors. Proc. 13th Int. Conf. Soil Mechanics and Foundation
Engineering, New Delhi, India, pp. 677–680.
Kellezi, L., Kudsk, G. and Hansen, P.B. (2005) FE modelling of spudcan-pipeline
interaction. Proc. 1st International Symposium on Frontiers in Offshore
Geotechnics, Taylor & Francis, Perth, Australia, pp. 551–557.
Kelly, R.B., Byrne, B.W., Houlsby, G.T. and Martin, C.M. (2003) Pressure chamber
testing of model caisson foundations in sand. Proc. 2nd British Geotechnical
Association International Conference on Foundations, 2–5 September, University
of Dundee, Scotland, UK, IHS BRE Press, pp. 421–432.
Kelly, R.B., Byrne, B.W., Houlsby, G.T. and Martin, C.M. (2004) Tensile loading of
model caisson foundations for structures on sand. Proc. 14th International
Offshore and Polar Engineering Conference, Toulon, Vol. 2, pp. 638–641.
Kelly, R.B., Houlsby, G.T. and Byrne, B.W. (2006a) A comparison of field and labo-
ratory tests of caisson foundations in sand and clay. Géotechnique, Vol. 56, no. 9,
pp. 617–626.
Kelly, R.B., Houlsby, G.T. and Byrne, B.W. (2006b) Transient vertical loading of model
suction caissons in a pressure chamber. Géotechnique, Vol. 56, no. 10, pp. 665–675.
Kennedy, J., Oliphant, J., Maconochie, A., Stuyts, B. and Cathie, D. (2013) CAISSON:
a suction pile design tool. Proc. 32nd International Conference on Ocean,
Offshore and Arctic Engineering, OMAE2013, 9–14 June, Nantes, France.
Kirsch, F., Richter, T., Coronel, M. (2014) Geotechnische Aspekte bei der
Gründungsbemessung von Offshore-Windenergieanlagen auf Monopfählen mit
sehr großen Durchmessern, Stahlbau Spez. 2014 – Erneuerbare Energien 83.
Kolk, H.J., Kay, S., Kirstein, A. and Troestler, H. (2001) North Nemba flare bucket
foundations. Proc. Offshore Technology Conference, Houston, Texas, USA, Paper
OTC-13057-MS.
292 References

Kolk, H.J. and Wegerif, J. (2005) Offshore site investigations: new frontiers. Proc.
1st International Symposium on Frontiers in Offshore Geotechnics, Taylor &
Francis, Perth, Australia, pp. 145–161.
Kuhlemeyer, R. (1979) Static and dynamic laterally loaded floating piles. ASCE
Journal of Geotechnical and Geoenvironmental Engineering, Vol. 105, 289–304.
Lambe, T.W. (1973) Predictions in soil engineering. Géotechnique, Vol. 23, no. 2, pp.
151–201.
Langford, T.E., Solhjell, E., Hampson, K. and Hondebrink, L. (2012) Geotechnical
design and installation of suction anchors for the Skarv FPSO, Offshore Norway.
Proc. 7th Int. Conf. Offshore Site Investigation and Geotechnics. OSIG 2012,
Society of Underwater Technology, London, pp. 613–620.
LeBlanc, C., Houlsby, G.T., Byrne, B.W. (2010) Response of stiff piles in sand to
long-term cyclic lateral loading. Géotechnique, Vol. 60, no. 2, pp. 79–90.
Lee, Y.C., Audibert, J.M.E. and Tjok, K.-M. (2005) Lessons learned from several
suction caisson installation projects in clay. Proc. 1st International Symposium on
Frontiers in Offshore Geotechnics, Taylor & Francis, Perth, Australia, pp.
235–241.
Lehane, B.M., Elkhatib, S. and Terzaghi, S. (2014) Extraction of suction caissons in
sand, Géotechnique, Vol. 64, no. 9, pp. 735–739.
Lehane, B.M., Schneider, J.A. and Xu, X. (2005) A review of design methods for
offshore driven piles in siliceous sand, University of Western Australia, Perth,
UWA Report, GEO:05358.
Lin, S.-S. and Liao, J.-C. (1999) Permanent strains of piles in sand due to cyclic lat-
eral loads. ASCE Journal of Geotechnical and Geoenvironmental Engineering,
Vol. 125, No. 9, pp. 798–802.
Lovera, A. (2019) Cyclic lateral design for offshore monopiles in weak rocks. PhD
Thesis, Université Paris-Est, Environmental Engineering, France. NNT:
2019PESC1016.
Lovera, A., Ghabezloo, S., Sulem, J., Randolph, M.F. and Palix, E. (2020) Extension
of the p-y curves framework for cyclic loading of offshore wind turbines mono-
piles. Proc. 4th International Symposium on Frontiers in Offshore Geotechnics,
Austin, Texas, USA.
Lunne, T., Robertson, P.K. and Powell, J.J.M. (1997) Cone penetration testing in
geotechnical practice, Blackie Academic & Professional, London.
Lunne, T. and St John, H.D. (1979) The use of cone penetrometer tests to compute
penetration resistance of steel skirts underneath North Sea gravity platforms.
Proc. 7th European Conference on Soil Mechanics and Foundation Engineering,
Brighton, England, September, Vol. 2, British Geotechnical Society, London, pp.
233–238.
Lunne, T.A., Myrovll, F. and Kleven, A. (1982) Observed settlements of five North
Sea gravity platforms, pp. 25, NGI Publication 139. ISBN: 0078-1193. A
Norwegian Geotechnical Institute (NGI) publication, Norway.
Madsen, S., Andersen, L.V. and Ibsen, L.B. (2012) Instability during installation of
foundations for offshore structures. Proc. 16th Nordic Geotechnical Meeting,
Copenhagen, Denmark, Vol. 2, Dansk Geoteknisk Forening, pp. 499–505.
Mana, D.S.K., Gourvenec, S. and Randolph, M.F. (2013) A novel technique to miti-
gate the effect of gapping on the uplift capacity of offshore shallow foundations.
Géotechnique, Vol. 63, no. 14, pp. 1245–1252.
References 293

Martinelli, M., Alderlieste, E.A., Galavi, V. and Luger, H.J. (2020) Numerical simula-
tion of the installation of suction bucket foundations using MPM. Proc. 4th
International Symposium on Frontiers in Offshore Geotechnics, Austin, Texas,
USA.
McManus, K.J. and Davis, R.O. (1997) Dilation-induced pore fluid cavitation in
sands. Géotechnique, Vol. 47, no. 1, pp. 173–177.
McNamara, A.P. (2000) Behaviour of soil around the internal stiffeners of suction
caissons and the effect on installation and pullout resistance, Geomechanics
Group Report No. G1512, University of Western Australia, Australia, June.
Meigh, A.C. and Wolski, W. (1979) Design parameters for weak rock. Proc. 7th
European Conference on Soil Mechanics and Foundation Engineering, British
Geotechnical Society, London, UK.
Metrikine, A.V., Tsouvalas, A, Segeren, M.L.A., Elkadi, A.S.K., Tehrani, F.S., Gómez,
S.S., Atkinson, R., Pisanò, F., Kementzetzidis, E., Tsetas, A., Molenkamp, T., van
Beek, K. and de Vries, P. (2020) GDP: a new technology for gentle driving of
(mono)piles. Proc. 4th International Symposium on Frontiers in Offshore
Geotechnics, article #3483; publication #1069 (IC-ISFOG21), 8–11 November
2021, Austin, Texas, USA, ASCE Geo-Institute and DFI (Proc. published in August
2020).
Mirza, U.A., Sweeney, M. and Dean, A.R. (1988) Potential effects of jack-up spudcan
penetration on jacket piles. Proc. of Offshore Technology Conference, Houston,
Texas, USA, OTC-5762-MS.
Muir Wood, A., Mackenzie, B., Burbury, D., Rattley, M., Clayton, C.R.I., Mygind,
M., Wessel Andersen, K., LeBlanc Thilsted, C. and Alberg Liingaard, M. (2015)
Design of large diameter monopiles in chalk at Westermost Rough offshore wind
farm. Proc. 3rd International Symposium on Frontiers in Offshore Geotechnics,
Taylor & Francis, Oslo, Norway, pp. 723–728.
Murff, J.D. (1975) Response of axially loaded piles. Journal of the Geotechnical
Engineering Division, Proceedings of the American Society of Civil Engineers, Vol.
101, no. GT3, pp. 356–360.
Murff, J.D. (1994) Limit analysis of multi-footing foundation systems. Proc. Int.
Conf. on Computer Methods and Advances in Geomechanics, Morgantown, 1,
pp. 223–244.
Murff, J.D. and Hamilton, J.M. (1993) P-ultimate for undrained analysis of laterally
loaded pile’. ASCE Journal of Geotnical Engineering, Vol. 119, no. 1, pp.
91–107.
Negro, V., López-Gutiérrez, J-S., Dolores Esteban, M., Alberdi, P., Imaz, M. and
Serraclara, J-M. (2017) Monopiles in offshore wind: preliminary estimate of main
dimensions. Ocean Engineering, Vol. 133, pp. 253–261.
Neubecker, S.R. and Randolph, M.F. (1995) Profile and frictional capacity of embed-
ded anchor chain. ASCE Journal of Geotnical Engineering, Vol. 121, no. 11, pp.
787–803.
Newlin, J.A. (2003a) Suction anchor piles for the Na Kika FDS mooring system. Part
1: site characterization and design. Proc. International Symposium on Deepwater
Mooring Systems: Concepts, Design, Analysis and Materials, American Society of
Civil Engineers, Houston, Texas, USA, pp. 28–54.
Newlin, J.A. (2003b) Suction anchor piles for the Na Kika FDS mooring system. Part
2: site installation performance. Proc. International Symposium on Deepwater
294 References

