Offshore Wind Farms

Offshore Wind Farms

Related titles Wind Energy Systems: Optimising Design and Construction for Safe and Reliable Operation (ISBN: 978-1-84569-580-4) Advances in Wind Turbine Blade Design and Materials (ISBN: 978-0-85709-426-1) Stand-Alone and Hybrid Wind Energy Systems: Technology, Energy Storage and Applications (ISBN: 978-1-84569-527-9)

Woodhead Publishing Series in Energy: Number 92

Offshore Wind Farms Technologies, Design and Operation

Edited by

Chong Ng and Li Ran

AMSTERDAM • BOSTON • CAMBRIDGE • HEIDELBERG LONDON • NEW YORK • OXFORD • PARIS • SAN DIEGO SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO Woodhead Publishing is an imprint of Elsevier

Woodhead Publishing is an imprint of Elsevier The Officers’ Mess Business Centre, Royston Road, Duxford, CB22 4QH, UK 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, USA The Boulevard, Langford Lane, Kidlington, OX5 1GB, UK Copyright © 2016 Elsevier Ltd. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. 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 A catalog record for this book is available from the Library of Congress ISBN: 978-0-08-100779-2 (print) ISBN: 978-0-08-100780-8 (online) For information on all Woodhead Publishing publications visit our website at https://www.elsevier.com/

Publisher: Joe Hayton Acquisition Editor: Cari Owen Editorial Project Manager: Alex White Production Project Manager: Omer Mukthar Designer: Vicky Pearson Esser Typeset by TNQ Books and Journals

Contents

List of contributors Woodhead Publishing Series in Energy Acknowledgments

Part One 1

2

3

Introduction to offshore wind energy and offshore wind farm siting

xi xiii xix

1

Introduction to offshore wind energy C. Ng, L. Ran 1.1 Wind energy 1.2 Offshore wind farm 1.3 Energy cost 1.4 Wind turbines 1.5 Disputable issues References

3

Economics of building and operating offshore wind farms P.E. Morthorst, L. Kitzing 2.1 Introduction 2.2 Investment costs 2.3 Operating costs 2.4 Key economic drivers for offshore wind energy 2.5 Levelised cost of energy 2.6 Future cost of offshore wind 2.7 Conclusions References

9

Wind resources for offshore wind farms: characteristics and assessment B.H. Bailey 3.1 Key issues in assessing wind resources 3.2 The nature of the offshore wind environment 3.3 Essential data parameters 3.4 Observational approaches

3 3 5 7 8 8

9 10 16 17 22 25 26 26

29 29 30 34 38

vi

4

Contents

3.5 Modeling approaches 3.6 Future trends Abbreviations Sources of further information References

44 52 53 54 55

Remote sensing technologies for measuring offshore wind M.S. Courtney, C.B. Hasager 4.1 Introduction 4.2 Conventional methods 4.3 Surface-based remote sensing 4.4 Space-borne RS 4.5 Case study e a near-coastal wind farm project 4.6 Future trends Sources of further information Abbreviations and Acronyms References

59

Part Two 5

6

7

Wind turbine components and design

Developments in materials for offshore wind turbine blades R. Nijssen, G.D. de Winkel 5.1 Key requirements for blade materials 5.2 Role of testing materials and structures in the blade design process 5.3 Case study on material selection and blade design 5.4 Future trends Abbreviations and nomenclature References

59 61 62 77 78 80 81 81 82

83 85 85 90 92 98 103 103

Design of offshore wind turbine blades P. Greaves 6.1 Introduction 6.2 Aerodynamics 6.3 Materials 6.4 Structural design 6.5 Manufacture Nomenclature References

105

Wind turbine gearbox design with drivetrain dynamic analysis S. McFadden, B. Basu 7.1 Introduction 7.2 WTGS gearbox design e concept stage 7.3 WTGS gearbox design e development stage

137

105 106 113 121 131 132 135

137 139 146

Contents

8

9

vii

7.4 WTGS gearbox design e production stage 7.5 Drivetrain dynamic analysis 7.6 Conclusions References

149 151 157 157

Design of generators for offshore wind turbines A. McDonald, J. Carroll 8.1 Introduction: key issues in generator design 8.2 Electrical generators: types and principles of operation 8.3 Practical design and manufacture of electrical generators 8.4 Selection of generators for offshore wind turbines 8.5 Future trends in offshore wind turbine generators Sources of further information References

159

Modelling of power electronic components for evaluation of efficiency, power density and power-to-mass ratio of offshore wind power converters R.A. Barrera-C ardenas, M. Molinas 9.1 Introduction 9.2 Semiconductors and switch valves 9.3 Filter inductors 9.4 Filter capacitors 9.5 Evaluation approach and design methodology 9.6 Evaluation example of a 1-MW 2L-VSC Nomenclature Symbols References

10 Design of offshore wind turbine towers R.R. Damiani 10.1 Introduction 10.2 Function and types of towers 10.3 Standards of reference 10.4 Design spiral process and loads’ analysis 10.5 Shell and flange sizing 10.6 Secondary steel, other structure details, and coatings 10.7 Optimization considerations 10.8 Final remarks Glossary List of symbols List of greek symbols Acknowledgments References

159 165 178 181 187 190 191

193 193 194 207 215 220 239 253 256 260 263 263 265 274 281 300 316 328 339 341 344 349 351 351

viii

Contents

11 Design of floating offshore wind turbines M. Collu, M. Borg 11.1 Introduction 11.2 Design of floating offshore wind turbines: main preliminary steps 11.3 Key issues in design of floating offshore wind turbines 11.4 Summary: case study 11.5 Future trends Nomenclature Sources of further information References

Part Three

Integration of wind farms into power grids

359 359 363 370 375 379 381 381 382

387

12 Offshore wind farm arrays O. Anaya-Lara 12.1 Fundamentals of offshore wind farm arrays 12.2 Design considerations 12.3 Main electrical components 12.4 Topologies 12.5 Converter interface arrangements and collector design 12.6 Wake farm arrangement e wake effects 12.7 Control objectives 12.8 Collector design procedure Abbreviations Acknowledgements References

389

13 Cabling to connect offshore wind turbines to onshore facilities Narakorn Srinil 13.1 Introduction 13.2 Offshore wind farm cables 13.3 Offshore cable installation, protection and challenges 13.4 Dynamic cables for floating wind turbines and substations 13.5 Some mechanical aspects of subsea cables 13.6 Outlook for offshore wind farm cables Abbreviations Acknowledgements References

419

14 Integration of power from offshore wind turbines into onshore grids O.D. Adeuyi, J. Liang 14.1 Introduction 14.2 Wind farm collection systems

441

389 390 390 397 399 410 411 412 415 416 416

419 420 426 431 434 437 438 439 439

441 441

Contents

14.3 14.4 14.5 14.6

ix

Offshore wind power transmission systems Voltage source converters Development of future submarine power transmission schemes Conclusions References

15 Energy storage for offshore wind farms D.A. Katsaprakakis 15.1 Introduction 15.2 The storage technologies 15.3 Indicative case studies: S-PSSs in Rhodes and Astypalaia 15.4 Conclusions Abbreviations References 16 Hydropower flexibility and transmission expansion to support integration of offshore wind N.A. Cutululis, H. Farahmand, S. Jaehnert, N. Detlefsen, I.P. Byriel, P. Sørensen 16.1 Introduction 16.2 Technologies 16.3 Summary e case study 16.4 Scenarios 16.5 Results 16.6 Conclusions References

Part Four

Installation and operation of offshore wind farms

17 Assembly, transportation, installation and commissioning of offshore wind farms M. Asgarpour 17.1 Introduction 17.2 Delivery of components 17.3 Onshore assembly 17.4 Offshore transport 17.5 Offshore installation 17.6 Tests and commissioning 17.7 Conclusions and future trends References

444 446 453 455 455 459 459 463 471 489 490 491

495

495 496 500 502 509 521 521

525 527 527 527 528 530 531 536 538 541

x

Contents

18 Condition monitoring of offshore wind turbines W. Yang 18.1 Reliability of offshore wind turbines 18.2 Challenges in offshore wind turbine operation and maintenance 18.3 Offshore wind turbine condition monitoring techniques 18.4 Offshore wind turbine condition monitoring systems 18.5 Signal processing techniques used for WT CM 18.6 Existing issues and future tendencies of WT CM References

543

19 Health and safety of offshore wind farms P.O. Lloyd 19.1 Limits of this chapter 19.2 Introduction 19.3 Legal framework 19.4 Safety management system 19.5 Plan, do, check, act 19.6 The offshore renewable energy industry 19.7 Plan 19.8 Do 19.9 Check 19.10 Act 19.11 For the future 19.12 Conclusion Abbreviations

573

20 Offshore wind turbine foundations e analysis and design B.C. O’Kelly, M. Arshad 20.1 Foundation options for offshore wind-turbine structures 20.2 System of loading on offshore foundations 20.3 General aspects of OWT monopile foundation system 20.4 Offshore design codes and methods 20.5 Investigation of monopileesoil behaviour 20.6 Design of OWT foundation 20.7 Future outlook and research needs Nomenclature Abbreviations References

589

Index

611

543 545 547 552 564 568 569

573 573 574 575 575 576 576 579 585 586 586 587 587

589 591 593 595 595 599 603 604 604 605

List of contributors

Cardiff University, Cardiff, United Kingdom

O.D. Adeuyi

O. Anaya-Lara

University of Strathclyde, Glasgow, United Kingdom

University of Engineering & Technology, Lahore, Pakistan

M. Arshad

M. Asgarpour Energy Research Centre of the Netherlands (ECN), Petten, The Netherlands; Aalborg University, Aalborg, Denmark AWS Truepower LLC, Albany, NY, United States

B.H. Bailey

R.A. Barrera-C ardenas Tsukuba, Tsukuba, Japan

Faculty of Pure and Applied Sciences, University of

B. Basu

Trinity College Dublin, Dublin, Ireland

M. Borg

Technical University of Denmark (DTU), Lyngby, Denmark

I.P. Byriel

Energinet.dk Tonne Kjærsvej 65 Fredericia

J. Carroll

University of Strathclyde, Glasgow, United Kingdom

M. Collu

Cranfield University, Bedford, United Kingdom

M.S. Courtney

Technical University of Denmark, Lyngby, Denmark

N.A. Cutululis

Technical University of Denmark, Roskilde, Denmark

R.R. Damiani

RRD Engineering, Arvada, CO, United States

N. Detlefsen

Danish District Heating Association Merkurvej 7, Denmark

G.D. de Winkel Knowledge Center Wind Turbine Materials and Constructions (WMC), Wieringerwerf, The Netherlands H. Farahmand Norwegian University of Science and Technology (NTNU), Trondheim, Norway P. Greaves Kingdom

Offshore Renewable Energy Catapult, Northumberland, United

C.B. Hasager Technical University of Denmark, Lyngby, Denmark S. Jaehnert

Sintef Energy Research, Trondheim, Norway

xii

List of contributors

D.A. Katsaprakakis Greece


L. Kitzing J. Liang

Technological Educational Institute of Crete, Heraklion,

Cardiff University, Cardiff, United Kingdom

P.O. Lloyd Siemens Centre of Competence EHS Offshore Beim Storhhause, Hamburg, Deutschland A. McDonald University of Strathclyde, Glasgow, United Kingdom Ulster University, Magee Campus, Northern Ireland, United

S. McFadden Kingdom

M. Molinas Department of Engineering Cybernetics, Norwegian University of Science and Technology, Trondheim, Norway Technical University of Denmark, Roskilde, Denmark

P.E. Morthorst C. Ng

Offshore Renewable Energy Catapult, Northumberland, United Kingdom

R. Nijssen Knowledge Center Wind Turbine Materials and Constructions (WMC), Wieringerwerf, The Netherlands B.C. O’Kelly L. Ran

Trinity College Dublin, Dublin, Ireland

University of Warwick, Coventry, United Kingdom

P. Sørensen


Narakorn Srinil W. Yang

Newcastle University, United Kingdom

Newcastle University, Newcastle upon Tyne, United Kingdom

Woodhead Publishing Series in Energy

1 2 3 4 5 6 7 8

9 10 11 12 13 14 15

Generating power at high efficiency: Combined cycle technology for sustainable energy production Eric Jeffs Advanced separation techniques for nuclear fuel reprocessing and radioactive waste treatment Edited by Kenneth L. Nash and Gregg J. Lumetta Bioalcohol production: Biochemical conversion of lignocellulosic biomass Edited by Keith W. Waldron Understanding and mitigating ageing in nuclear power plants: Materials and operational aspects of plant life management (PLiM) Edited by Philip G. Tipping Advanced power plant materials, design and technology Edited by Dermot Roddy Stand-alone and hybrid wind energy systems: Technology, energy storage and applications Edited by John K. Kaldellis Biodiesel science and technology: From soil to oil Jan C. J. Bart, Natale Palmeri and Stefano Cavallaro Developments and innovation in carbon dioxide (CO2) capture and storage technology Volume 1: Carbon dioxide (CO2) capture, transport and industrial applications Edited by M. Mercedes Maroto-Valer Geological repository systems for safe disposal of spent nuclear fuels and radioactive waste Edited by Joonhong Ahn and Michael J. Apted Wind energy systems: Optimising design and construction for safe and reliable operation Edited by John D. Sørensen and Jens N. Sørensen Solid oxide fuel cell technology: Principles, performance and operations Kevin Huang and John Bannister Goodenough Handbook of advanced radioactive waste conditioning technologies Edited by Michael I. Ojovan Membranes for clean and renewable power applications Edited by Annarosa Gugliuzza and Angelo Basile Materials for energy efficiency and thermal comfort in buildings Edited by Matthew R. Hall Handbook of biofuels production: Processes and technologies Edited by Rafael Luque, Juan Campelo and James Clark

xiv


16

Developments and innovation in carbon dioxide (CO2) capture and storage technology Volume 2: Carbon dioxide (CO2) storage and utilisation Edited by M. Mercedes Maroto-Valer Oxy-fuel combustion for power generation and carbon dioxide (CO2) capture Edited by Ligang Zheng Small and micro combined heat and power (CHP) systems: Advanced design, performance, materials and applications Edited by Robert Beith Advances in clean hydrocarbon fuel processing: Science and technology Edited by M. Rashid Khan Modern gas turbine systems: High efficiency, low emission, fuel flexible power generation Edited by Peter Jansohn Concentrating solar power technology: Principles, developments and applications Edited by Keith Lovegrove and Wes Stein Nuclear corrosion science and engineering Edited by Damien Féron Power plant life management and performance improvement Edited by John E. Oakey Electrical drives for direct drive renewable energy systems Edited by Markus Mueller and Henk Polinder Advanced membrane science and technology for sustainable energy and environmental applications Edited by Angelo Basile and Suzana Pereira Nunes Irradiation embrittlement of reactor pressure vessels (RPVs) in nuclear power plants Edited by Naoki Soneda High temperature superconductors (HTS) for energy applications Edited by Ziad Melhem Infrastructure and methodologies for the justification of nuclear power programmes Edited by Agustín Alonso Waste to energy conversion technology Edited by Naomi B. Klinghoffer and Marco J. Castaldi Polymer electrolyte membrane and direct methanol fuel cell technology Volume 1: Fundamentals and performance of low temperature fuel cells Edited by Christoph Hartnig and Christina Roth Polymer electrolyte membrane and direct methanol fuel cell technology Volume 2: In situ characterization techniques for low temperature fuel cells Edited by Christoph Hartnig and Christina Roth Combined cycle systems for near-zero emission power generation Edited by Ashok D. Rao Modern earth buildings: Materials, engineering, construction and applications Edited by Matthew R. Hall, Rick Lindsay and Meror Krayenhoff Metropolitan sustainability: Understanding and improving the urban environment Edited by Frank Zeman Functional materials for sustainable energy applications Edited by John A. Kilner, Stephen J. Skinner, Stuart J. C. Irvine and Peter P. Edwards Nuclear decommissioning: Planning, execution and international experience Edited by Michele Laraia

17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36


37 38 39 40 41 42 43 44 45 46 47 48 49

50 51 52 53 54 55

xv

Nuclear fuel cycle science and engineering Edited by Ian Crossland Electricity transmission, distribution and storage systems Edited by Ziad Melhem Advances in biodiesel production: Processes and technologies Edited by Rafael Luque and Juan A. Melero Biomass combustion science, technology and engineering Edited by Lasse Rosendahl Ultra-supercritical coal power plants: Materials, technologies and optimisation Edited by Dongke Zhang Radionuclide behaviour in the natural environment: Science, implications and lessons for the nuclear industry Edited by Christophe Poinssot and Horst Geckeis Calcium and chemical looping technology for power generation and carbon dioxide (CO2) capture: Solid oxygen- and CO2-carriers Paul Fennell and E. J. Anthony Materials’ ageing and degradation in light water reactors: Mechanisms, and management Edited by K. L. Murty Structural alloys for power plants: Operational challenges and high-temperature materials Edited by Amir Shirzadi and Susan Jackson Biolubricants: Science and technology Jan C. J. Bart, Emanuele Gucciardi and Stefano Cavallaro Advances in wind turbine blade design and materials Edited by Povl Brøndsted and Rogier P. L. Nijssen Radioactive waste management and contaminated site clean-up: Processes, technologies and international experience Edited by William E. Lee, Michael I. Ojovan, Carol M. Jantzen Probabilistic safety assessment for optimum nuclear power plant life management (PLiM): Theory and application of reliability analysis methods for major power plant components Gennadij V. Arkadov, Alexander F. Getman and Andrei N. Rodionov The coal handbook: Towards cleaner production Volume 1: Coal production Edited by Dave Osborne The coal handbook: Towards cleaner production Volume 2: Coal utilisation Edited by Dave Osborne The biogas handbook: Science, production and applications Edited by Arthur Wellinger, Jerry Murphy and David Baxter Advances in biorefineries: Biomass and waste supply chain exploitation Edited by Keith Waldron Geological storage of carbon dioxide (CO2): Geoscience, technologies, environmental aspects and legal frameworks Edited by Jon Gluyas and Simon Mathias Handbook of membrane reactors Volume 1: Fundamental materials science, design and optimisation Edited by Angelo Basile

xvi


56

Handbook of membrane reactors Volume 2: Reactor types and industrial applications Edited by Angelo Basile Alternative fuels and advanced vehicle technologies for improved environmental performance: Towards zero carbon transportation Edited by Richard Folkson Handbook of microalgal bioprocess engineering Christopher Lan and Bei Wang Fluidized bed technologies for near-zero emission combustion and gasification Edited by Fabrizio Scala Managing nuclear projects: A comprehensive management resource Edited by Jas Devgun Handbook of Process Integration (PI): Minimisation of energy and water use, waste and emissions Edited by Jir í J. Klemes Coal power plant materials and life assessment Edited by Ahmed Shibli Advances in hydrogen production, storage and distribution Edited by Ahmed Basile and Adolfo Iulianelli Handbook of small modular nuclear reactors Edited by Mario D. Carelli and Dan T. Ingersoll Superconductors in the power grid: Materials and applications Edited by Christopher Rey Advances in thermal energy storage systems: Methods and applications Edited by Luisa F. Cabeza Advances in batteries for medium and large-scale energy storage Edited by Chris Menictas, Maria Skyllas-Kazacos and Tuti Mariana Lim Palladium membrane technology for hydrogen production, carbon capture and other applications Edited by Aggelos Doukelis, Kyriakos Panopoulos, Antonios Koumanakos and Emmanouil Kakaras Gasification for synthetic fuel production: Fundamentals, processes and applications Edited by Rafael Luque and James G. Speight Renewable heating and cooling: Technologies and applications Edited by Gerhard Stryi-Hipp Environmental remediation and restoration of contaminated nuclear and NORM sites Edited by Leo van Velzen Eco-friendly innovation in electricity networks Edited by Jean-Luc Bessede The 2011 Fukushima nuclear power plant accident: How and why it happened Yotaro Hatamura, Seiji Abe, Masao Fuchigami and Naoto Kasahara. Translated by Kenji Iino Lignocellulose biorefinery engineering: Principles and applications Hongzhang Chen Advances in membrane technologies for water treatment: Materials, processes and applications Edited by Angelo Basile, Alfredo Cassano and Navin Rastogi

57 58 59 60 61 62 63 64 65 66 67 68

69 70 71 72 73 74 75


76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96

xvii

Membrane reactors for energy applications and basic chemical production Edited by Angelo Basile, Luisa Di Paola, Faisal Hai and Vincenzo Piemonte Pervaporation, vapour permeation and membrane distillation: Principles and applications Edited by Angelo Basile, Alberto Figoli and Mohamed Khayet Safe and secure transport and storage of radioactive materials Edited by Ken Sorenson Reprocessing and recycling of spent nuclear fuel Edited by Robin Taylor Advances in battery technologies for electric vehicles Edited by Bruno Scrosati, J€urgen Garche and Werner Tillmetz Rechargeable lithium batteries: From fundamentals to applications Edited by Alejandro A. Franco Calcium and chemical looping technology for power generation and carbon dioxide (CO2) capture Edited by Paul Fennell and Ben Anthony Compendium of hydrogen energy Volume 1: Hydrogen production and purificiation Edited by Velu Subramani, Angelo Basile and T. Nejat Veziroglu Compendium of hydrogen energy Volume 2: Hydrogen storage, transmission, transportation and infrastructure Edited by Ram Gupta, Angelo Basile and T. Nejat Veziroglu Compendium of hydrogen energy Volume 3: Hydrogen energy conversion Edited by Frano Barbir, Angelo Basile and T. Nejat Veziroglu Compendium of hydrogen energy Volume 4: Hydrogen use, safety and the hydrogen economy Edited by Michael Ball, Angelo Basile and T. Nejat Veziroglu Advanced district heating and cooling (DHC) systems Edited by Robin Wiltshire Microbial electrochemical and fuel cells: Fundamentals and applications Edited by Keith Scott and Eileen Hao Yu Renewable heating and cooling: Technologies and applications Edited by Gerhard Stryi-Hipp Small modular reactors: Nuclear power fad or future? Edited by Daniel T. Ingersoll Fuel flexible energy generation: Solid, liquid and gaseous fuels Edited by John Oakey Offshore wind farms: Technologies, design and operation Edited by Chong Ng and Li Ran Uranium for nuclear power: Resources, mining and transformation to fuel Edited by Ian Hore-Lacy Biomass supply chains for bioenergy and biorefining Edited by Jens Bo Holm-Nielsen and Ehiaze Augustine Ehimen Sustainable energy from salinity gradients Edited by Andrea Cipollina and Giorgio Micale Membrane technologies for biorefining Edited by Alberto Figoli, Alfredo Cassano and Angelo Basile

This page intentionally left blank

Acknowledgments

Dr Chong Ng would like to expresses his sincere gratitude to the Offshore Renewable Energy (ORE) Catapult, Mr. Ignacio Marti for allowing him the time and resources to work on this book. He is extremely grateful for the support given by Mr. David Southern, Mr. Hyunjoo Lee, Mr. Jonathan Hughes, Dr. Kirsten Dyer, and Mr. Tony Fong with their specialist knowledge in the reviewing process.