Mooring Systems: Concepts, Design, Analysis and Materials, American Society of

Civil Engineers, Houston, Texas, USA, pp. 55–75.
NGI (1997) Skirted foundations and anchors in clay – state of the art report,
Norwegian Geotechnical Institute Report 524071-1, Norway, 15 December,
p. 158.
Nguyen, T.C., van Lottum, H., Hölscher, P. and van Tol, A.F. (2012) Centrifuge mod-
eling of rapid load tests with open-ended piles. Proc. 9th Int. Conf. on Testing and
Design Methods for Deep Foundations, Japan, pp. 265–272.
Nichols, G. (2009) Sedimentology and stratigraphy, 2nd edn, Wiley-Blackwell,
Canada, 432 pp.
Norwegian Geotechnical Institute (1997) Skirted foundations and anchors in clays
- State-of-the-art, NGI Report No. 52407-1, December.
Norwegian Geotechnical Institute (2004) Windows program HVMCap. Version 3.0:
theory, user manual and certification. Norwegian Geotechnical Institute Report
524096-7, Rev. 2.
Novello, E.A. (1999) From static to cyclic p-y data in calcareous sediments. Proc.
2nd International Conference on Engineering for Calcareous Sediments, 15–18
March, Routledge, Perth, pp. 17–27.
NPD (1992) Guidelines concerning loads and load effects to regulations concerning
loadbearing structures in the petroleum activities. Issued by the Norwegian
Directorate, 7 February.
O’Loughlin, C.D., Randolph, M.F. and Richardson, M. (2004) Experimental and
theoretical studies of deep penetrating anchors. Proc. Offshore Technology
Conference, Houston, Texas, USA, Paper OTC-16841-MS.
O’Riordan, N.J. (1991) Effects of cyclic loading on the long-term settlements of
structures, Chapter 9. In: Cyclic loading of soils: from theory to design, O’Reilly,
M.P. and Brown, S.F. (Eds.), Blackie & Son Ltd, UK, pp. 411–433.
Osman, A.S. and Randolph, M.F. (2012) Analytical solution for the consolidation
around a laterally loaded pile. ASCE International Journal of Geomechanics, Vol.
12, no. 3, pp. 199–208.
OTRC (2008) FALL16 PROGRAM. Foundation analysis by the limit load method --
16th in a series of limit analysis programs. Suction caisson analysis under inclined
load, Excel document FALL16Rev3Mar2008_master.xls.
Paikowsky, S.G. (1989) A static evaluation of soil plug behavior with application to
the pile plugging problem, PhD. Thesis, Massachusetts Institute of Technology,
USA.
Palix, E., Chan, N., Yangrui, Z. and Haijing, W. (2013) Liwan 3-1: how deepwater
sediments from South China Sea compare with Gulf of. Guinea sediments. Proc.
Offshore Technology Conference, Houston, Texas, USA, Paper OTC-24010-MS.
Palix, E. and Lovera, A. (2020) Field testing for monopile to be installed in weak
carbonated rock. Proc. 4th International Symposium on Frontiers in Offshore
Geotechnics, Austin, Texas, USA.
Palix, E., Willems, T. and Kay, S. (2010) Caisson capacity in clay: VHM resistance
envelope – Part 1: 3D FEM numerical study. Proc. 2nd International Symposium
on Frontiers in Offshore Geotechnics, Perth, Australia, pp. 753–758.
Panagoulias, S. (2016) Critical pressure during installation of suction caissons in
sand, MSc. Thesis, Delft University of Technology, The Netherlands.
Panagoulias, S., van Dijk, B.F.J., Drummen, T., Askarinejad, A. and Pisanò, F. (2017)
Critical suction pressures during installation of suction caissons in sand. Proc. of
References 295

the 8th Int. Conf. Offshore Site Investigation and Geotechnics. OSIG 2017,
Society of Underwater Technology, London, pp. 570–577.
PanGeo (2010) Mitigating risk for offshore installations, presentation to The
Hydographic Society, UK, 21 October 2010, 54 pp. https://www.ths.org.uk/event_
details.asp?v0=234.
Peck, R.B. (1969) Advantages and limitations of the observational method in applied
soil mechanics. Géotechnique, Vol. 19, no. 2, pp. 171–187.
Pells, P.J.N., Rowe, R.K. and Turner, R.M. (1980) An experimental investigation into
side shear for socketed piles in sandstone. Proc. International Conference on
Structural Foundations on Rock, Sydney, 7–9 May, Balkema, Vol. 1, pp. 291–302.
Pile Dynamics, Inc. (2010) GRLWEAP, wave equation analysis of pile driving,
Computer Program Package.
Poulos, H.G. and Davis, E.H. (1980) Pile foundation analysis and design, John Wiley
and Sons, New York, Series in Geotechnical Engineering, 410 pp.
Poulos, H.G. and Davis, E.H. (1991) Elastic solutions for soil and rock mechanics,
Series in Soil Engineering, Originally published 1974 by John Wiley & Sons (ISBN
10: 0471695653 ISBN 13: 9780471695653), republished in 1991 by Centre for
Geotechnical Research, Sydney University. Available online http://www.usucger.
org/PandD/PandD.htm.
Poulos, H.G. and Randolph, M.F. (1983) Pile group analysis: a study of two
methods. ASCE, Journal of Geotechnical Engineering, Vol. 109, no. 3, March,
pp. 355–372.
Puech, A., Becue, J.P. and Colliat, J.-L. (1988) Advances in the design of piles driven
into non-cemented to weakly cemented carbonate formations. Proc. 1st
International Conference on Calcareous Sediments, 15–18 March 1988, Perth,
A.A. Balkema, Rotterdam, pp. 305–312.
Puech, A., Colliat, J-L., Nauroy, J.F. and Meunier, J. (2005) Some geotechnical speci-
ficities of Gulf of Guinea deepwater sediments. Proc. 1st International Symposium
on Frontiers in Offshore Geotechnics, Taylor & Francis, Perth, Australia, pp.
1047–1053.
Puech, A., Poulet, D. and Boisard, P. (1990) A procedure to evaluate pile drivability
in the difficult soil conditions of the southern part of the Gulf of Guinea. Proc.
Offshore Technology Conference, Houston, Texas, USA, Paper OTC-6237-MS.
Puech, A. and Quiterio-Mendoza, B. (2019) Caractérisation des massifs rocheux
pour le dimensionnement de pieux forés en mer (Characterization of rock masses
for designing drilled and grouted offshore pile foundations). Revue Française de
Géotechnique, Vol. 158, no. 5. doi:10.1051/geotech/2019011.
Randolph, M.F. (1981) Piles subjected to torsion. Journal of the Geotechnical
Engineering Division, ASCE, Vol. 107, no. GT8, pp. 1095–1111.
Randolph, M.F. (2018) Potential damage to steel pipe piles during installation.
International Press-in Association (IPA), Vol. 3, no. 1, March 2018, pp. 8, Special
Contribution. https://www.press-in.org/en/page/specialcontributions.
Randolph, M.F. and Gourvenec, S. (2011) Offshore geotechnical engineering, Spon
Press, London.
Randolph, M.F. and Houlsby, G.T. (1984) The limiting pressure on a circular pile
loaded laterally in cohesive soil. Géotechnique, Vol. 34, no. 4, pp. 613–623.
Randolph, M.F. and House, A.R. (2002) Analysis of suction caisson capacity in
clay. Proc. Offshore Technology Conference, Houston, Texas, USA, Paper
OTC-14236-MS.
296 References

Randolph, M.F., Martin, C.M. and Hu, Y. (2000) Limiting resistance of a spherical
penetrometer in cohesive material. Géotechnique, Vol. 50, no. 5, pp. 573–582.
Randolph, M.F., O’Neill, M.P. and Stewart, D.P. (1998) Performance of suction
anchors in fine-grained calcareous soils. Proc. Offshore Technology Conference,
Houston, Texas, USA, Paper OTC-8831-MS.
Randolph, M.F. and Wroth, C.P. (1978) Analysis of deformation of vertically loaded
piles. Journal of the Geotechnical Engineering Division, American Society of Civil
Engineers, Vol. 104, no. GT12, pp. 1465–1488.
Reese, L.C. (1997) Analysis of laterally loaded piles in weak rock. ASCE Journal
of Geotechnical and Geoenvironmental Engineering, Vol. 123, no. 11, pp.
1010–1017.
Reese, L.C., Cox, W.R. and Koop, F.D. (1974) Analysis of laterally loaded piles
in sand. Proc. Offshore Technology Conference, Houston, Texas, USA, Paper
OTC-2080-MS.
Romp, R.H. (2013) Installation-effects of suction caissons in non-standard soil con-
ditions, MSc. Thesis, Delft University of Technology, The Netherlands.
Rosenberg, P. and Journeaux, N.L. (1976) Friction and end bearing tests on bed-
rock for high capacity socket design. Canadian Geotechnical Journal, Vol. 13,
pp. 324–333.
Rowe, R.K. and Armitage, H.H. (1987) A design method for drilled piers in soft
rock. Canadian Geotechnical Journal, Vol. 24. pp. 126–142.
Salgado, R., Lyamin, A.V., Sloan, S.W. and Yu, H.S. (2004) Two and three-dimen-
sional bearing capacity of foundations in clay. Géotechnique, Vol. 54, no. 5, pp.
297–306.
Sassi, K., Kuo, M.Y.-H., Versteele, H., Cathie, D. and Zehzouh, S. (2017) Insights
into the mechanisms of anchor chain trench formation. Proc. 8th Int. Conf.
Offshore Site Investigation and Geotechnics, OSIG 2017, Society of Underwater
Technology, London, pp. 963–970.
Sassi, K., Zehzouh, S., Blanc, M., Thorel, L., Cathie, D., Puech, A. and Colliat-
Dangus, J-L. (2018) Effect of seabed trenching on the holding capacity of suction
anchors in soft deepwater Gulf of Guinea clays – Centrifuge testing and numerical
modelling. Proc. Offshore Technology Conference, Houston, Texas, USA, Paper
OTC-28756-MS.
Saviano, A. (2015) Effects of misalignment on the undrained HV capacity of suction
anchors in clay, MSc. Thesis, TU Delft, The Netherlands.
Saviano, A. and Pisanò, F. (2017) Effects of misalignment on the undrained HV
capacity of suction anchors in clay. Ocean Engineering, Vol. 133, pp. 89–106.
Schneider, J.A., White, D.J. and Lehane, B.M. (2007) Shaft friction of piles in sili-
ceous, calcareous and micaceous sands. Proc. 6th Int. Conf. Offshore Site
Investigation and Geotechnics. OSIG 2007, Society of Underwater Technology,
London, pp. 367–382.
Seibold, S. and Berger, W.H. (2010) The seafloor: an introduction to marine geology,
3rd edn, Springer-Verlag, Germany, 272 pp.
Seidel, J.P. and Collingwood, B. (2001) A new socket roughness factor for predic-
tion of rock socket shaft resistance. Canadian Geotechnical Journal, Vol. 38, no.
1, pp. 138–153.
Senders, M. (2005) Tripods with suction caissons as foundations for offshore wind
turbines on sand. Proc. 1st International Symposium on Frontiers in Offshore
Geotechnics, Taylor & Francis, Perth, Australia, pp. 397–405.
References 297