Part One Introduction to offshore wind energy and offshore wind farm siting


Introduction to offshore wind energy

1

C. Ng Offshore Renewable Energy Catapult, Northumberland, United Kingdom L. Ran University of Warwick, Coventry, United Kingdom

1.1

Wind energy

The history of utilising wind energy in the form of windmills started thousands of years ago. In modern terms, wind power generation is a process of utilising wind energy to generate electricity. Wind turbines are used as the mechanism to convert the kinetic energy in the wind into mechanical work and then into electricity through a generator. Since the first offshore wind farm was deployed in Denmark in the early 1990s, utilising stronger and steadier wind energy offshore to generate electricity has always been part of the wind industry development agenda. With the confidence and technical competency that has been accumulated through the experience of onshore wind development, it can be seen that the offshore wind industry started to grow significantly in the middle of the 2000s, doubling the total capacity every 2e4 years. Fig. 1.1 shows an analysis presented by the European Wind Energy Association (EWEA) on offshore wind installations in Europe since 1993 [1]. Based on the Global Wind 2014 statistics [2], more than 90% of all offshore wind installations were in European waters, spread across the North Sea (63.3%), the Atlantic Ocean (22.5%) and the Baltic Sea (14.2%). The United Kingdom accounts for over half of the total European offshore wind capacity installed to date, with an accumulative capacity of 4494 MW. Outside Europe, countries have set aggressive plans to promote their wind industry and offshore wind has become a new focus. China in particular installed close to 230 MW of offshore wind in 2014 alone, which makes it the third largest annual market globally after the United Kingdom and Germany.

1.2

Offshore wind farm

Offshore wind power development, in simple terms, can be split into two levels: wind farm level and wind turbine level. At the wind farm level, the power generated by individual wind turbines is collected through an inter-array connection to an offshore substation, or sometimes to more than one substations. The electricity generated is transmitted to the shore in either alternative current (AC) or direct current (DC)

Offshore Wind Farms. http://dx.doi.org/10.1016/B978-0-08-100779-2.00001-5 Copyright © 2016 Elsevier Ltd. All rights reserved.

9000

1600

8000

1400

7000

1200

6000

1000

5000

800

4000

600

3000

400

2000

200

1000

Cumulative (MW)

Annual (MW)

4

1800

-

Figure 1.1 Cumulative and annual offshore wind installations (MW).

Offshore Wind Farms

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2.0 4.0 50.50 170.0 276.20 89.70 90.0 92.50 318.40 373.48 576.90 882.70 873.55 1165.5 1567.0 1483.3 2.80 5.0 16.80 Cumulative 4.950 6.950 11.950 28.750 28.750 31.550 31.550 35.550 86.050 256.05 532.25 621.95 711.95 804.45 1122.8 1496.3 2073.2 2955.9 3829.4 4994.9 6561.9 8045.2 Annual


5

form through a subsea transmission scheme, which sometimes consists of multiple links to increase the availability and security. Details on offshore power collection, onwards transmission, cabling and planning and operation of the onshore power systems with offshore wind are covered in Part III of this book, consisting of five self-contained chapters: • • • • •

Chapter 12 Offshore wind farm arrays Chapter 13 Cabling to connect offshore wind turbines to onshore facilities Chapter 14 Integration of power from offshore wind turbines into onshore power grids Chapter 15 Energy storage for offshore wind farms Chapter 16 Hydropower flexibility and transmission expansion to support integration of offshore wind

The earlier offshore wind farms were typically within 10 km of the shore, with a water depth of less than 20 m. However, as the availability of such sites has been exhausted, new offshore wind farms have moved to further and deeper locations. For example, the ‘Dogger Bank’ wind farm, one of the world’s largest wind farms under development in the United Kingdom, is located more than 100 km from the shore with what is currently the longest edge-to-edge distance of 260 km. Generally, developing larger wind farms further offshore could allow a higher rate of energy harvesting and hence better financial returns. Increases in farm size and distance to shore are both inevitable in future offshore wind farm developments. To reduce the dependency on shallow water sites and to explore high wind resources in further offshore and deeper water regions, floating wind turbines have been proposed and have achieved good development breakthrough in the past few years. As described in detail in Chapter 11 ‘Design of floating offshore wind turbines’, despite the leaps in recent development, there are challenges on issues such as load stress reductions, design margin calculation, operational stability and so on, and these are yet to be addressed to realise the practicality of floating wind turbines.

1.3

Energy cost

It is well known that the success of the entire renewable power generation industry is highly dependent on the levelised cost of energy (LCoE). In 2015, the Offshore Renewable Energy (ORE) Catapult presented an assessment result, Fig. 1.2, showing an LCoE reduction for offshore wind, from £136/MWh for the 2010e2011 completion projects to £131/MWh for the 2012e2014 completion projects. The LCoE was projected to be £121/MWh for 2012e2014 final investment decision projects [3]. Efficiencies, including design efficiency, system efficiency and operational efficiency, together with system availability that is dependent on the subsystem reliability, are the key elements in the offshore wind energy cost battle and are the main focus of this book. Amongst other parameters, the annual availability and OPEX (operation and maintenance expenditure) of the power plants would be

6

Offshore Wind Farms

160 Levelised cost of energy (£/MWh, 2011 prices)

140

136

131 121

120 (2020 Target)

100 80 60 40 20 0 Works completion 2010–2011

Works completion 2012–2014

FID 2012–2014

Figure 1.2 Quantitative LCoE assessment summary [3].

one of the easiest measurable performance indicators that will influence energy cost. The offshore wind industry, in addition to its high capital expenditure (CAPAX), also suffers from higher OPEX as compared to its onshore counterpart. Lifetime OPEX of offshore wind is close to 90% of its CAPAX. Unlike the large mechanical drive-train components, such as gearboxes and bearings, the design philosophy adopted by the power electronics system industry, that is, power converters, power conditioners, etc., to overcome the fragile nature and maintain high system availability, is usually to modularise the system with easily swappable subsystems. This concept has proven to be effective in many of the onshore projects as on-site repairs can be performed without the need for major system replacement. However, when it comes to offshore, due to the extremely costly offshore logistics and highly weather-dependent vessel scheduling, any intermittent failures of power electronic systems that require manual reset or component replacement would have a significant impact, possibly as great an impact as failures of mechanical components, to the plant O&M in terms of OPEX. In recent years, there have been a number of UK- and European-funded research projects looking into system robustness improvement, health condition monitoring and lifetime prognosis methodologies, to improve overall wind turbine availability. Deployment further offshore in deeper waters would potentially further increase the OPEX. How condition monitoring techniques can help on these issues is addressed in Chapter 18 ‘Condition monitoring of offshore wind turbines’, in Part IV of the book, which includes three other chapters to discuss other aspects about the installation and operation of offshore wind farms: • • •

Chapter 17 Assembly, transportation, installation and commissioning of offshore wind farms Chapter 19 Health and safety of offshore wind farms Chapter 20 Offshore wind-turbine foundations: analysis and design


7

These are complemented by three chapters in Part I of the book which addresses the resource and siting criteria for offshore wind: • • •

Chapter 2 Economics of building and operating offshore wind farms Chapter 3 Wind resources for offshore wind farms: characteristics and assessment Chapter 4 Remote sensing technologies for measuring offshore wind

1.4

Wind turbines

Upwind horizontal axis with geared high-speed doubly fed induction generators (DFIG), medium-speed geared permanent magnet synchronous generators (PMSGs) and low-speed direct-drive PMSGs are the three main wind turbine configurations utilised by the offshore wind industry to date. The vast majority of the offshore wind turbines rated below 4 MW deployed so far are still of the DFIG configuration. There is a trend of moving towards larger, hybrid, medium-speed configurations and there are also major players who have announced during the past 2 years their development and intention to deploy large direct-drive PMSG-type wind turbines in the near future. In the wind industry, as analysed in detail in Chapter 2, scale is important to achieve the economic benefits. There is a constant increase in wind farm size and, at the same time, larger wind turbines are being promoted and deployed offshore to improve the return of investments. Larger wind turbines, however, have introduced technical challenges in the substructures or subcomponents, such as rotor blades, towers and potentially the foundation designs. The rotor blade, and the building materials, as some of the heaviest components in the existing wind turbines are discussed in Chapter 5 ‘Developments in materials for offshore wind turbine blades’ and Chapter 6 ‘Design of offshore wind turbine blades’, respectively. These chapters in Part II of the book also address the issues of how to improve the mass density and reliability, and discuss designs to accommodate larger wind turbines for the future offshore wind industry. Increasing turbine size and head mass will have a direct impact on the tower as well as the foundation. This, together with the harsh offshore environment, wind and wave loadings, as discussed in Chapter 10 ‘Design of offshore wind turbine towers’ and Chapter 19 ‘Health and Safety of Offshore Wind Farms’, would create new design challenges to the offshore wind turbine developers. Turbine technologies are mostly covered in Part II of this book. In addition to Chapters 5, 6, 10 and 11, Part II also includes the following three chapters on other key components: • • •

Chapter 7 Wind turbine gearbox design with drive-train dynamic analysis Chapter 8 Design of generators for offshore wind turbines Chapter 9 Modelling of power electronic components for evaluation of efficiency and power density of offshore wind power converters

8

1.5

Offshore Wind Farms

Disputable issues

One of the challenges to develop larger wind farms further offshore in the future is to effectively collect the generated electricity from individual turbines and transmit the power back to shore. To collect electricity in a large wind farm, at a scale even larger than Dogger Bank (1.2 GW), with widespread wind turbines would require multiple secondary collection platforms to reduce the cable length. In the past 1e2 years, operating interarray at a higher collection voltage, for example, 66-kV AC to replace the current 33-kV AC voltage level for offshore power collection has been at the centre of discussions [4]. Carbon Trust-funded research in the United Kingdom suggested that 1.5% of the cost of energy can be saved by simply moving the interarray voltage level from 33 to 66 kV. Multiterminal HVDC, on the other side, has been suggested by a number of developers and manufacturers as a longer-term solution for the offshore wind industry. Detailed discussion of the technologies and features can be found in Chapter 12 ‘Offshore wind farm arrays’. Given the uncertainties involved in the future development of the offshore wind industry, many aspects of current practice in terms of planning, design, deployment and operation could be open for argument. Most of the chapters in this book include discussions on the factors that affect the current practice and that may vary in the future.

References [1] European Wind Energy Association (EWEA), The European Offshore Wind Industry e Key Trends and Statistics 2014, January 2015. [2] Global Wind Energy Council (GWEC), Global Wind Report e Annual Market Update 2014, March 2015. [3] O.R.E. Catapult, Cost Reduction Monitoring Framework- Summary Report to the Offshore Wind Programme Board, February 2015. [4] A. Ferguson, P.D. Villiers, B. Fitzgerald, J. Matthiesen, Benefits in moving the inter-array voltage from 33 kV to 66 kV AC for large offshore wind farms, in: European Wind Energy Conference (EWEC), April 2012. Copenhagen.

Economics of building and operating offshore wind farms

2

P.E. Morthorst, L. Kitzing Technical University of Denmark, Roskilde, Denmark

2.1 2.1.1

Introduction Expectations to offshore wind power

Expectations of the development of offshore wind power are high, especially in Europe. According to the National Renewable Energy Allocation Plans (NREAPs) prepared by European Union (EU) member states, a total offshore wind capacity of 43 GW is expected to be implemented in the EU by 2020 (Green and Vasilakos, 2011). However, in reality the development thus far has not been that fast. By the end of 2014 a total offshore capacity of 8759 MW was installed worldwide, the vast majority in Europe (8045 MW or 91%) (GWEC, 2015) e still a long way to go if the NREAP target for 2020 is to be reached. Faced with this fact the European Wind Energy Association (EWEA) has recently downgraded its 2020 expectations from being in line with the NREAP target to a significantly lower level, ranging between 19.5 and 27.8 GW by 2020. To reach these levels a growth rate of between 20% and 27% p.a. is required, much in line with the realised growth rates of recent years, 33% in 2012, 31% in 2013 and 23% in 2014.