Senders, M. (2008) Suction caissons in sand as tripod foundations for offshore wind
turbines, PhD thesis, School of Civil and Resource Engineering, UWA, Australia.
Senders, M. and Kay, S. (2002) Geotechnical suction pile anchor design in deep
water soft clays. Proc. 7th Annual Conference, Deepwater Risers, Moorings and
Anchorings, IBC, London, UK, 50 pp.
Senders, M. and Randolph, M.F. (2009) CPT-based method for the installation of
suction caissons in sand. ASCE Journal of Geotechnical and Geoenvironmental
Engineering, Vol. 135, no. 1, January, pp. 14–25.
Senders, M., Randolph, M.F. and Gaudin, C. (2007) Theory for the installation of
suction caissons in sand overlain by clay. Proc. 6th Int. Conf. Offshore Site
Investigation and Geotechnics. OSIG 2007, Society of Underwater Technology,
London, pp. 429–438.
Senpere, D. and Auvergne, G.A. (1982) Suction anchor piles – a proven alternative to
driving or drilling. Proc. Offshore Technology Conference, Houston, Texas, USA,
Paper OTC-4206-MS.
Shonberg, A., Harte, I., Aghakouchak, A., Brown, C.S.D, Andrade, M.P. and
Liingaard, M.A. (2017) Suction bucket jackets for offshore wind turbines: applica-
tions from in situ observations. Proc. TC 209 Workshop - 19th ICSMGE, 20
September 2017, Seoul, pp. 65–77.
Siciliano, R.J., Hamilton, J.M., Murff, J.D. and Phillips, R. (1990) Effects of jackup
spudcans on piles. Proc. Offshore Technology Conference, Houston, Texas, USA,
Paper OTC-6467-MS.
Simons, N., Menzies, B. and Matthews, M. (2002) A short course in geotechnical site
investigation, Thomas Telford Publishing, London, ISBN 9780727729484.
Skempton, A.W. (1951) The bearing capacity of clays. Proc. Building Research
Congress, Vol. 1, pp. 180–189.
Smith, E.A.L. (1960) Pile-driving analysis by the wave equation. Journal of the Soil
Mechanics and Foundations Division ASCE, Vol. 86, no. SM4, pp. 35–61.
Smith, I.M. and Chow, Y.K. (1982) Three-dimensional analysis of pile drivability.
Proc. First Int. Conf. Num. Methods in Offshore Piling, ICE, London, UK, 20 pp.
Smith, I.M. and Griffiths, D.V. (1998) Programming the finite element method, 3 rd
edn, Wiley, USA, 534 pp.
Smith, I.M., Hicks, M.A., Kay, S. and Cuckson, J. (1988) Undrained and partially
drained behaviour of end bearing piles and bells founded in untreated calcarenite.
Proc. 1st International Conference on Calcareous Sediments, 15–18 March 1988,
Perth, Vol. 2, A.A. Balkema, Rotterdam, pp. 663–679.
SOLCYP (2017) Design of piles under cyclic loading – SOLCYP recommendations,
Puech, A. and Garnier, J. (Ed.), ISTE – WILEY, France.
Sorensen, S.P.H., Ibsen, L.B., Augustesen, A.H. (2010) Effects of diameter on initial
stiffness of p-y curves for large-diameter piles in sand, in: numerical Methods in
Geotechnical Engineering. CRC Press, pp. 907–912.
Sørlie, E. (2013) Bearing capacity failure envelope of suction caissons subjected to
combined loading, MSc. Thesis, Norwegian University of Science and Technology,
Trondheim.
Stavropolou, E., Dano, C. and Boulon, M. (2019) Experimental investigation of
the mechanical behaviour of soft carbonate rock/grout interfaces for the design
of offshorewind turbines. Proc. E3S Web of Conferences 7th Int. Symp.
Deformation Characteristics of Geomaterials, Vol. 92, IS-Glasgow. doi:10.1051/
e3sconf/20199213006.
298 References

Stevens, J.B., Audibert, J.M. (1979) Re-examination of PY curve formulations. Proc.

Offshore Technology Conference, Houston, Texas, USA, Paper OTC-3402-MS.
Stevens, R.S., Wiltsie, E.A. and Turton, T.H. (1982) Evaluating pile drivability for
hard clay, very dense sand and rokc. Proc. Offshore Technology Conference,
Houston, Texas, USA, Paper OTC-4205-MS.
Stewart, D. (2005) Influence of jack-up operation adjacent to a piled structure. Proc.
1st International Symposium on Frontiers in Offshore Geotechnics, Taylor &
Francis, Perth, Australia pp. 543–550.
Sturm, H. (2017) Design aspects of suction caissons for offshore wind turbine foun-
dations. Proc. TC 209 Workshop – 19th ICSMGE, 20 September 2017, Seoul,
pp. 45–63.
Sturm, H., Nadim, F. and Page, A. (2015) A safety concept for penetration analyses
of suction caissons in sand. Proc. 3rd International Symposium on Frontiers in
Offshore Geotechnics, Taylor & Francis, Oslo, Norway, pp. 1393–1398.
Supachawarote, C. (2006) Inclined load capacity of suction caissons in clay, PhD
thesis. University of Western Australia, Australia.
Supachawarote, C., Randolph, M.F. and Gourvenec, S. (2004) Inclined pull-out
capacity of suction caissons. Proc. 14th International Offshore and Polar
Engineering Conference ISOPE, 23–28 May 2004, Toulon, France, International
Society of Offshore and Polar Engineers (ISOPE), Cupertino, pp. 500–506.
Suroor, H. and Hossain, J. (2015) Effect of torsion on suction piles for subsea and
mooring applications. Proc. 3rd International Symposium on Frontiers in Offshore
Geotechnics, Taylor & Francis, Oslo, Norway, pp. 325–330.
Svanø, G., Eiksund, G., Kavli, A., Langø, H., Karunakaran, D. and Tjelta, T.I. (1997)
Soil-structure interaction of the Draupner E bucket foundation during storm con-
ditions. BOSS, Vol. 97, pp. 163–176.
Taiebat, H.A. and Carter, J.P. (2000) Numerical studies of the bearing capacity of
shallow foundations on cohesive soil subjected to combined loading. Géotechnique,
Vol. 50, no. 4, pp. 409–418.
Taiebat, H.A. and Carter, J.P. (2002) A failure surface for the bearing capacity of
circular footings on faturated clays. Proc. 8th Int. Symp. Numerical Models in
Geomechanics (NUMOG), 10–12 April 2002, Rome, Italy, pp. 457–462.
Taiebat, H.A. and Carter, J.P. (2005) A failure surface for caisson foundations in
undrained soils. Proc. 1st International Symposium on Frontiers in Offshore
Geotechnics, Taylor & Francis, Perth, Australia, pp. 289–295.
Taiebat, H.A. and Carter, J.P. (2010) A failure surface for circular footings on cohe-
sive soils. Géotechnique, Vol. 60, no. 4, pp. 265–273.
Terente, V., Torres, I., Irvine, J. and Jaeck, C. (2017) Driven pile design methods in
weak rock. Proc. 8th Int. Conf. Offshore Site Investigation and Geotechnics.
OSIG 2017, Society of Underwater Technology, London, pp. 652–657.
Terzaghi, K., Peck, R.B. and Mesri, G. (1996) Soil Mechanics in Engineering Practice,
3rd edn, Wiley, New York.
Tho, K.K., Chan, N. and Paisley, J. (2015) Comparison of coupled and decoupled
approaches to spudcan-pile interaction. Proc. 3rd International Symposium on
Frontiers in Offshore Geotechnics, Taylor & Francis, Oslo, Norway, pp. 1317–1322.
Tho, K.K., Leung, C.F., Chow, Y.K. and Swaddiwudhipong, S. (2013) Methodologies
for evaluation of spudcan-pile interaction. Proceedings of the 14th International
Conference on Jack-Up Platform Design, Construction and Operation, London,
September.
References 299

Tjelta, T.I. (1995) Geotechnical experience from the installation of the europipe
jacket with bucket foundations. Proc. Offshore Technology Conference, Houston,
Texas, USA, Paper OTC-7795-MS.
Tjelta, T.I. (2015) The suction foundation technology. Proc. 3rd International
Symposium on Frontiers in Offshore Geotechnics, Taylor & Francis, Oslo,
Norway, pp. 85–93.
Tran, M.N. (2005) Installation of suction caissons in dense sand and the influence of
silt and cemented layers, PhD. Thesis, Department of Civil Engineering, University
of Sydney, Australia.
Tran, M.N., Randolph, M.F. and Airey, D.W. (2004) Experimental study of suction
installation of caissons in dense sand. Proc. 23rd International Conference on
Offshore Mechanics and Arctic Engineering, 20–25 June 2004, Vancouver,
Canada, Paper OMAE2004-51076.
Tran, M.N., Randolph, M.F. and Airey, D.W. (2007) Installation of suction caissons
in sand with silt layesr. ASCE Journal of Geotechnical and Geoenvironmental
Engineering, Vol. 133, no. 10, October 1, pp. 1183–1191.
Van den Berg, P. (1994) Analysis of soil penetration, PhD. Thesis, TU Delft, The
Netherlands.
Van Dijk, B.F.J. (2015) Caisson capacity in undrained soil: failure envelopes with
internal scooping. Proc. 3rd International Symposium on Frontiers in Offshore
Geotechnics, Taylor & Francis, Oslo, Norway, pp. 337–342.
Versteele, H., Jaeck, C., Silvano, R. and Bignold, D. (2015) Sinkage analysis of
A-shaped jack-up mudmats using the coupled Eulerian Lagrangian approach.
Proc. 3rd International Symposium on Frontiers in Offshore Geotechnics, Taylor
& Francis, Oslo, Norway, pp. 1323–1328.
Versteele, H., Kuo, M.Y.-H., Cathie, D., Sassi, K. and Zehzouh, S. (2017) Anchor
chain trenching – numerical simulation of progressive erosion. Proc. 8th Int. Conf.
Offshore Site Investigation and Geotechnics, OSIG 2017, Society of Underwater
Technology, London, pp. 971–977.
Vicente, V., López-Gutiérrez, J-S., Dolores Esteban, M., Alberdi, P., Imaz, M. and
Serraclara, J-M. (2017) Monopiles in offshore wind: preliminary estimate of main
dimensions. Ocean Engineering, Vol. 133, pp. 253–261.
Villalobos, F.A. (2007) Installation of suction caissons in sand. Proceedings 6th
Chilean Conf. on Geotechnics (Congreso Chileno de Geotecnia), Chile.
Vivatrat, V., Valent, P.J. and Ponterio, A.A. (1982) The influence of chain friction on
anchor pile design. Proc. Offshore Technology Conference, Houston, Texas, USA,
Paper OTC-4178-MS.
Vulpe, C., Gourvenec, S. & Power, M. (2014) A generalized failure envelope for
undrained capacity of circular shallow foundations under general loading.
Géotechnique Letters, Vol. 4, no. July–September, pp. 187–196.
Waltham, T. (2009) Foundations of Engineering Geology, 3rd edn, CRC Press,
London. doi:10.1201/9781315273488.
Wang, M.C., Demars, K.R. and Nacci, V.A. (1977) Breakout capacity of model suc-
tion anchors in soil. Canadian Geotechnical Journal, Vol. 14, no. 2, pp. 246–257.
Wang, M.C., Demars, K.R. and Nacci, V.A. (1978) Applications of suction anchors
in offshore technology. Proc. Offshore Technology Conference, Houston, Texas,
USA, Paper OTC-3203-MS.
Wang, M.C., Nacci, V.A. and Demars, K.R. (1975) Behavior of underwater suction
anchor in soil. Ocean Engineering, Vol. 3, no. 1, pp. 47–62.
300 References

Watson, P.G., Gaudin, C., Senders, M. and Randolph, M.F. (2006) Installation of
suction caissons in layered soil. Proc. 6th International Conference on Physical
Modeling in Geotechnics, Hong Kong, pp. 685–692.
Whitehouse, R.J.S., Harris, J.M., Sutherland, J. and Rees, J. (2011) The nature of
scour development and scour protection at offshore windfarm foundations.
Marine Pollution Bulletin, Vol. 62, no. 1, pp. 73–88.
Williams, A.F. and Pells, P.J.N. (1981) Side resistance rock sockets in sandstone,
mudstone, and shale. Canadian Geotechnical Journal, Vol. 18, no. 4, pp.
502–513.
Wind Europe (2018) Offshore wind in Europe - Key trends and statistics 2017,
February, 37 pp.
Wind Europe (2020) Offshore wind in Europe - Key trends and statistics 2019,
https://windeurope.org/data-and-analysis/product/offshore-wind-in-europe-
key-trends-and-statistics-2019/.
XG Geotools (2014) SPCalc Manual.
Xie, Y. (2009) Centrifuge model study on spudcan-pile interaction, PhD. Thesis,
Department of Civil and Environmental Engineering, National University of
Singapore, Singapore.
Xie, Y, Leung, C.F. and Chow, Y.K. (2017) Centrifuge modelling of spudcan-pile
interaction in soft clay overlying sand. Géotechnique, Vol. 67, no. 1, pp. 69–77.
Zienkiewicz, O.C., Chang, C.T. and Bettess, P. (1980) Drained, undrained, consoli-
dating and dynamic behaviour assumptions in soils. Géotechnique, Vol. 30, no. 4,
pp. 385–395.
Index

Page numbers in italics refer to figures and those in bold refer to tables.

A in-place response, 7, 8
internal stiffeners, 21, 42–44, 119
accuracy of prediction, 4–5, 210, 258 loads and load transfer, 25, 26, 27,
active suction, 23 28
adhesion factors, 95, 104 optimum lug level (OLL), 60, 216,
outer for torsional loading, 196–197 217
pile-rock, 225, 226, 228, 229, 229 partial factor design approach, 72,
pile-soil, 182, 185 76–77, 78
adjacent infrastructure, 256–261 riser tower anchors (RTA), 41–42;
alpha method, 120 see also suction foundations
anchor chain, 190, 221 anchor holding capacity, 41, 269–270,
boulder contact, 130–131 271
trenching, 48–49, 50, 178, 209, 210, API RP 2A, 74, 76, 165, 228, 248
212, 269–274 API RP 2GEO, 9, 56, 74, 76, 103, 106
failure model, 17, 272 API RP 2SK, 9, 56, 76
anchor foundations, 1, 2, 41–46 API RP 2T, 76
anchor piles, 42–44, 239 axial failure models, 15–17, 19,
geometry, 118–119, 166 176–182, 184, 192–194
groups, 238 axial friction, 61
in-place resistance, 212–213 axial pull-out load, 178, 179
installation hazards, 44–48 axial resistance, 187–194, 209–212
misalignment, 45, 45 drilled and grouted piles in rock,
over-penetration, 87–88 226–232
pull-out capacity, 59–60 driven piles in weak rock, 224–226
tilt, 45, 220–223 tensile, 262
buoy anchors, 42 axial shear stress, 249
design considerations, 7, 59, 60 axial soil resistance, 91
external stiffeners, 21
factors of safety, 72, 78
FPSO anchors, 42, 220–223 B
in-place resistance, 7, 8, 176–179,
180 back analysis, 95, 143, 248, 266
analysis, 198, 202, 209, 212–213 base failure, 23, 111–112, 117–120
caisson anchor, 65–66, 67 base moment, 249, 250
FPSO anchor design example, bearing capacity factor, 182, 184–186,
220–223 191
resistance reduction, 209–211, bearing capacity method, 90–91,
212 103–105, 114–120
301
302 Index

bearing failure, 63 carbonate sands, 88, 89, 163

best estimate, 53 carbonate soils, 93
installation parameters, 81 cemented layers/rock, 49, 89, 171
in sand, 140–141 lateral resistance of piles, 232–234
penetration resistance, 92–93 monopile installation, 83
reverse punch-through failure model, resistance of driven piles, 225
161–162 shaft friction, 229
soil resistance, 85, 86 centrifuge tests, 245, 271
boreholes, 56, 170–171 chain load, 221
Bothkennar soils, 274 chalk, 164–165, 225–226, 229, 251
boulders, 50, 89, 121–131 chipping, 232–234
braced foundations, 80, 235, 236–237, clay, 32, 53
238, 266 with boulders, 126–128
British Standards Institution (BSI), 55 calcareous/carbonate, 88
buckling, 50, 83, 125, 127 competent, 112, 127, 129, 145
cylinder buckling, 46 drained/undrained, 35–36
impact driving in weak rock, failure models, 12–17, 19, 179
165–168 foundation geometry, 20, 21
pile tip integrity, 128–129 friction set-up, 121
reducing risk of, 83, 165–166, 168 group foundations, 235
buoy anchors, 41, 42 in-place resistance, 176, 182–183,
186, 239
analysis, 198, 201–204, 205
C maximum lateral resistance,
caissons 187–189, 191
anchor foundations, 65–66, 67, in-place response, 245–247
117–120 cyclic loading, 248–250
in-place resistance, 65–66, 67, loading conditions, 35, 36
214–215 normally consolidated
large diameter, 139–142 failure modes, 179
lateral resistance, 190–194 friction set-up, 120–121
loading conditions and soil drainage, geotechnical investigation, 56
35, 39 installation hazards, 41–46
lug level, 190–194 operational hazards, 48–49, 50,
offset legs, 27 269–274
reverse punch-through failure, over sand, 61–62, 89, 146–147
158–163 overconsolidated, 80, 117–118, 121
suction caissons, 3, 44, 108–111 plug, 146–148
in mixed clay and sand layers, seams, 88–89
146–162 sedimentation environments, 32
trenching seabed, 273 settlements, 242
support foundations, 216–220 with shallow/dissolved gas, 261–262
wall protuberances, 21 weak, 43, 87–88, 126–127; see also
CAISSON_VHM, 9, 192–193, 200, 200 installation in clay
comparisons, 212–213 closed-ended piles, 37, 38, 184–185
FPSO anchor design example, cluster foundation, 236–237
220–223 coated/painted steel, 20, 42, 87, 107
functionality, 200, 200 coefficient of variation (COV), 52, 85
reduction factors, 209–212 competent soil, 106, 190
support foundation design example, clay, 112, 127, 129, 145
216–220 embedment ratios, 61, 112, 142
calcareous sediment, 88, 165, 225, reverse punch-through, 160–163
232–234 sand, 48, 61, 87, 142, 143
Index 303