2.1.2

Development of offshore wind power

In a number of countries, offshore turbines are taking on an increasingly important role in the development of wind power, particularly in the north-western part of Europe. Without a doubt, the main reasons are that on-land sitings are limited in number and the utilisation of these sites, to a certain extent, is exposed to opposition from the local population. This, seen in relation to a significantly higher level of energy production from offshore turbines compared to on-land sitings, has paved the way for strong interest in offshore development. As for onshore turbines, the wind regime, where the offshore turbines are sited determining the production of power, is the single most important factor for the cost per generated unit of electricity. In general, the wind regime offshore is characterised by higher average wind speeds and more stability than onshore wind. At the Danish Horns Reef wind farm, a wind speed corresponding to a utilisation time of more than 4200 h per year was measured (adjusted to a normal wind year), thus giving a capacity factor close to 50%, which is comparable to many relatively small conventional power plants. For most offshore wind farms, a utilisation time of more than Offshore Wind Farms. http://dx.doi.org/10.1016/B978-0-08-100779-2.00002-7 Copyright © 2016 Elsevier Ltd. All rights reserved.

10

Offshore Wind Farms

3000 h per year is to be expected, significantly higher than that for on-land sited turbines and, therefore, to a certain extent compensating for the additional costs of offshore plants. Offshore development is dominated by a handful of countries, most located in Western Europe. In Fig. 2.1, the cumulated offshore capacity and the growth in 2014 clearly indicate the countries driving the development: the UK installed 47% of the new capacity in 2014, followed by Germany (31%), China (13%) and Belgium (8%). The UK seems set to maintain its leading position in the years to come as quite a number of new offshore wind farm installations are either under construction or in the planning process. Although onshore wind power is being developed fast in China, offshore development has not reached the same pace, despite that a strong potential for offshore wind does exist there. With 230 MW installed in 2014, total offshore wind capacity in China has reached 658 MW (GWEC, 2015). China has an offshore target of 5 GW by 2015, and although a number of new offshore projects are on the way it seems most unlikely that this target will be met.

2.2 2.2.1

Investment costs Development in investment costs

Offshore wind farms are capital-intensive. Upfront investment costs make up approximately 75% of the total lifetime cost of an offshore wind farm, which is extremely high in comparison to other electricity generation technologies e typically investment costs constitute around 40% of the cost of energy from a conventional power plant. From the investment per MW point of view, offshore wind is still some 50% more expensive than onshore wind. The higher capital costs of offshore are due to the larger structures and complex logistics of installing the towers. The costs of offshore foundations, construction, installations, and grid connection are significantly higher than for onshore. Typically, offshore turbines are 20% more expensive, and towers and foundations cost more than 2.5 times the price for a project of similar size onshore. As shown in Fig. 2.21 the specific investment costs of offshore wind farms (MV/MW) have in general been increasing, despite significantly larger wind farms and thus the expected economies of scale. The main reasons for this are to be found in increasing distances to shore and water depths. Also, in the late 2000s, supply bottlenecks and increasing component prices occurred. Overall, the specific investment cost of offshore wind power has not decreased at the same pace as previously seen for onshore turbines. 1

A comprehensive data-base consisting of 45 large European offshore wind farms compiled of data from Risø DTU, KPMG and 4C is utilised for the analyses in this chapter. The data-base includes all major offshore wind farms, constituting 96% of all offshore wind power capacity by 2013. Thus it is found to be representative of the total European offshore wind power capacity.

New capacity in 2014

800 700

United Kingdom 51%

600 MW

Belgium 12%

900

2014


Ireland Sweden Finland 0% 3% Japan 0% Netherlands Others 1% 0% China 8% 2% Germany 8%

500 400 300 200 100 0

Denmark 15%

United Kingdom

Germany

China

Belgium

Figure 2.1 Offshore development. Left: Distribution of total installed capacity by the end of 2014. Right: New installed capacity in 2014. GWEC, 2015. Global Wind Report. Annual Market Update 2014. Global Wind Energy Council, Brussels, Belgium.

11

Specific investment cost (m€/MW) 2

1

0 2000 2002 2004 2006 2008 2010 2012 2014 Nordsee Ost (2014)

Northwind (2014)

Anholt (2013)

Thornton Bank II (2013)

London Array 1 (2012)

Walney 2 (2012)

Baltic I (2011)

Rødsand II (2010)

Alpha Ventus (2010)

Belwind 1 (2009)

Lillgrunden (2008)

Burbo bank (2007)

Princess Amalia (2006)

Kentich flat (2005)

Scroby sands (2004)

Samsø (2003)

Horns rev I (2002)

3

Middelgrunden (2000)

5

Yttre Stengrund (2001)

4 M€/MW

12

6

5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

Figure 2.2 Specific investment costs of European offshore wind farms, MV/MW. Left: All wind farms in data-base. Right: Selected wind farms. Offshore Wind Farms


13

5 4.5 4

m€/MW

3.5 3 2.5 2 1.5 1 0.5 0 Be lg iu m

an

(4 )

s

)

6) y(

nd

)

(7

(2

k

rla

1)

he

(2

m er

et

G

N

K

ar

en

m

ed

en

U

D

Sw

(3 )

Figure 2.3 Average specific investment cost categorised by country. In parentheses is shown the number of wind farms represented in the data-base.

Worth noticing also is that some countries are in general found to have significantly lower investment costs than others, see Fig. 2.3. Belgium is found to be the country with the highest specific investment costs, followed by Germany and the Netherlands. Sweden2 is the country with the lowest specific investment costs in the past, followed by Denmark. Some divergence between countries can be explained by differences in the age of wind farms, in the size of turbines, in water depth and distance to shore.

2.2.2

Investment costs split into cost components

The investment cost of an offshore wind farm is typically split into a limited number of cost components, including the turbine itself, foundation, cabling etc. Table 2.1 shows a way to categorise these costs and explains what is included. Turbines are the most costly element of an offshore wind project, accounting for between 40% and 60% of the total investment cost. Turbine blades and towers make approximately half the total turbine costs (Ports, 2014). Installation is typically the largest cost item after the turbines themselves, with roughly a quarter of the total investment cost. Installation costs are however often not visible in cost breakdowns, as they often are shown as an integrated part of turbine, foundation and cable costs, respectively. The third-largest cost component, foundations, accounts for approximately 20% of the investment costs. Overall, approximately one third of the total investment cost is labour costs, another third is materials and the last third comprises services, insurance and other overheads (RAB, 2010). Key commodities in the material cost are fibreglass, steel, iron and copper, which make upto 90% of the material cost of a turbine (RAB, 2010). 2

Sweden is represented with two wind farms in the data-base, one of which is very small. Thus, the Swedish data may not be representative.

14

Table 2.1

Offshore Wind Farms

Investment costs split into a number of cost components

Cost components

Examples for detailed cost items

Development and project management

Design, management, consenting, consultancy fees

Turbines

Tower, rotor blades, rotor hub, rotor bearings, main shaft, main frame, gear box, generator, yaw system, pitch system, power converter, transformer, brake system, nacelle housing, cables

Foundations

Foundations, transition piece

Electrical installations (offshore) (including balance of plant)

Collection system, integration system, offshore substation. Transmission system, reactive power compensation system, electrical devices, export cable (main cable to coast)

Grid connection (onshore)

Dedicated cables onshore, isolators, switchgear under control of onshore network operator

Installation (of turbines, foundations, cables and electrical equipment)

Transportation cost, vessel chartering costs, labour cost

Financial costs

Financing, bank fees, securities

Miscellaneous

Services, insurance and other overheads

The split into cost components will of course vary considerably between wind farm projects. However, to illustrate more thoroughly the shares of these cost components in total costs, two examples are given: (1) an average of the two Danish offshore wind farms, Horns Reef I and Rødsand I and (2) the Swedish offshore wind farm, Lillgrunden. Horns Reef I, completed in 2002, is located approximately 18 km off the west coast of Jutland (west of Esbjerg). It is equipped with 80 2 MW machines, a total capacity of 160 MW. Rødsand I, completed in 2003, is located south of the isle of Lolland e 11 km from the shore e and consists of 72 2.3 MW turbines, a total capacity of 165 MW. Both wind farms have their own transformer station located at the sites, which through transmission cables are connected to the high-voltage grid at the coast. The farms are operated from onshore control stations and no staff is required at the offshore sites. Lillgrunden is located south of the Øresund-bridge connecting Copenhagen and Malm€ o, approximately 8 km off the Swedish coast. Lillgrunden is equipped with 48 2.3 MW turbines, in total 110 MW. The average investment costs related to the above-mentioned wind farms, split into main components, are shown in Table 2.2. The total cost of each of the two Danish offshore farms is close to 260 million V, while the Swedish cost around 215 million V.

16

Offshore Wind Farms

2.3

Operating costs

Next to investment cost, operation and maintenance (O&M) costs constitute a sizeable share of the total costs of an offshore wind turbine. Thus, O&M costs may easily make up 25e30% of the total levelised cost per kWh produced over the lifetime of the turbine. If the turbine is fairly new, the share may only be 20e25%, but this might increase to at least 30e35% by the end of a turbine’s lifetime. As a result, O&M costs are attracting greater attention, as manufacturers attempt to lower these costs significantly by developing new turbine designs that require fewer regular service visits and less turbine downtime, which is especially important for offshore turbines. Condition monitoring is also being developed as a technology that can be added to existing and new designs. Offshore O&M costs are related to a limited number of cost components, including: • • • • • •

Insurance; Regular maintenance; Repair; Spare parts; Access to platform and turbines; and Administration.

For offshore wind farms, part of these O&M costs is rather difficult to estimate. Due to weather-related access restrictions and longer distances to offshore turbines, labour-related costs tend to be significantly higher and less predictable than for onshore plants. Whereas standard long-term contracts for insurance and regular maintenance can be obtained for onshore turbines, this is not necessarily the case for offshore turbines. Finally, costs for repair and related spare parts are always difficult to predict. Although all cost components tend to increase as the turbine gets older, costs for repair and spare parts are particularly influenced by turbine age, starting low and increasing over time. In the past, lifetime average O&M cost have been estimated spanning a broad range from 15 to 49 V/MWh. Table 2.3 shows several different O&M estimations from the

Estimates of O&M costs for offshore turbines (all prices converted to Euros, real 2012)

Table 2.3

O&M cost (V/MWh) Operating farms, EU, 2002e2009 (Morthorst et al., 2009)

18

Danish technology data catalogue, 2015 (DEA, 2014)

19

German projects, 2010 (KPMG, 2010)

27

Offshore farms in Europe (IRENA, 2012)

25e49


15

Average investment costs per MW related to offshore wind farms at Horns Rev, Nysted and Lillgrunden (split into main components)

Table 2.2

Horns Reef I and Rødsand I

Lillgrunden

Investments 1000 V/MW

Share %

Investments 1000 V/MW

Share %

Turbines ex work, including transport and erection

872

49

1074

57

Transformer station and main cable to coast

289

16

244

13

Internal grid between turbines

91

5

e

e

Foundations

375

21

361

19

Design, project management

107

6

60

3

Environmental analysis etc.

54

3

e

e

Other contractors

e

e

80

4

Miscellaneous

11

20 m

3.8

Water depth Average for: 0e10 m

2.2

11e20 m

3.2

>20 m

3.7

Utilising the data-base for simple partial calculations, Table 2.4 is established, indicating quite strong increases in cost due to further distance to shore and/or increased water depth. Note that due to the partial nature of these calculations, several other parameters influence the results and the big differences cannot be related to the distance to shore and water depth only. For the same reason, they cannot be directly compared to the results of the statistical analyses mentioned above.