under-penetration, 87, 224 sign conventions, nomenclature and

water pocket model, 153–158 reference point, 68–69
compressive V loads, 180, 184 desk studies, 53, 55
conductor wells, 50, 261 displacement, 27, 68, 69, 81, 107
cone penetration test (CPT), 36–37, 40, dynamic and cyclic loads, 241,
56 248–250
cone resistance, 90–92 excessive, 63
foundation tip pore pressure immediate, 242, 243
coefficient, 95–101 rock socket, 230
penetration resistance assessment, soil displacement, 62, 258, 259
90–101 static loads, 241–248; see also plug
unit end-bearing resistance, 92–95 heave
unit skin friction resistance, 92–95 DNV (1992), 40, 74, 77, 92–93
consolidation settlement, 242, 245–247 DNV-OS-J101, 243, 259
constant normal stiffness (CNS) DNV-RP-E303, 9, 112, 114, 119, 120
conditions, 230–231 DNV-RP-H103, 92, 106
continuum solutions, 243, 244, 254, DNVGL-RP-C212, 105–106, 248, 249
267 drained soil response, 11, 15, 18,
coral outcrops, 49 35–40, 184, 223–224,
coring, 177–181, 184, 192–193 239–240
torsional resistance, 194–195 dredge pumps, 87
creep settlement, 89, 142, 242, 247 drilled and grouted piles
cross plates, 20–21 axial resistance, 226–232
cross-over anvils, 82 CNS conditions, 230–231
currents, 34, 43, 249 hole stability, 227
cyclic loading, 227, 232, 234 installation, 83–84, 169–171,
displacement, 241, 248, 250 226–227
accumulated, 242, 252–253 removal/retrieval, 172–173
dynamic/non-dynamic, 241 in rock, 226–232
cylinder buckling, 46 weak rock, 168–171, 251–252
rock-grout interface, 230–231
drilling, 83–84, 168–171, 226–227
D Drive-Drill-Drive, 83, 168–169
data quality, 4–5 driven piles, 61
dead loading, 72 axial resistance, 224–226
decommissioning, 172–173 in clay, 182
deep (pile) foundations, 1, 7, 8, 9, 59, Drive-Drill-Drive, 83, 168–169
258 Gentle Driving of Piles (GDP), 82
failure modes/models, 15–17, 19 impact driving, 81, 82, 84, 163–168
installation stop criteria, 21 lateral resistance, 232
load and soil resistance factors, removal/retrieval, 172
72–73, 76–77 in sand, 142–143
deep water, 3, 33, 56–58, 198 skin friction resistance, 142–143
depositional environments, 31, 32, 33 in weak rock, 163–168, 224–226,
design basis, 59–61 250–253
foundation stiffness and fixity, driving shoe, 165, 168
69–72 dropstone, 126–127
grouped foundations, 235
in-place resistance, 63–66, 67 E
in-place response, 68, 241–242
installation/retrieval/removal, 61–62 effective area method, 9, 12
load and material factors, 72–73, effective stresses, 35, 131, 132, 225,
74–77, 78 239
304 Index

embedment ratios, 8, 21–23, 59–60 reverse punch-through failure,

in clay, 112, 113 158–163
competent soils, 61, 112, 142 rotational failure, 9, 11, 13, 14, 60,
and maximum lateral resistance, 65, 270
187–189 shallow foundations, 9–12, 15, 17
offshore wind turbines (OWTs), 20, suction installation, 43–44, 61–62
22 tensile V loads, 176–179, 181–182,
in sand, 142, 144 184
suction foundations, 20, 21, 112, undrained soil, 13–17
113 water pocket model, 146–158
in weak rock, 61 field vane shear tests, 56
end-bearing failure, 111, 264–265 finite element analysis (FEA), 12, 14,
end-bearing resistance, 92–95, 103, 179, 201–202, 267–270
110–111, 178, 227 fins, 21
factors, 105 fixed head anchor pile, 191–194, 270,
and skin friction, 186–187 272
undrained clay soil response, 182 fixed head resistances, 187–194,
water pocket model, 147, 149 214–215
environmental (live) loading, 72 fixed platforms, 3, 56–57, 71
Eurocode, 144 floating distribution semi-submersible
external stiffeners, 21 (FDS), 45
extrusion buckling, 83, 166 floating production storage and
off-loading, see FPSO anchors
foundation tip
F hydraulic leak, 178
facies, 31 integrity, 121, 127–129, 267
factor of safety (FOS), 7, 53, 59–60, 72, pore pressure, 95–103
74–77, 78, 87, 114, 200 resistance, 132
failure modes/models, 9–19, 62, 175 FPSO anchors, 41, 42, 178, 213,
anchor chain trenching, 178, 270, 220–223, 235
272, 273 chain trenching, 269, 271
axial failure, 15–17, 19, 176–182, free head resistance, 187–191, 214–215
184, 192–194 free-fall penetration, 107–111
base failure, 23, 111–112, 117–120 free-fall risk, 80–81
combined VHM loading, 13–17, 18, friction, 42
192–194 friction fatigue, 164
compressive V loads, 180, 184 inner friction, 80, 131, 132, 148,
coring, 88, 177–181, 184 149
deep foundations, 15–17, 19 set-up, 120–121, 142–143
drained soils, 18 shaft friction, 164, 224–228,
end-bearing failure, 111, 264–265 229–231
installation in clay, 61–62, 88, soil response, 104, 105; see also skin
111–114, 146–163 friction
installation in sand, 10, 15, 17, 18,
131–134, 146–163 G
intermediate foundations, 12–15,
16–17, 18 gapping, 176, 209–212
leaking, 177–181 gas, see oil and gas applications, see
over-penetration, 88 shallow gas
overturning HM loads, 65 Gentle Driving of Piles (GDP), 82
plugged, 88, 177–181, 184, 192–195 geohazards, 31–34, 49–50, 89, 121–131
reverse end-bearing failure, 23, 62, geophysical surveys, 31, 33–34, 56–57
111–112, 177, 185 geotechnical analysis, 25–30, 257–258
Index 305

geotechnical data, 51–52 hydraulic fracture, 61, 62, 145, 146

geotechnical design, 52, 53, 261–262 hydraulic leaks, 23, 33, 133, 178
geotechnical loads, conversion, 28–30
geotechnical surveys, 33, 40, 56–57
Girassol (offshore Angola), 41–42, 43 I
glacial till, 33, 46, 87, 89, 121, 127– impact driving, 81, 82, 84, 163–168
128, 129 in-place resistance, 7, 8, 175–240
Gorm Field (North Sea), 47–48, 133 and adjacent infrastructure, 259,
ground investigation, 53 261
group resistance, 235–238 anchor foundations, 177–179
grouting, 169, 171, 268–269; see also analysis, 198, 202, 209, 212–213,
drilled and grouted piles 220–223
Gulf of Guinea FPSO anchor design example,
clays, 93, 94, 95 220–221
anchor chain trenching, 48–49, resistance reduction, 209–211,
50, 269–274 212
assessments, 85, 86
H analysis methods, 198–200
models/software, 9, 200, 200; see
hazards also VHM(T) resistance
from human activity, 51 envelope methods
installation, 41–48, 49–52, 54, axial resistance, 187–194, 209–211
165–168 drilled and grouted piles in rock,
in clay, 41–46, 49 226–232
cylinder buckling, 46 driven piles in weak rock,
difficult soil profiles, 88–89 224–226
excessive misalignment, tensile, 262
45, 45 clay soils, 176, 182–183, 186, 239
impact driving in rock, 163 maximum lateral resistance,
low penetration resistance, 41–46 187–189, 191
pile buckling, 83 maximum torsional resistance,
in sand, 46–48, 49, 50, 61, 62 194–195
sand plug liquefaction, 46–48, 61, twist, 196–197
62 combined VHM(T) loads, 13–17,
underpressures close to/above 18, 197, 198, 199
critical, 48 design basis, 63–66, 67
natural geohazards, 31–34, 49–50, design examples, 214–221
89, 121–131 drained soil response, 18, 184, 201,
operational, 52–53 223–224, 239–240
anchor chain trenching, 48–49, failure modes/models, 9–19, 175
50, 178, 209, 210, 212, combined VHM loading, 13–17,
269–274 18, 192–194
excessive scour, 48, 52 compressive V loads, 180, 184
risk minimisation, 52, 57 deep foundations, 15–17, 19
uncertainties, 51–52 drained soils, 18
high estimate, 48, 53 intermediate foundations, 12–15,
installation parameters, 81 16–18
in sand, 141, 142 pure axial (pull out) V load, 177
penetration resistance, 92–93 pure V load, 193
reverse punch-through failure model, shallow foundations, 9–12
161, 162 tensile V loads, 176–179,
soil resistance, 85, 86 181–182, 184
hole stability, 227 undrained soils, 13–17
306 Index

group resistance, 235–239 immediate displacement, 242, 243

lateral, 187–194, 209–211, 232–235 primary consolidation, 242,
lateral loads, 176 245–247
loading conditions and soil response, regional settlements, 242, 248
175 secondary compression (creep), 242,
maximum axial resistance, 176–187, 247
209–211 shallow/dissolved gas, 262
bearing capacity factor, 182, soil reaction, 249, 250
184–186 and soil type, 242
clay soil response, 182–183 static loads, 241–248
drained soil, 184 support foundations, 253–254
end-bearing, 186–187 weak rock, 250–253
sand soil response, 183–184 inner friction, 131, 132
skin friction, 182, 186–187 water pocket model, 148, 149
undrained soil, 177–184 installation, 7, 79–173
maximum lateral resistance, assessment, 79–81, 85, 90–105, 111,
209–211 120
in clay, 187–189, 191 observational method, 144–145
and lug level, 190–194 presentation, 171–172
suction anchor piles, 190–194 in clay, see installation in clay
support foundations, 187–189 data, 132–134, 135
maximum torsional resistance, back analysis, 95, 143, 248, 266
194–195 defined, 23, 79
non-co-planar MH loads, 64–67 depth, 56–57
reduction, 209–211 design basis, 59, 61–62
sand, 183–184, 201, 223–224, drilled and grouted piles, 83–84,
239–240 169–171, 226–227
and scour, 256 feasibility, 80, 84, 88–89
shallow penetration, 224 maximum pump underpressure,
suction anchor piles, 195–197 90
support foundations, 180 soil profiles, 88–89
analysis, 198, 210, 212, 217–218, free-fall penetration, 107–111
219–220 hard ground, 83, 84
design example, 216–219 hazards, see hazards
maximum lateral resistance, hybrid methods, 83, 168–169
187–189 impact driving, 81, 82, 84, 163–168
resistance reduction, 209–211, jack-up spudcan-foundation
212 interaction, 257–259, 260
tension cracks/gapping, 176 landing on seafloor, 105–106,
tilt and twist, 195–197, 220–221 263–265
undrained soils, 13–17, 175, lowering velocity, 105–106
177–184, 193 maximum pump underpressure, 90
clay, 182–183, 195, 198–200, model parameters, 81
201–204, 205, 239 monopiles, 81–84, 108–111
sand, 183–184 noise emissions, 81
VHM(T) interaction effects, 239, over-penetration, 85, 87–88
240 penetration resistance, 90–105
weak rock, 224–234 resistance, see installation
in-place response, 7, 8, 9, 241–254 resistance
and adjacent infrastructure, risk minimisation, 52–53, 57, 83–85
257–261 in rock, see installation in (weak)
cyclic loads, 241, 242, 247–250, rock
252–253 in sand, see installation in sand
design basis, 68 seafloor conditions, 33–34
Index 307

self-weight penetration, see self- shallow/dissolved gas, 261–262

weight penetration stop criteria, 21
shallow water, 90, 132, 133 suction pile, 129–130
and soil conditions, 31, 33 water injection, 87
and soil profiles, 88–89 water pocket formation model,
soil reactions, 267–268 146–157
stop criteria, 21 weak clay, 43, 88, 126–127, 163
suction assisted, 33, 79–81, 113–114 installation in sand, 173
under-penetration, 85, 87 back analysis of installation data,
and vent design, 263–264 143
vent resistance, 108–110, 264 competent sand, 48
water depth, 56–57, 90 design challenges, 131–134,
installation in clay, 111, 173 145–146
anchor pile caisson, 117–120 foundation tip resistance, 132
base failure, 111–113, 117–120 piping, 105–106, 131, 133, 134
boulders, 121–131 reduced inner friction, 132
anchor chain contact, 130–131 sand plug heave, 131, 133, 134
forces on, 122–125 sand plug liquefaction, 131–134
refusal, 126–128 scouring, 105–106
suction piles, 129–130 difficult soil profiles, 88–89
design challenges, 111–114, embedment ratios, 142, 144
145–146 failure modes/models, 10, 15, 17,
difficult soil profiles, 88–89 61–62, 131–134, 146–163
embedment ratios, 112, 113 friction set-up, 142–143
failure modes/models, 61–62, 88, hazards, 46–48, 49–51, 61, 62
111–114, 146–163 installation resistance, 134, 141
free-fall penetration, 107–111 models, 81, 134–142
friction set-up, 120–121 reverse punch-through failure,
hazards, 41–44, 49–51 158–163
installation/retrieval model, 114–120 water pocket model, 146–158
internal friction, 80 observational method, 144–145
model parameters, 81 penetration resistance, 90, 92, 93,
normally consolidated clay, 88, 94, 95–101
95, 163 permanent passive suction, 262–263
anchor chain trenching, 48–49, retrieval/removal, 142
50, 178, 209, 211, 212, sand plug heave, 62, 96, 131, 133,
269–274 134
excessive misalignment, 44–46, scour, 48, 52, 105, 256
45 self-weight penetration, 107, 263
free-fall penetration, 108–111 stop criteria, 21
low penetration resistance, 41–44 suction pump design, 265–266
over sand, 61–62, 89, 145–146 under clay, 89, 145–146
reverse punch-through failure, reverse punch-through failure,
158–163 158–163
water pocket model, 146–158 water pocket model, 146–158
over-penetration, 87–88 water flow, 80, 95–103
overconsolidated clay, 117–120 installation in (weak) rock, 81–84,
penetration resistance, 90, 91, 93–95 163–171
pile tip forces, 122–125 crushed/remoulded rock zone,
plug heave, 62, 88, 113–114, 146, 164–165, 224, 225, 251–252
147 drilled and grouted, 83–84, 169–171
plug hydraulic fracture, 62, 145, 146 drilling, 83–84, 168–171, 226–227
reverse punch-through failure, Drive-Drill-Drive (DRD), 83,
158–163 168–169
308 Index

embedment ratios, 61 ISO 19901-4:2016, 4, 9, 56, 61, 63, 68,

impact driving, 81, 163–168 75, 77, 103, 104, 106, 112,
refusal, 165 120, 158, 182, 184, 186, 201,
risk of buckling, 165–168 224, 241, 248, 256
resistance, 224 ISO 19901-7:2013, 72, 76
vibratory, 81–83, 171 ISO 19901-8:2014, 56
installation resistance, 85 ISO 19902:2007, 74, 77, 105, 228
in clay, 118, 119 ISO 19903-1:2019, 269
and over-penetration, 88 ISO 19905-1:2016, 182, 185, 258, 259
in sand, 134, 141
water pocket model, 148
instrumentation, 266 J
intermediate foundations, 1–4, 8 jack-up spudcans, 257–259, 260, 261
applications, 3 jacket structures, 163, 238, 248, 266,
compared to deep and shallow 268
foundations, 7–17, 72–73 jacket movement, 259
defined, 7, 8, 23 structural analysis, 69–70
design considerations, 7, 255–274 with suction buckets, 2, 3, 183
failure modes/models, 12–15, 16–18
geometry, 20–22, 43, 59–60, 114
anchor piles, 118–119, 166 K
optimum, 216–223 Kelvin-Voigt model, 253
support foundations, 213
in-place response, 7, 8
installation stop criteria, 21 L
interaction with adjacent
L6-B platform, 236–237
infrastructure, 256–261
laboratory tests, 40, 51, 53, 58, 183,
loading conditions and soil drainage,
263
35–40
Laminaria (Timor Sea), 41–45
types, 1, 2
large diameter foundations, 139–142,
uncertainties and risk minimisation,
165, 166, 196, 248–249, 265
51–53, 57
lateral failure models, 15, 16, 18, 19
uses, 1; see also design basis;
lateral fixity, 71–72
installation
lateral resistance
internal friction, 80
increasing with fins, 21
internal scoop, 217
modifying, 210–212
internal stiffeners, 20–21, 42–44, 80,
in rock, 232–235
119
soil reaction, 249
driven piles, 165, 166
support foundations, 187–190
and plug heave, 113–114
laterally loaded foundations, 28, 248
investigation, 55–58
pile groups in clay, 238–239
desk studies, 53, 55
leaking, 177–181, 192–193
geophysical surveys, 31, 33–34,
limit equilibrium methods, 61, 63, 185
56–57
liquefaction, see sand plug liquefaction
geotechnical data, 51–53
live loading, 72
geotechnical design, 52, 53, 261–262
Load and Resistance Factor Design
geotechnical surveys, 33, 40, 56–57
(LRFD), 72, 73–76, 77
ground investigation, 52
loads, 8, 25–30
in-situ tests, 40, 261
asymmetric, 63
inadequate, 51
axial pull-out load, 178, 179, 180
laboratory tests, 40, 51, 53, 58, 263
centric loads at seafloor, 27–28,
seafloor conditions, 33–34
68–69
uncertainties and risk minimisation,
co-planar, 63–65
51–53, 57
Index 309