2.5

Levelised cost of energy

The total cost per kWh produced (unit cost) is calculated by discounting and levelising investment and O&M costs over the lifetime of the turbine, and then dividing them by the annual electricity production; this is called the levelised cost of energy (LCOE). Thus, LCOE is calculated as an average cost over the turbine’s lifetime. In reality, actual costs will be lower than the calculated average at the beginning of the turbine’s life, due to low O&M costs, and will increase over the period of turbine use. The turbine’s power production is the single most important factor for the cost per unit of


23

power generated. The profitability of a turbine depends largely on whether it is sited at a good wind location. The costs for offshore wind farms are considerably higher than for onshore turbines. However, this is to a certain degree moderated by higher total electricity production (capacity factor) from the turbines due to higher offshore wind speeds. For an on-land installation utilisation time is normally around 2000e2300 h per year, while a typical offshore installation has a utilisation time of 3000 h per year or above. As an example the calculated costs per kWh of electricity generated as a function of the wind regime at the chosen sites are shown in Fig. 2.85 where the turbine’s investment cost is treated on a sensitivity basis (left), together with a sensitivity analysis on the used discount rate (right). The sensitivity analysis covers investment costs for offshore wind farms ranging from 3200 US$/kW to 5000 US$/kW using a discount rate of 7% p.a. (left figure). The sensitivity analysis on the discount rate ranges from 3% p.a. to 10% p.a. using the average investment cost of 3900 US$/kW (right figure). As shown, LCOE changes considerably depending on the capacity factor and thus how windy the chosen site is. For an offshore standard installation with an investment cost of 3900 US$/kW the cost ranges from approximately 15 UScent/kWh (13.5 cV/kWh) at sites with average offshore wind speeds (capacity factor of 35%) to approximately 11e12 UScent/kWh (10e11 cV/kWh) at excellent offshore sites (capacity factor of 50%). The Danish wind farm Horns Reef I has a capacity factor of 50%. The sensitivity analysis on investment cost shows that investments ranging from 3200 to 5000 US$/kW imply that LCOE ranges from approximately 13 UScent/kWh (11.5 cV/kWh) to approximately 18 UScent/kWh (16 cV/ kWh). The level of the discount rate also has a significant influence; increasing the discount rate from 3% to 10% p.a. increases the LCOE from approximately 12 UScent/kWh (10.5 cV/kWh) to almost 18 UScent/kWh (16 cV/kWh) with a capacity factor of 35%. As shown in Fig. 2.8 the cost of energy from offshore turbines exceeds the cost for onshore turbines considerably. At good coastal positions, onshore turbines can produce at approximately 6e7 UScent/kWh (5e6 cV/kWh). In Europe coastal positions such as these are mostly to be found on the coasts of the UK, Ireland, France, Denmark and Norway. The cost of balancing the power production from the turbines is not included in the above-mentioned estimates, normally these costs are borne by the farm owners. According to previous Danish experience, balancing requires an equivalent cost of approximately 3.5 UScent/kWh (3 cV/MWh). Also, balancing costs are subject to high uncertainty and might differ substantially between countries. The above-mentioned costs are calculated as simple national economic ones, thus these costs will not be those of a private investor, which will have higher financial 5

In Figure 2.8 the capacity factor is used to represent the wind regime. The capacity factor is defined as the number of full load hours per year divided by the total number of hours per year (8760). Full load hours are calculated as the turbine’s average annual production divided by its rated power. The higher the capacity factor (correspondingly the number of full load hours), the higher the wind turbine’s production at the chosen site.

24

35

35

Offshore discount rate = 10%

Offshore USD 3900/kW

30

Levelized cost of energy (UScent2005/kWh)

Levelized cost of energy (UScent2005/kWh)

Offshore USD 5000/kW

Offshore USD 3200/kW Onshore USD 2100/kW

25

Onshore USD 1750/kW Onshore USD 1200/kW

20

15 Europe offshore projects

10

5

Offshore discount rate = 7%

30

Offshore discount rate = 3% Onshore discount rate = 10%

25

Onshore discount rate = 7% Onshore discount rate = 3%

20

15

10

5

China European Low-Medium Wind Areas

0 15

US Great Plains

0 20

25

30

35

Capacity factor (%)

40

45

50

15

20

25

30

35

40

45

50

Capacity factor (%) Offshore Wind Farms

Figure 2.8 Estimated levelised cost of on- and offshore wind energy, 2009: (left) as a function of capacity factor and investment cost and (right) as a function of capacity factor and discount rate. Wiser, R., Yang, Z., Hand, M., Hohmeyer, O., Infield, D., Jensen, P.H., Nikolaev, V., O’Malley, M., Sinden, G., Zervos, A., 2012. Wind energy. In: Edenhofer et al. (Eds.), Renewable Energy Sources and Climate Change Mitigation, p. 588 (Chapter 7, Figure 7.23).


25

costs, and require a risk premium and a profit. How much a private investor will add on top of the simple costs will, among other things, depend on the perceived technological and political risk of establishing the offshore wind farm and on the competition between manufacturers and developers.

2.6

Future cost of offshore wind

While future costs are hard to predict and previous expectations about cost reductions from the late 1990s and early 2000s have given way to actual cost increases (UKERC, 2010), the offshore wind industry is still optimistic, expecting significant cost reductions during the next 10 years: An industry survey (Ports, 2014) with 200 executives from the European offshore industry revealed that investment cost reductions of on average 23% are expected until 2023. In several addresses at the EWEA Offshore in 2013, it was argued that the industry is on track to achieve the targets of 40% cost reduction by 2020. A study for the German offshore wind sector (Hobohm et al., 2013) came to the conclusion that cost reductions of up to 39% of the levelised costs can be achieved over the next 10 years in optimum market conditions. The most important drivers for future cost reductions in the offshore wind supply chain are economies of scale, higher-capacity turbines and technology innovation (Ports, 2014). Additionally, improvements in logistic infrastructure, such as geographic concentration of the supply chain and faster ships, can play a role (Hobohm et al., 2013). All these factors will create reductions in different cost components. Looking at the investment cost components, it is expected that the three largest elements (turbines, installation, foundations) are also the ones with the highest cost reduction potentials, namelye5e7% each in the short term (Ports, 2014). Turbine costs will mostly be reduced through economies of scale, amongst other things from higher-capacity turbines. Above, we show the previous development of turbine sizes in European offshore wind farms, with the largest installed turbines of 6.15 MW (by the end of 2014). In the near future, we expect to see 8 MW and even 10 MW turbines being installed in European waters. Also, some greater competition in the offshore wind turbine sector is expected. Regarding installation costs, it is expected that a more efficient installation process alone could result in a 1.65% decrease in the total investment cost of an offshore wind farm (Ports, 2014). Furthermore, larger projects and higher-capacity turbines will contribute to the reduction of per MW installation costs. Foundation costs are expected to be reduced through economies of scale, higher-capacity turbines, technology innovation and standardization of foundation designs (Ports, 2014). Also, a holistic design optimisation can contribute to minimising overall costs. Recently developed jacket-based foundations can, for example, be installed within 12 h, rather than in up to 5 days, which are typically required for a monopole foundation (Ports, 2014). Thus, foundation design can significantly influence installation costs as well. Longer-term cost reduction options in investment costs are HVDC connections to shore and even offshore grid networks, which will benefit multiple offshore wind farms at a time.

26

Offshore Wind Farms

On the operations side, potentials for O&M cost reduction and performance optimisation are expected to come from three main drivers: reliability, maintainability and operations management (RAB, 2010). The reliability of offshore wind farms is expected to improve from advanced condition monitoring and failure avoidance. Concepts for increasing the maintainability could, for example, include a minimisation of the need for the use of jack-ups for repairs. Improvements in operations management could include condition-based maintenance and scheduling of activity. Due to the highly specific O&M cost for offshore wind farms, cost reduction potentials are difficult to quantify on a general basis.

2.7

Conclusions

Expectations for the development of offshore wind power are high, especially in Europe, where a total offshore wind capacity of 43 GW is expected to be implemented in the EU by 2020. At present the offshore development is dominated by a handful of countries, most of these located in Western Europe. However, in reality the development until now has not been that fast, mainly because considerable cost reductions are still needed. Offshore wind farms are capital-intensive. Upfront investment costs make up approximately 75% of the total lifetime cost of an offshore wind farm, which is extremely high in comparison to other electricity generation technologies e typically investment costs constitute around 40% of the cost of energy from a conventional power plant. From the investment per MW point of view, offshore wind is still some 50% more expensive than onshore wind. The higher capital costs of offshore are due to the larger structures and complex logistics of installing the towers. The costs of offshore foundations, construction, installations and grid connection are significantly higher than for onshore. However, looking into the LCOE, the higher investment costs are to a certain degree moderated by higher production for offshore wind turbines. For a standard offshore installation with a capacity factor of 35%, LCOE typically ranges from approximately 13 UScent/kWh (11.5 cV/kWh) to approximately 18 UScent/ kWh (16 cV/kWh). In comparison a standard on-land installation with a capacity factor of 25%, LCOE typically ranges from approximately 7 UScent/kWh (6 cV/kWh) to approximately 11 UScent/kWh (10 cV/kWh). So there is still quite a long way to go before offshore wind can compete economically on its own. However, significant potential seems to exist for further cost developments of offshore installations and it is important to analyse how this potential can best be exploited.

References DEA, 2014. Technology Data for Energy Plants Generation of Electricity and District Heating, Energy Storage and Energy Carrier Generation and Conversion, May 2012 e Certain Updates Made October 2013 and January 2014. Danish Energy Agency and Energinet.dk.


27

EEA, 2009. Europe’s Onshore and Offshore Wind Energy Potential, an Assessment of Environmental and Economic Constraints. Technical report, No 6/2009. ISSN: 1725-2237. European Environment Agency, EEA. Green, R., Vasilakos, N., 2011. The economics of offshore wind. Energy Policy 39, 496e502. GWEC, 2015. Global Wind Report. Annual Market Update 2014. Global Wind Energy Council, Brussels, Belgium. Hobohm, J., Krampe, L., Peter, F., Gerken, A., Heinrich, P., Richter, M., 2013. Cost Reduction Potentials of Offshore Wind Power in Germany. Short Version (Report Commissioned by the German Offshore Wind Energy Foundation, Fichtner and Prognos: Berlin and Stuttgart, Germany). IRENA, June 2012. ‘Wind Power’, Renewable Energy Technologies: Cost Analysis Series, IRENA Working Paper. In: Volume 1: Power Sector, Issue 5/5. International Renewable Energy Agency. Krohn, S., Morthorst, P.E., Awerbuch, S., 2009. The Economics of Wind Energy: A Report by the European Wind Energy Association. European Wind Energy Association EWEA, Brussels. KPMG, 2010. Offshore Wind in Europe: 2010 Market Report. KPMG AG, Advisory, Energy & Natural Resources. Morthorst, P.E., Auer, H., Garrad, A., Blanco, I., 2009. Wind Energy the Facts. Part III: The Economics of Wind Power. European Wind Energy Association, Routledge, Taylor & Francis Group. Ports, P.D., 2014. Offshore Wind Project Cost Outlook. 2014 Edition. Publisher: Clean Energy Pipeline, VB/Research Ltd, London, UK. RAB, 2010. Value Breakdown for the Offshore Wind Sector. Report Commissioned by the Renewables Advisory Board. RAB (2010) 0365. UKERC, 2010. Great Expectations: The Cost of Offshore Wind in UK Waters e Understanding the Past and Projecting the Future. A report by the Technology and Policy Assessment Function of the UK Energy Research Centre, ISBN 1903144094. Wiser, R., Yang, Z., Hand, M., Hohmeyer, O., Infield, D., Jensen, P.H., Nikolaev, V., O’Malley, M., Sinden, G., Zervos, A., 2012. Wind energy. In: Edenhofer, et al. (Eds.), Renewable Energy Sources and Climate Change Mitigation, Chapter 7, pp. 535e607.