combined VHM(T) loads, 7, 8, 12, M

13, 27–28, 68–69, 71,
197–198, 205 marine geology/sediments, 31–34,
compressive, 180, 184 88–89
conversion (structural to misalignment, 45, 45, 195–197
geotechnical), 28–30 mitigation measures, 83, 134, 168, 176,
dead/live, 72 274
design approaches, 72, 73–76, 77 monopiles, 2, 3
and displacement, 243 diameter, 20, 22, 39
extreme loading conditions, 72 embedment ratios, 20, 22
and failure modes/models, 9–19 foundation geometry, 20, 22
foundation fixity, 71–72 in-place response, 248–250
foundation stiffness, 69–72, 243, installation methods, 81–84
244 laterally loaded, 243
foundation type comparison, 7, 8, pile free-fall penetration, 108–111
9–19 retrieval/removal, 172–173
group, 235 risk of buckling, 46, 166, 167, 168;
HM loads, 18, 63–66, 67, 197, see also offshore wind turbines
216–217, 243 (OWTs)
in-place capacity, 59, 60 mooring systems, 41–45, 269–274
installation assessment, 80 design and investigation, 55, 61, 213
load interaction, 1, 7, 59, 240 loads, 27, 195, 235, 238–239
loading conditions and soil drainage, mud, 227
35–40, 63, 175 mudmats, 72, 200, 238
dynamic loading of solid piles, Mya North Field (Bay of Bengal), 216
37–38
effect of loading rate on bearing N
response, 36–37
laterally loaded pile, 38–39 Na Kika (Gulf of Mexico), 44–45, 45
maximum axial load, 180 natural frequency, 27, 248
mooring systems, 27, 195, 235, noise emissions, 81–82
238–239 non-carbonate rock, 95, 228, 252
non-co-planar MH loads, 63–66, 67 North Rankin “A” platform (North
overturning, 63–66, 67, 215, 217 West Shelf, Australia), 4–5
partial load factors, 72, 73–76, 77 North Sea
point of load transfer, 1, 26, 27–28, clays, 93, 94
68–69 installation data, 132–133
pure V loading, 9–12, 176–182 installation model examples
restoring, 63–66, 67 reverse punch-through, 160–163
seafloor VHM(T) loads, 8, 27–28, sand installation, 139–142
68–69, 71, 187, 205 water pockets, 153–157
sign conventions, 27, 28, 30 platforms, 236–237, 245
and soil resistance, 73–76, 85, 86 suction anchor piles, 46–48, 47 133
structural vs. geotechnical North West Shelf (Australia), 4–5
perspectives, 25, 26, 28
tensile, 177–179, 180, 181–182, 184 O
torsional T load, 195–198, 199
units, 25, 27, 30; see also cyclic observational method, 144–145
loading obstructions, 34, 87, 126, 133
lowering velocity, 105–106 offset legs, 27–28, 27–28, 69
Luce Bay, 33 offshore wind turbines (OWTs), 2, 3, 4
lug levels, 59, 60, 190–194, 216, 217 analysis methods, 201
lumped factor of safety (FOS) design, 248–249
approach, 72, 73–76, 77 embedment ratios, 20, 22
310 Index

environmental conditions, 248, 249 P

foundation geometry, 22
in-place response, 243, 248–250 p-y curves, 232, 233, 234, 248–251,
installation feasibility, 84 253, 258, 259
installation hazards, 45 padeye level, see lug levels
installation methods, 81–84 paint, 20, 42, 87, 107
drilled and grouted, 83–84 partial drainage, 36, 39, 40
impact driving, 81, 82, 84 partial factor design (PFD) approach,
vibratory, 81–83, 84 72, 73–76, 77
jack-up spudcan interaction, passive suction, 23, 177–178, 262–263,
257–261 272
lateral loading, 243 peat, 88–9, 242
pile free-fall penetration, 108–111 penetration depth, 85, 87–88
retrieval/removal, 172–173 resistance at shallow penetration,
shaft friction, 226 224
size, 20, 22 and reverse punch-through, 161–162
substructures, 3 and water pocket formation,
oil and gas applications, 3, 48–49, 51, 154–157
167, 253–254 penetration rate, 36–37
Girassol (offshore Angola), 41–42, penetration resistance, 33
45 assessment, 90–105
Gorm Field, 133 classical bearing capacity method,
Laminaria (Timor Sea), 41–45 90–91, 103–105
Mya North Field (Bay of Bengal), CPT method, 90–101
216 involving water flow, 95–103
Na Kika (Gulf of Mexico), 44–45, high, 48
45 installation in clay, 114–115, 118
Serpentina Field (Gulf of Guinea), installation in sand, 135–142
48, 49, 270 low, 41–45
Skarv Field, 89; see also protuberances, 104
platforms permeable soils, 95–101
open-ended piles, 37, 164–166, 167, pile cutting, 172–173
184–185 pile driving, 81–83, 163–168
Ophir well head platform, 257, 259, pile dynamics, 107–108
260 pile foundations, see deep (pile)
optimum lug level (OLL), 60, 190–194, foundations
216, 217 piles
over-penetration, 85, 87–88 anchor piles, 42–45, 239
overconsolidated clay, 80, 117–118, geometry, 118–119, 166
121 groups, 238–239
overconsolidated sand, 32, 33 in-place resistance, 214
overpressure installation hazards, 44–46
defined, 23 misalignment, 44, 45
installation in sand, 134 over-penetration, 87–88
installation/retrieval in clay, pull-out capacity, 59–60
117–120 tilt, 45, 221–223
maximum allowable, 111 buckling, 46, 50, 83, 125, 127–129,
retrieval, 118, 119, 119 165–168
retrieval/removal, 61, 79, 269 closed-ended, 37, 38, 184–185
and top plate vent design, 263–264 drilled and grouted, 83–84, 169–
overturning HM loads, 63–66, 67, 215, 171, 226–232, 251–252
217 dynamic loading, 37–38
Index 311

geometry, 43, 118–119, 166 plug separation, 61, 62

installation plugged failure model, 88, 177–181,
free-fall penetration, 107–111 184, 192–195
hazards, 45–48, 165–168 Pluto (offshore Australia), 227
impact driving, 81, 82, 84, pore pressure, 63, 95–103, 225, 272
163–168 in chalk, 225
self-weight penetration, 110–111 dissipation, 39, 120
jack-up spudcan interaction, 259, preload measurements, 266
260 primary consolidation settlement, 242,
large diameter, 165, 166, 196, 245–247
248–249 protuberances, 21, 88, 104, 195, 228
laterally loaded, 38–39 pseudo code
open-ended, 37, 165 free-fall model, 107–108
driven in chalk, 164 installation in sand, 137–139
plugged, 184–185 installation/retrieval in clay,
tip distortion, 165–166, 167 115–117
partial load and soil resistance water pocket model, 152–153
factors, 75–76 pull-out capacity, 59–60, 172
pile tip forces, 122–125, 128 punch-through, 10, 50, 158–163
pile tip geometry, 166, 167 putty chalk, 225
pile tip integrity, 128–129
refusal, 126–130, 165
socket piles, 228–232 R
suction anchor piles, 45, 46
lateral capacity, 190–194 reference point, 68–69
load and material factors, 72, refusal, 33, 46, 85
75–76 boulders, 50, 126–128
tilt and twist, 195–197 cemented layers/rock, 50
suction piles, 3, 43, 45, 46, 259, 260 impact driving in weak rock, 165
refusal, 126–127, 129–130 regional settlements, 51, 242, 248
tension cracks/gapping, 176 remoulded rock, 164–165, 224, 225,
torpedo piles, 2, 107–108 251–252
weight, 88, 118; see also driven removal, 23, 61, 79, 81, 172–173
piles; monopiles friction set-up in sand, 142–143, 172
pipelines, 51, 261 grouted foundations, 269
piping, 23, 61, 62, 105–106, 131, 133, OWT monopiles, 172–173
134 soil reactions, 267–268; see also
piping failure, 133 retrieval
PISA (Pile Soil Analysis), 249, 250–251, renewables, see offshore wind turbines
250 (OWTs)
platforms, 3, 48–9, 56–57, 71 restoring HM loads, 63–66, 67
jack-up spudcan interaction, retrieval, 23, 61, 79, 80
257–259, 260 assessment, 119–120, 172–173
L6-B, 236–237 clay installation/retrieval model,
North Rankin “A”, 4–5 114–120
North Sea, 236–237, 245 grouted foundations, 269
Ophir well head platform, 257, 259, OWT monopiles, 172–173
260 resistance in clay, 118, 119
support platforms, 253, 254 soil reactions, 267–268
plug heave, 23, 48, 61–62, 88, 180 suction foundations, 118–120, 172;
in clay, 62, 88, 113–114, 146, 147 see also removal
in sand, 62, 96, 131, 133, 134 Reverse Circulation Drilling (RCD), 83
312 Index