Wind resources for offshore wind farms: characteristics and assessment

3

B.H. Bailey AWS Truepower LLC, Albany, NY, United States

3.1

Key issues in assessing wind resources

The wind is the fuel that wind turbines tap into to generate both electricity and revenue. The wind is also one of the environmental forces that offshore wind farms must endure to perform reliably over their planned lifetime. Wind resource assessment is the process of characterizing the atmospheric environment through measurement and modeling to address the many questions raised during the development, construction, and operational phases of a wind farm. These questions relate to site selection, energy production potential, turbine suitability and layout, the balance-of-plant design, site accessibility, and other project elements. Air temperature, precipitation, humidity, pressure, and other atmospheric variables are integral to wind resource assessment. They influence both the amount of power available in the wind as well as the efficiency by which wind turbines capture and covert this power. Ocean waves, currents, surface temperature and other waterrelated parameters are influencing factors too. Not only do they impose major loads on foundations and challenges to vessels, they also directly influence the nature of the overlying atmosphere. Ultimately, the study of the physical and operating design environment of wind farms must be approached in an integrated fashion; meteorological and oceanographic (metocean) factors are interactive. For example, it is the concurrence of extreme winds and extreme waves from severe storms that can define the design envelope to which wind farms must be designed. Fig. 3.1 illustrates many of the metocean factors with which offshore wind turbines must contend. The greatest challenge to offshore resource characterization is the marine environment itself. Physical measurements are logistically difficult and expensive, which explains why they are relatively sparse. To compensate, strong emphasis is placed on weather satellites and numerical weather prediction models to characterize the ocean environment for many marine activities. While they are effective for special purposes, such as navigation and commercial fishing, their value is more qualitative than quantitative for wind energy applications. This is because the layer of the atmosphere relevant to large-scale wind turbinesdextending at least 150 m above the water surfacedis not addressed by most measurements, which focus on the ocean surface and a few meters above and below it. Further, because wind turbines are attached to


30

Offshore Wind Farms

Figure 3.1 Illustration of various metocean factors on a floating offshore wind turbine. Source: National Renewable Energy Laboratory.

sea bottom-fixed or floating foundations, measurements of the water column, which are largely absent from observational networks, are essential too.

3.2

The nature of the offshore wind environment

Among the most obvious distinctions of the open ocean environment relative to land is the low surface roughness and the lack of terrain. Generally speaking, this contributes to stronger winds with greater horizontal uniformity, smaller changes of wind speed with height (ie, wind shear), and lower levels of turbulence. Fig. 3.2 depicts mean wind conditions over the world’s ocean derived from satellite imagery. A surface roughness length of approximately 0.001 m is representative of a sea surface with small waves; by comparison, most vegetated land surfaces have a roughness length of 0.03e1.0 m, depending on the vegetation type and height. Surface roughness length represents the bulk effects of surface roughness elements and has a value of approximately one-tenth the height of those elements (Sutton, 1953; Stull, 1988). IEC-specified operational conditions for offshore wind turbines assume a power law wind shear exponent of 0.14, with a lower value of 0.11 specified for extreme


31 W/m2 1400 1200 1000 850 750 650 550 450 350 270 190 110 70 30

60N 40N 20N EQ 20S 40S 60S 0

40E

80E

120E

160E

160W

120W

80W

40W

0 W/m2 1400 1200 1000 850 750 650 550 450 350 270 190 110 70 30

60N 40N 20N EQ 20S 40S 60S 0

40E

80E

120E

160E

160W

120W

80W

40W

0

Figure 3.2 Wind power density over the global oceans in winter and summer. Source: nasa.gov.

conditions (IEC, 2009). In actuality, wind shear varies with atmospheric conditions, and average values between 0.06 and 0.16 have been observed in the North Sea and western Atlantic (Berge et al., 2009; Brower, 2012). Typical wind shear values over land average significantly higher: 0.14e0.30. Turbulence intensity (TI), which is defined as the standard deviation of wind speed samples relative to the recorded mean, is commonly observed in the offshore environment to average in the range of 0.05e0.10. High waves produced by strong winds will increase the surface roughness, which in turns increases the TI. TI values are essentially twice as high over land. Close to coastlines and islands, land influences on offshore winds grow in importance and come in many forms. When the wind has an offshore component, a transition zone within the marine atmospheric boundary layer (MABL) is created that can extend seaward for several or tens of kilometers. This transition zone, also known as an internal boundary layer, begins at the landewater interface and propagates the land’s shear and turbulence traits downwind until they are eventually mixed away (see Fig. 3.3). Winds blowing over long distances parallel to a coastline, particularly one comprised of high terrain, can be channeled to form a coastal barrier jet, which is a zone of higher-speed winds than would otherwise develop if the coast were not present. Flows between islands tend to concentrate the wind and generate higher wind speeds, while

32

Offshore Wind Farms

External boundary layer

External boundary layer

z Internal boundary layer (IBL) ∼150 m z02

z01 Land

Sea

5 km

Figure 3.3 Illustration of an internal boundary layer formed as wind blows from land to ocean. Source: Delft University of Technology OpenCourseWare.

lighter winds are experienced both upwind and downwind of the islands due to their barrier effect. The contrast in temperature between the water surface and the overlying atmosphere is an important feature of the ocean environment. This contrast impacts the stability, or the vertical mixing tendency, of the MABL. When warm air moves over cooled water, as often occurs in spring and summer in the middle latitudes of the northern hemisphere, the lower layers of the MABL become thermally stable, or less buoyant. The resulting suppression of mixing leads to stratification and the decoupling of upper air winds from near-surface winds. Sea fog and the suppression of cloud formation are signatures of this condition. This situation can also lead to the formation of low-level jets, which are zones of relatively high wind speeds in the upper layers of the MABL. The depth of a stable layer depends on the duration of the contributing weather conditions, the wind fetch, and other factors, and can be on the order of the height of wind turbines. Consequently, turbine rotors can experience high wind shears during stable conditions, which can last for days or weeks at a time. If the measurement of winds does not probe high enough to represent the entire rotor plane (up to 150e200 m above the surface), these high-shear conditions can go undetected. The same conditions can cause sea breeze circulations, especially when regional pressure gradients are weak (see Fig. 3.4). This usually occurs under the influence of a high-pressure system and relatively clear skies. During the day, strong solar heating of the land along the coast causes the overlying air to rise. Air over the adjacent water flows inland to replace the rising air, and within a few hours a circulation develops that penetrates deeper inland while drawing in air from distances farther offshore. This phenomenon can enhance offshore wind speeds up to 40 km from shore for several hours, typically from mid-afternoon to early evening; calm zones further offshore can develop as well (Steele et al., 2014). As sunset approaches, the land begins to cool and the sea breeze weakens. Overnight, a reverse but weaker circulation can develop, which is known as a land breeze. The semipermanent Bermuda High is a well-known summer feature that enables frequent sea breeze events in the eastern


33

Cooler air sinks Rising air helps form clouds Warm air rises

Sea breeze

Land heats up (heat source)

Sinking air spreads along surface Ocean is cooler compared to land (cold source, aka heat sink)

Figure 3.4 Illustration of sea breeze circulation. Source: noaa/comet program.

United States. The coincidence of strong offshore winds when daytime air conditioning loads are highest can be an attractive benefit of offshore wind development in this region (Bailey and Wilson, 2014; Dvorak et al., 2013). The opposite contrast of air and water temperaturesdcold air flowing over warm waterdis common in the fall and winter (in the northern hemisphere). This situation creates unstable conditions in the lower MABL, with vertical mixing (thermal convection) that is often earmarked by enhanced cloud formation. Vertical mixing also promotes homogeneity within the MABL, as evidenced by relatively low wind shear. Small airewater temperature differences promote near-neutral stability, which generally occurs most frequently compared to stable and unstable conditions. Because of water’s high heat capacity, the MABL’s qualities are slow to change with time compared to the boundary layer over land. The land surface quickly heats and cools during the day and night, causing the stability and wind shear to vary diurnally. Strong storms deliver the extreme wind and wave conditions experienced in the offshore environment. The two main types of storm systems are tropical cyclones and extratropical cyclones. Tropical cyclones are non-frontal warm-core storms that derive their energy from the release of latent heat of condensation. They originate over warm waters and can assume extratropical characteristics as they move poleward (see Fig. 3.5). Strong tropical cyclones are classified as hurricanes, typhoons, or cyclonesddepending on their locationdonce they attain sustained winds of 33 m/s or greater. Hurricanes occur in the Atlantic and northeast Pacific, typhoons in the northwest Pacific, and cyclones in the South Pacific and Indian Ocean. Tropical cyclones occur with greatest frequency in the summer and early fall seasons when ocean temperatures are warmest. Extratropical cyclones, which include nor’easter events that occur along the east coast of the United States, have cold-air cores and frontal features, and derive most

34

Offshore Wind Farms

Figure 3.5 Satellite image of northbound hurricane Sandy, October 28, 2012, which became an extratropical storm. Source: earthobservatory.nasa.gov.

of their energy from the temperature contrast between different air masses. They can originate over land as well as water. Peak winds can reach hurricane strength. Extratropical cyclones are often larger in radial size than tropical cyclones, with strong winds extending further from the storm center. They occur with greater frequency than tropical cyclones and prevail from mid-fall through mid-spring. Because extratropical cyclones involve colder air temperatures, they can also involve more precipitation types, which in frozen form can accumulate on structures and blades.

3.3

Essential data parameters

Sound planning and design of offshore wind plants depend on a thorough understanding of the local metocean environment. This environment is comprised of a spectrum of atmospheric and oceanographic conditions that vary in time and space. The nature of these conditions, including the extremes that may be encountered over a plant’s lifetime, must be addressed in advance to ensure the plant’s reliable long-term delivery of energy and storm survivability. The data parameters used to define metocean conditions can be grouped into three categories: wind and other meteorological variables; water- and sea bed-related variables; and joint characteristics. Some parameters are measured directly while others are derived from one or more observations. The parameters identified in this section reflect the recommendations obtained from a cross-section of leading international standards and guidelines, industry best practice


35

documents, turbine manufacturer suitability forms, and other industry experience (IEC, 2005a,b, 2009; ABS, 2013a,b; ISO, 1975; API, 2007; DNV, 2013). However, differences exist among these sources in terms of parameter measurement and modeling approaches, analytical methods, and time scales.

3.3.1

Wind and other meteorological variables

Within the atmosphere, the measurement of horizontal wind speed and direction is of paramount importance, especially at the intended turbine hub height and ideally at multiple heights, including across the height span of the turbine rotor. The differences in wind speed and direction with height lead to the derivation of wind shear and wind veer, respectively. Standard wind measurement protocols employ a sampling rate of 1e2 s and a recording/averaging interval of 10 min. The standard deviation of sampled speeds within each averaging interval divided by the mean speed for the same interval yields the turbulence intensity (TI), which is another derived parameter. Extreme gusts are derived from the sampled data; statistical methods such as the Gumbel Generalized Extreme Value are used to estimate extreme values for given return periods (typically 50 and 100 years). Other important direct measurements of the atmosphere include air temperature, atmospheric pressure, and relative humidity. All three are used to determine air density, which directly impacts turbine performance; as such, they should be measured at turbine hub height. However, if this is not feasible, height adjustments can be made using simple assumptions. Relative humidity and temperature also influence the corrosion potential of materials and coatings. Air temperature is also a factor in determining the probability of a turbine exceeding its safe operating and survival envelope (Stout, 2013). The vertical profile of temperature can be used to estimate the thermal stability of the atmosphere, as can the temperature differential between the sea surface and the overlying air. For cases where wind and other meteorological variables are not observed at hub height, extrapolation techniques are available to adjust values from other heights. For example, measured wind speeds can be adjusted to hub height using the logarithmic wind profile assumption or the power law using an appropriate shear exponent. Both extrapolation techniques are sensitive to the atmospheric stability and work most reliably under near-neutral stability conditions. To minimize the uncertainties associated with extrapolation, measurements should be taken as close to the desired height as possible. Additional weather variables that can impact wind farms and their operation include precipitation, solar radiation, lightning, and visibility. Precipitation comes in many formsdrain, freezing rain, hail, snow, etc.dand can impact turbine performance, such as from ice buildup on blades. Solar radiation data can be used to approximate blade deterioration rates and for sizing ancillary power supplies; it can also be an input for some atmospheric stability classification methods. Lightning frequency and characteristics data are useful in estimating damage and downtime risks. Visibility data are relevant to wind farm navigation marking requirements, for assessing visual impacts from shore, and to support vessel operations for construction and operations.

36

Offshore Wind Farms

Table 3.1 lists recommended meteorological data parameters to be measured and derived. Most nonwind meteorological parameters are sampled and recorded at the same frequency as wind data. Cloud-to-ground lightning statistics can be obtained from government- or privately run detection networks. Hurricane/typhoon/cyclone statistics are available from government meteorological offices.