reverse end-bearing, 177–179, 271–274 design challenges, 131–134

reverse end-bearing failure, 23, 62, installation model, 140–142
111–112, 177, 185 sand waves, 33–34, 51, 52
reverse flow, 134 scour, 48–49, 52, 105, 256
reverse punch-through failure, 158–163 seafloor, 33–34, 51, 56, 63, 69
ring stiffeners, 21, 42–45, 80, 113–114, end-bearing failure, 264–265
119 landing on, 105–106, 263–265
riser tower anchors (RTA), 41–42 scour, 48–49, 52, 105, 256
risks slopes, 33, 34, 69
buckling, 46, 83, 165–168 trenching, 49, 50, 269–274
during drilling, 83–84, 170 seafloor velocity, 108, 109, 110
free-fall, 80–81 seafloor VHM(T) loads, 8, 27–28,
minimisation, 52–53, 57, 83–85 68–69, 71, 187, 205
monopile installation, 83–84, 166, secondary compression (creep)
167, 168 settlement, 242, 247
rotation, 45, 69; see also tilt; twist sediment types, 31, 32, 33–34
rotational failure, 9, 11, 13, 14, 60, 65, seepage flow, 95–103, 131–133,
270 140–142
rotational fixity, 71–72 seismic surveys, 31, 57
rotational resistance (base moment), self-weight penetration (SWP), 33, 42,
249, 250 80–81, 110–111
anchor piles in weak clay, 118
minimum, 106–107
S over-penetration, 87
sand, 32, 53 plug heave in clay, 113–114
aquifer flowrate, 147, 150, 155, 157 sand installation, 142
boulders in, 129–130 top plate vent design, 263–265
carbonate, 89 sensitivity analyses, 100, 157, 158
competent, 48, 61, 87, 142, 143 Serpentina Field (Gulf of Guinea), 49,
drained soil response, 35, 36, 184, 50, 269
201, 223–224, 239–240 set-up, 81
foundation geometry, 21 in clay, 120–121
in-place resistance, 183–184, 201, defined, 23
223–224, 239–240 design basis, 61
loading conditions in sand, 142–143
cyclic loading, 248–250 settlements, 51, 241–242, 245–248
short-term loading, 240 shaft friction, 164, 224–228, 230–231
and soil drainage, 35–40 shaft resistance coefficient (SRC),
sustained load, 239–240 229–230
loose, 88 shallow foundations, 1, 7, 8, 9
over clay, 88 adjacent interactions, 261
overconsolidated, 32, 33 groups, 238
plug, 146–157; see also sand plug load and displacement, 27
liquefaction loading conditions and soil drainage,
plug heave, 62, 96, 131, 133, 134 35
seafloor, 33–34, 51, 52, 105–106 partial load and soil resistance
settlements, 242 factors, 73–74
undrained soil response, 15, 35, 36, point of load transfer, 27–28
183–184, 240; installation in VHM(T) resistance envelopes, 202
sand shallow gas, 51, 261–262
sand plug liquefaction, 23, 47–48, 61, shallow penetration, 224
62 shallow rotational failure, 60, 65
Index 313

shallow water, 33, 56–57, 90, 132, 133 soil plug

shear keys, 228 failure, 61–62, 111, 112, 117–120
shear strength, 182–183, 188, 190–191, forces on, 245–247
249 hydraulic fracture of clay plug, 62,
undrained shear strength, 40–42, 145, 146
274 permeability, 96–103
short-term loading, 240 stability ratio, 112; see also plug
sign conventions, 27, 28, 30, 68–69 heave
silt, 36–37, 42–45 soil reaction, 249, 250, 267–268
Skarv Field suction anchors, 89 soil resistance, 73–76, 85, 86, 88, 147,
skin friction, 92–95, 104, 105, 110– 148, 155, 157
111, 120–121 soil resistance to driving (SRD), 164,
driven piles in sand, 142–143 165
driven piles in weak rock, 164 soil wedge, 14, 19, 42, 44
and end-bearing, 186–187 South China Sea, clays, 94, 95
inner, 247 spudcan foundations, 36–37, 51,
outer, 245, 247 257–259, 260, 261
primary consolidation settlement, squeezing, 10
245–247 Statnamic Rapid Pile Load Testing, 37–38
socket piles, 228 static (dead) loading, 72, 241–248
undrained clay soil response, 182 steady state conditions, 101–103
skirted foundations, 12, 33, 245 steel design, 267
penetration resistance, 91–92 stick-up, 23, 69
piping around, 105–106 stiffness, 7, 69–72, 243, 244
slenderness ratio, 20, 22 matrices, 68–69, 71, 243, 244
socket piles, 228–232 risk of buckling, 165, 166
soil strains, 30
drainage response, 35–40, stresses, 30
63, 95 effective, 35, 131, 132, 225, 239
flow mechanisms around internal soil reaction, 267–268
stiffeners, 44 steel, 267, 268
gassy soils, 51, 261–262 structural analysis, 25–30, 69–71
lateral variability of top layers, structural loads, conversion, 28–30
31–34, 57 suction assisted penetration, 113–114,
movement induced by spudcans, 132–133
257–259 suction foundations, 2, 3, 46
North Sea soil profiles, 47 anchor piles, 45, 46
parameters, 268 lateral capacity, 190–194
partially drained, 36, 39, 40 load and material factors, 72,
properties, 31, 94 75–76
bearing modulus, 36–37, 40 tilt and twist, 195–197
coefficient of consolidation, anchors, 2, 3, 56, 89, 245
35–37, 40 anchor chain trenching, 49, 50,
consolidation, 38–39 271
drained/undrained, 35–40 embedment ratios, 112
dynamic drainage factor, 37–38 FPSO mooring, 213, 235
permeability, 95–103 twist, 195–196
pore pressure dissipation, 39 buckets, 2, 3
and response to loading, 35–40, caissons, 3, 44, 108–111
241–242, 249, 250 in mixed clay and sand layers,
type, 31, 32 146–162
uncertain geotechnical data, 51–52 trenching seabed, 273
314 Index

in clay, 42–45, 46, 111–118, 245 tension cracks, 176

boulders, 121–131 tilt, 23, 45, 80, 195, 221–223, 248
friction set-up, 120–121 Timoshenko beam element, 250
plug heave, 113–114 tip distortion, 165–166, 167
defined, 23 tip integrity, 50, 121, 127–129, 267
embedment ratios, 20, 21, 112, 113 top plates, 20, 61, 88, 183, 263–265
free-fall penetration, 108–111 torpedo piles, 2, 107–108
geometry, 20, 21, 43 torsion T load, 195–198, 199
groups, 235–239 torsional resistance, 194–195
installation, 33, 41–45, 46, 47–48, total stress methods, 225
61–62, 173 twist, 23, 80, 195–197
difficult soil profiles, 88–89
and retrieval, 114–120
investigation, 56–58 U
jack-up spudcan interaction, 259, uncertainties, 51–52, 95
260 unconfined compressive strength (UCS),
laboratory testing, 53 225, 226, 228–229, 231
lateral capacity, 190–194 uncontrolled penetration, 80–81
in mixed clay and sand layers, under-penetration, 85, 87
146–162 under-reaming, 83, 168, 227
over-penetration, 85, 87–88 underbase grouting, 268–269
pile refusal, 126–127, 129–130 underconsolidation, 51
primary consolidation settlement, underpressure, 23, 33
245 close to/above critical, 48
retrieval/removal, 118–120, 172 critical, 100–111, 134, 135
in sand, 47–49, 131–134, 173, 183 difficult soil profiles, 88–89
soil flow mechanisms, 44 installation in clay, 111, 118, 119
soil response to loading, 35, 36 over sand, 154–156, 157, 159,
suction piles, 3, 43, 45, 46, 259, 260 161
tilt and twist, 195–197 installation in sand, 131–136,
under-penetration, 85, 87 140–142
suction pump design, 90, 265–266 maximum allowable, 111–112, 132,
support foundations, 1, 2, 216–220 133
design considerations, 7, 59–60 maximum pump underpressure, 90
factors of safety, 72, 77 permanent, 262–263
geometry, 213 pump design and flow rate, 90,
in-place resistance, 180 265–266
analysis, 200, 210, 213, 216–220 reverse punch-through failure, 159,
maximum lateral resistance, 161
187–190 and under-penetration, 87
resistance reduction, 210–212 water pocket formation, 154–156, 157
in-place response, 245–247, undrained shear strength, 40–42, 274
253–254 undrained soil response, 11, 13–17,
internal stiffeners, 20–21 35–40, 175, 177–184, 193
loads and load transfer, 25, 26, clay, 12–14, 182–183, 195, 200,
27–28 201–205, 208, 239
partial factor design, 72, 77 sand, 15, 35, 36, 183–184, 240
rotational failure, 60 units, 25, 27, 30
V capacity, 59 upward seepage flow, 131–133, 145

T
V
target penetration depth (TPD), 85,
87–88, 114, 268 vacuum anchors, 263
tensile V loads, 176–179, 181–182, 184 vent design, 263–265
Index 315

vent resistance, 108–110, 264 water flow, 91, 95–101

VHM(T) loads, 7, 12, 13, 197–198 installation in sand, 131–133,
seafloor, 8, 27–28, 68–69, 71, 187, 140–142
205 water pocket model, 146–157
VHM(T) resistance envelope methods, water injection, 87
9, 201–224, 239–240 water plug, 263–266
design examples, 215–223 water pockets, 62, 100, 145, 162–163,
caisson support foundation, 268–269
216–220 model, 146–158
FPSO anchors, 221–223 wave loading, 36, 39, 40
drained sand soil response, 11, 15, weak clay, 43, 87–88, 126–127
18, 223–224 weak rock
equations, 208–209 drilled and grouted piles, 168–171,
MH ellipses, 64–65, 205–208, 210, 251–252
211, 212, 215–216 driven piles, 163–168, 224–226,
reduction factors, 210–212 251–253
resistance comparisons, 213–214 fracturing, 164–165
shallow foundations, 202 impact driving, 163–168
undrained clay soil response, 11, 12, in-place resistance, 176, 224–235
14, 17, 202–205, 208, 239 axial resistance, 224–226
V-Hmax ellipsoids, 193–194, 205, lateral resistance, 232–235
211, 211, 216, 217 in-place response, 250–253
Vlug-Hlug envelopes, 192–194, 270, remoulding, 164–165, 224, 225,
271 251–252; see also cemented
yield function, 209 layers/rock; installation in
vibratory installation, 81–83, 84 (weak) rock
vibratory retrieval/removal, 173 Wihelmshaven (Germany), 46
wind loading, 39, 85
W
Y
wash out, 226, 261
water depth measurements, 33 yield function, 160, 200, 209