3.3.2

Water- and sea bed-related variables

Waves, currents, and water levels comprise the primary hydrographic parameters. Short-term wave characteristics are given by a wave spectrum, which is used to determine the wave energy contained at varying wave frequencies or directions. These parameters are derived from observations of significant wave height, wave period, and wave direction. Wave steepness and breaking waves are special cases particularly relevant to wave slamming impacts on foundations (Zang et al., 2015; Peng, 2014). A long-term wave climatology is typically derived by fitting a distribution function to data observed at a site, eg, Rayleigh, Weibull, or Gumbel. Extreme value statistics may be calculated using observed parameter measurements together with empirical formulas, or by fitting observations to distribution models and projecting return times based on the observed frequency of events over a given reference period. Currents within the water column, which can vary with depth, are also referred to as current profiles. A current profile consists of wind-generated near-surface currents; subsurface currents induced by tidal fluctuations, large-scale circulations, or density gradients; and near-shore currents. The water level and its range consist of an astronomical tidal fluctuation and any storm surge. The juxtaposition of the two results in the maximum range of water levels expected at a site. Other relevant water-related variables include water temperature, density, salinity, conductivity, and ice. The parameters affecting the density of sea water, such as salinity and temperature, also affect the structural loading due to the water flow. Table 3.1

Recommended meteorological data parameters

Measured • • • • • • • • • • •

Horizontal wind speed at multiple heights Wind direction at multiple heights Vertical wind speed Inclined (off-axis) flow Barometric pressure Relative humidity Temperature Lightning (cloud-to-ground flashes) Precipitation Solar radiation Visibility

Derived • Wind speed distribution and standard deviation • Turbulence intensity • Wind shear • Extreme operating gust • Extreme coherent gust with direction change • Wind direction distribution (wind rose) • Wind veer • Air density • Thermal stability • Hurricane/typhoon/cyclone category frequencies


Table 3.2

Recommended oceanographic data parameters

Measured • • • • • • • • • • • • • • •

37

Wave height Dominant wave period Mean wave speed Wave direction and directional spectrum Current at multiple depths Still water level Tidal datum Seabed movement Temperature Salinity Conductivity Ice thickness (and other qualities) Bathymetry Soil type Scour

Derived • • • • •

Significant wave height Frequency spectrum Storm surge Water density Seismic and tsunami risk

In addition, the presence of sea ice and its physical properties can greatly affect structural loading in cold climates. Corrosion potential may be estimated from observation of water chemistry or pollution, and salinity. Water conductivity measurements are frequently used to estimate water salinity, given known or assumed proportions of dissolved salts. Water temperature also affects corrosion rates, in addition to influencing structural loading characteristics through marine growth. Estimates of storm surge and sea ice properties are ocean surface observations. For all other parameters listed here, observations throughout the depth of the water column are essential for accurately gauging conditions at development sites. Table 3.2 summarizes the recommended set of oceanographic data parameters relevant to offshore wind farms.

3.3.3

Joint characteristics

The combination of concurrent metocean factors drives the design loads analysis process as well as the wind farm’s turbine layout and energy production. Several combinations of windemeteorological, windewater, and waterewater parameter analyses are required. For example, one design load condition may evaluate the coincident severe wave height and wind speed. Another may evaluate fatigue loading under conditions when wind and wave directions vary from each other. A wind farm’s layout is strongly influenced by the joint speed-direction frequency distribution, which is used to optimally arrange and space turbines in order to minimize production losses due to wakes. Wind turbine output is a function of three coincident factors: wind speed, air density, and turbulence intensity.

38

Offshore Wind Farms

Table 3.3

Recommended joint data parameters

Joint distributions • Wind directionewave direction • Wind directionewind speed • Significant wave heightepeak spectral perioddby wave direction • Wind generated currentewind speed • Wave heightewind speed

Table 3.3 lists relevant joint metocean parameters. The desired parameters, their applications, and the available measurement and modeling technologies will likely expand over time, so it is imperative to remain abreast of offshore wind industry developments and new data needs. It is worth noting that many of the parameters presented in Tables 3.1e3.3 are common to multiple references, recommended practices, and applications; however, the means of measuring, calculating, and/or analyzing them can vary significantly. Design standards and guidelines provide procedures for certain parameters where consensus or industry best-practices are available. Among these references, however, differences exist in procedures and requirements, and many do not address all parameters equally. Some meteorological parameters, for example, may be called out as relevant for design consideration, with no guidance on how to collect, analyze or interpret them. In other cases, characteristics of certain metocean parametersdrelevant measurement frequency, return period, extrapolation method, etc.dwill be affected by project location and application.

3.4

Observational approaches

The measurement of offshore wind conditions poses particular challenges due to the hostile marine environment and the need for suitable platforms from which to take reliable observations. Historically, moored weather buoys have comprised the primary source of in situ wind measurements in offshore locales (see Fig. 3.6). They are commonly operated and maintained by government organizations to report on wind, wave and other metocean conditions at strategic locations for purposes of ocean navigation, search and rescue operations, and scientific research. However, their relatively sparse distribution and near-surface wind measurement height (typically 3e5 m above the water surface) limits their value for offshore resource assessment purposes since turbine hub heights are in the vicinity of 100 m. Weather data collected by moving ships are another source of marine data. The data are transmitted to various national meteorological services as part of the World Meteorological Organization’s Voluntary Observing Ships (VOS) program (www.vos. noaa.gov). Approximately 4000 ships participate in the program today, which is down by almost half from the peak of participation in the mid-1980s. Data are heavily


39

Figure 3.6 National data buoy center discus buoy located off the coast of Georgia, USA. Adapted from: PMEL Carbon Group http://www.pmel.noaa.gov/co2/.

concentrated along the major shipping routes in the North Atlantic and North Pacific Oceans. The quality of wind observations from ships is highly suspect because they can be recorded from visual estimates of sea state using the Beaufort scale, from observed wind effects on shipboard objects, or from anemometry which can be strongly influenced by the ship’s superstructure. Oil rigs are another source of offshore wind data, but they are concentrated in certain regions of the world, such as the North Sea and Gulf of Mexico. Only a fraction of rigs maintain continuous wind measurements, and studies have shown that measurement quality is compromised due to the distortion effects of the platforms on the wind field (Berge et al., 2009). Consequently, rig data should be applied with caution.

3.4.1

Satellite

Since the late 1980s, specialized weather satellites have provided information about the ocean’s surface winds. These mostly polar-orbiting satellites use microwave sensors to derive surface wind motions by detecting the amount of microwave radiation

40

Offshore Wind Farms

emitted or reflected by small wavelets. Because the wind is primarily responsible for the creation of these wavelets, methods have been developed to relate microwave observations to surface wind conditions. Weather buoys have been used as the main standard of comparison from which to derive statistical relationships between observed wind conditions and microwave measurements. There are three main types of sensors used on satellites that detect ocean winds. Passive microwave radiometers measure different microwave frequencies which are passively emitted from the ocean’s surface. The spatial resolution is on the order of 25 km, which limits the ability to resolve winds within this distance of a coast or island due to the corrupting influence of land. The first satellite of this typedSSM/Idwas launched in 1987 by NASA, and several others have been launched since. Other satellites of this type include TMI, AMSR-E, and WindSAT. Scatterometers are another sensor type which emit microwave pulses at the earth’s surface and receive the return signal using a common antenna. Their resolution is also approximately 25 km, which similarly limits their application in the vicinity of coastlines and islands. Satellites using scatterometers include Quick SCAT/Sea Winds and ASCAT/METOP-1. The third sensor type, which also actively emits and receives pulses, is synthetic aperture radar (SAR). In addition to surface wind definition, this technology is also used for wave measurements, oil spill detection, and other applications. SAR has the advantage of much finer resolution (generally 25 m) than the other sensor types and can therefore observe offshore winds close to shore. A disadvantage is that SAR has a much narrower field of view and provides coverage of the same given area less frequently. Satellites using SAR include: RADARSAT-1, ERS-2/SAR, ALOS, and RADARSAT-2. In all cases, the accuracy of satellite-derived wind speed estimates at 10 m above the ocean surface is on the order of 1e2 m/s. Extrapolation to heights approaching 100 m introduces significant uncertainty when estimating hub height wind speeds. It should also be pointed out that the frequency of satellite imagery for any one particular area is usually limited to twice a day or less. This frequency does not provide detail about the diurnal nature of the wind. In general practice for offshore wind energy purposes, satellite information is best used as a first-order siting tool and an indicator of regional wind speeds.

3.4.2

Measurements

Apparent from the foregoing discussion about existing marine wind measurements should be that they are unsuitable alone for characterizing the wind conditions for proposed offshore wind projects. This database of knowledge should therefore be regarded as guidance to facilitate project siting, first-order energy production estimation, and regional wind flow modeling. Investments in new measurements that are tailored to the needs and locations of offshore wind projects are a prerequisite to new project development. Purpose-built meteorological towers are the primary method for measuring site-specific wind conditions at a proposed offshore project site. The most common


41

Figure 3.7 (a) FINO3 platform with 100-m lattice-type tower. (Source: fino-offshore.de.); (b) Cape Wind platform with 60-m tapered tubular-type tower. (Source: AWS Truepower LLC.)

design is a self-supporting lattice structure, as shown in Fig. 3.7(a); Fig. 3.7(b) depicts an alternative tapered tubular design. The objective of these structures is to provide a safe and stable platform from which wind, ocean, and meteorological data can be taken for one or more years. Below the tower portion is a foundation section that is attached securely to the seabed. In addition to a large suite of sensors and mounting booms, the towers must also accommodate other features: a power supply (including battery storage), data logging and communications equipment, aviation obstruction lighting, aids to navigation, docking capabilities, biological monitoring systems, and personnel safety features. The entire tower system must be designed to withstand hurricane force winds, extreme waves and currents, and ice loading. Corrosion of structural steel must be addressed, as must the potential for scour in the design of the foundation. International standards do not yet exist to specify how offshore wind measurements must be taken, but industry best-practices encourage high levels of redundancy and sensor robustness to ensure reliable data capture. For example, three sets of anemometry are recommended for each measurement level, as opposed to two sets for land-based measurements, to achieve the desired target of 90% or greater data recovery. This is because offshore environments are harsher and more difficult to access when maintenance is required. Offshore towers are also more massive and,

42

Offshore Wind Farms

consequently, more likely to experience flow distortion issues. Sensors extending off of multiple faces of a tower can facilitate the detection and subsequent correction of flow distortions. IEC specifications (IEC 61400-12-1) recommend that wind sensors be placed on booms extending at least three to six tower widths away from the sides of towers, depending on tower solidity, to minimize distortion effects. While wind measurement programs for land-based wind projects employ multiple towers, offshore projects will most likely invest in at most one. For the most part, this is because of the high cost of an offshore tower ($10 million 50%, depending on locale and other factors), which is roughly two orders of magnitude greater than a typical land-based tower. Another consideration is that the need for multiple towers is much less justified offshore because of terrain and surface roughness uniformity within proposed project areas. In practice, a singular tower is usually positioned just outside, and upwind of, the perimeter of the proposed turbine array. This placement allows the tower to continue supplying useful observations even after the array is installed and commissioned. Numerical modeling is used to extrapolate the tower’s observed conditions throughout the project area. The high cost of offshore towers and the practical limits of their installation in waters deeper than approximately 40 m have opened the door to alternative measurement approaches. The leading candidate is profiling lidar, which can be mounted on fixed or floating platforms; sodar has been used on occasion but is suitable only on fixed platforms. On fixed platforms (such as short towers or jack-up barges), measurements taken by either remote sensing technology have shown a high degree of comparability and correlation with concurrent wind observations taken from adjacent tall towers (Cox, 2014; Hung et al., 2014; Barthelmie et al., 2003). These technologies can complement tower-based measurements by sampling winds above the top of the tower, such as across the upper portion of the turbine rotor span. Unless an existing fixed platform is available, it may be cost-prohibitive to invest in a new, purposely built one to support lidar or sodar. Buoy-mounted floating lidars (see Fig. 3.8) are now commercially available that have demonstrated the ability to measure winds aloft with essentially the same accuracy as tall-tower measurements. However, more extensive validations on long-term measurement reliability are needed before the offshore wind industry fully accepts floating lidar as a replacement for tall towers. A drawback to reliance on lidar data alone is that, unless more than one colocated lidar is used, there is no backup measurement system should there be an operational problem. Furthermore, lidar is unable to measure temperature profiles, which can readily be observed from a tall tower. A measurement period of at least one full year, and preferably two or more, is required to understand the magnitude and variability of winds and other conditions across all seasons. Compared to the 20þ year life of a wind project, this duration is relatively short and will not adequately represent long-term conditions. The measure-correlate-predict (MCP) method is commonly used to derive a long-term climatology of wind conditions from the onsite observations. The MCP process works by correlating the site’s wind observations with concurrent data from a high-quality long-term reference located within the same region, such as a coastal weather station or model-generated reanalysis grid data. The established relationship (such as a


43

Figure 3.8 Example of a buoy-mounted floating lidar. Source: AXYS.

regression equation) between the two is then applied to the reference’s historical record to predict the long-term average wind characteristics for the project site. This approach assumes, of course, that the past is a reliable indicator of the future. This has been found to be generally true within an uncertainty margin of approximately 2% for well-correlated sites (r2 > 80%) where the duration of the reference station’s wind records is at least 10 years (Brower, 2012). The lack of available wind speed measurements at or near the hub height of modern offshore wind turbines contributes significant uncertainty to wind speed estimation. The majority of publicly available offshore wind data are collected by buoys at anemometer heights of 5 m or less. Satellite-derived estimates of ocean winds are available at a 10 m height. Fig. 3.9 shows a broad range in hub height speed estimates that would result from using a range of power law shear exponents to extrapolate a known wind speed value from 5 m above the surface up to 120 m. The 5-m wind speed value of 6.7 m/s was the measured annual average observed by a north Atlantic buoy in 2013, while estimated 80-m wind speeds varied from 7.5 to 11 m/s. The average shear exponents represent a range of values that are representative of an offshore environment: • • • • •

0.05: 0.08: 0.11: 0.14: 0.17:

Extreme storm conditions, such as Nor’easters Low end of mean annual offshore wind conditions IEC-specified shear for extreme conditions for offshore wind turbines (IEC, 2009) IEC-specified operational conditions for offshore wind turbines (IEC, 2009) Mean annual near-shore wind conditions (high offshore value)

Because the shear exponent that should be used cannot be precisely determined without directly measuring the shear, and the shear may also change with height, buoys or satellite-based estimates alone are insufficient for deriving reliable information about hub height wind conditions.

44

Offshore Wind Farms

140 Height above surface (m)

7.85

8.64

10.45

9.50

11.50

120 100 Modern turbine 120 m hub height

80 60

0.05 0.08

40

0.11

Surface buoy 5 m anemometer height

20

0.14 0.17

6.70

0 6

7

8

9 Wind speed (m/s)

10

11

12

Figure 3.9 Wind speeds extrapolated to hub height from surface measurements with the power law. Source: AWS Truepower LLC.

3.5

Modeling approaches

Numerical modeling is commonly used to assimilate disparate sources of observed data to create an integrated approximation of atmospheric and oceanographic conditions within a defined time and space domain. It is commonly applied to produce weather maps, forecasts, and other products to serve the needs of the general public as well as specialized industries. The skill of numerical models for analysis and forecasting applications has advanced dramatically in recent decades, due largely to huge gains in affordable computing power and to the availability of new data inputs from sources like weather satellites. This section reviews how numerical modeling is used within the offshore wind energy sector to simulate metocean conditions, in particular winds, waves, and currents. The discussion will emphasize the modeling of winds, which not only drive the operation of wind turbines but also the generation of most waves and currents. Meteorological phenomena occur over a wide range of time and space scales. Fig. 3.10 gives an example of atmospheric processes ranging from seconds to weeks, and from meters to thousands of kilometers. The four space scalesdmicroscale, mesoscale, synoptic, and globaldrefer to the horizontal dimension of atmospheric motions, which range from short-lived microscale phenomena, such as turbulent eddies and wind gusts, to much longer-lasting global long waves and trade winds. All these scales of atmospheric motion interact with each other as well as with the land, the oceans (and other water bodies), and sea ice. In atmospheric sciences, numerical models are built around the equations of fluid dynamics, namely the NaviereStokes equations, with varying degrees of complexity (or nonlinearity). The equations may include conservation of mass, momentum,


Long waves in the westerlies

Global scale

Spatial scale

20,000 km Synoptic scale 2000 km Mesoscale 2 km Microscale 0 km

45

Small turbulent eddies

Seconds

Thunderstorms; tornadoes; waterspouts

Minutes

Sea breezes; katabatic winds

Hours

Highs & lows; weather Hurricanes; fronts tropical storms

Days

Weeks-months

Temporal scale (typical life span)

Figure 3.10 Time and size scales of atmospheric motion. Source: AWS Truepower LLC.

energy, and moisture, as well as an equation of state for air based on the ideal gas law. Numerical weather prediction (NWP) models and large-eddy simulations (LES) solve all of these equations. Due to computational runtime, cost, or other constraints, some (simpler) models solve only a subset of the equations. Although the atmosphere is always evolving and various weather variables are changing in intensity, not all numerical models are able to step forward in time. Prognostic models are ones that simulate the evolution of atmospheric conditions over time, while diagnostic models simulate steady-state conditions. Models of different types operate at different time and space scales, depending on the application. For example, climate models predict long-term changes in atmospheric properties (such as mean temperature, precipitation, and winds) over large portions of the globe (ie, at the synoptic and global scales). NWP models simulate short-term changes within smaller regions, such as portions of countries (ie, the mesoscale and synoptic scale); this scale is consistent with the size of modern wind farms. Microscale models are applied to processes in even smaller areas, such as within individual wind farms at the scale of individual turbines. Using a finite data set, essentially all models represent the environment with a three-dimensional grid. Most atmospheric models incorporate multiple vertical layers, some extending up to several kilometers in altitude. Grid resolution, particularly in the horizontal dimension, is generally consistent with the model space scale, with much coarser-resolution grid spacing employed by climate models compared to mesoscale NWP models or microscale models. The selection of grid spacing and domain size in a modeling exercise is critical when attempting to represent the flow phenomena of interest. Physical processes such as turbulence or cumulus clouds that are too small to be explicitly resolved by a model within its grid scale need to be approximated using some sort of parameterization scheme. Physical features such as mountains, islands, or irregular coastlines that are smaller than the model’s grid resolution will generally be ignored. A standard strategy to capture small features or small-scale processes with a numerical model is to run a finer-resolution grid nested inside a coarser-resolution grid. Typically, the latter

46

Offshore Wind Farms

covers a much larger region than the finer-resolution grid (similar to a box inside a box). Grid nesting is used to downscale coarse resolution information to a finer-resolution grid while ensuring proper energy transfers in the atmosphere.

3.5.1

Numerical weather prediction models

NWP models have been developed primarily for weather forecasting purposes over different time horizons ranging from hours to days. These models heavily rely on observations of initial surface and atmospheric conditions, which include surface weather stations, buoys, ships, radiosondes (weather balloons), radars, aircraft, and satellites (visible, infrared, and microwave bands). Mesoscale NWP models are well-equipped for simulating wind flows accurately in offshore environments. Several studies have demonstrated their ability to represent many of the complex wind phenomena found in offshore environments: mountain and island blocking, gap flows, coastal barrier jets, internal boundary layer growth, stability transitions, sea breeze circulations, and so on (eg, Colle and Novak, 2010; Freedman et al., 2010; Steele et al., 2013; Gilliam et al., 2004). The root mean square error (RMSE) of wind speed data from NWP models is typically around 2e3 m/s in offshore regions (Jimenez et al., 2007; Berge et al., 2011; Beaucage et al., 2007; Dvorak et al., 2010). In addition to wind speed components at several heights, NWP models can output almost any atmospheric variable. The typical model resolution for most mesoscale simulations is on the order of a few kilometers, ie, near the interface between the microscale and mesoscale. Since this scale does not provide a very detailed picture of wind conditions within a large wind farm, coupling with a microscale model is often done to obtain the desired detail. It has been demonstrated that a coupled mesoscale NWP and microscale model shows improvement over a mesoscale model alone. Examples of coupled mesoscale and microscale models include AWS Truepower’s MesoMap and SiteWind systems (Brower, 1999), Risø National Laboratory’s KAMM-WAsP system (Frank et al., 2001), and Environment Canada’s AnemoScope system (Yu et al., 2006). Coupled model approaches have been used to create relatively high-resolution wind maps and atlases of the globe. Fig. 3.11 is a representation of the annual average wind speed of the United States, including a 90-km wide zone of offshore winds, at 100 m above the surface at a spatial resolution of 2 km (Elliott et al., 2010; Schwartz et al., 2010). In northern Europe, the NORSEWIND (NORthern Seas Wind Index Database) project, which began in 2008, has develop wind atlases for the Irish Sea, the North Sea, and the Baltic Sea using offshore wind measurements, satellite data, and numerical model data (Hasager et al., 2010). Mesoscale models take into account subgrid scale effects and physics parameterizations for solar radiation, land surfaceeatmosphere interaction, the planetary boundary layer (PBL), turbulence, cloud convection, and cloud microphysics. Since they incorporate the dimensions of both energy and time, NWP models are capable of simulating such phenomena as thermally driven mesoscale circulations (eg, sea breezes, thunderstorms) and atmospheric stability, or buoyancy. In the world of mesoscale modelingdas in the real worlddthe wind is never in equilibrium with the surface because of the constant exchange of energy. This exchange occurs through solar


47

United States—land-based and offshore annual average wind speed at 100 m

Wind speed m/s >10.5 10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 fi1 > > > 8 9 > > > > > > > > > < > = > = < > < 1 = 0 0 1 b i hi x y x 0 y 0 ¼ þG « « i2m i1m : > > > > ; li i1m i2m > > > > > 1 > > > > > > > : ; ; : fi2 fi2

1

3 5

1 [6.50]

fug ¼ ½K1 ffg with boundary condition of arbitrary node set to 0. With the warping ordinate known we can then calculate the position of the shear centre relative to the elastic centre as shown in Eq. [6.51], as well as the warping constant relative to the shear centre.

128

Offshore Wind Farms

EAu ¼

i¼n X 1

2

i¼1

bi ðui1 þ ui2 Þli hi E

EAxu ¼

i¼n X

1 bi ð2xi1m þ xi1m Þui1 þ ðxi1m þ 2xi1m Þui2 li hi E 6 i¼1

EAyu ¼

i¼n X

1 bi ð2yi1m þ yi1m Þui1 þ ðyi1m þ 2yi1m Þui2 li hi E 6 i¼1

EAu uk ¼ EA x0sc ¼

[6.51]

EAyu EIxx0

y0sc ¼

EAxu EIyy0

fug ¼ fug uk þ fxgy0sc fygx0sc The torsional stiffness constant GJ can then be calculated using Eq. [6.52]. It is composed of an open section and a closed section component. If the section is closed (as is typically the case for wind turbine blades) then the closed section constant will usually be dominant. The value of GJ, along with the location of the shear centre, is often useful as an input for aeroelastic software. GJ ¼ GJopen þ GJclosed GJopen ¼

i¼n X 1 i¼1

GJclosed ¼

3

b i li h3 G i

[6.52]

i¼n X b i hi ðrti li þ ui1 ui2 Þ rti G i¼1

where rti ¼

x0i1m

x0sc

y0i2m y0i1m x0i2m x0i1m 0 0 yi1m ysc li li

We can then calculate shear flows for use in Eq. [6.36], which will allow us to calculate the stress state in any ply. There are shear flows due to shear forces and shear flows due to torsion. In order to calculate the shear flows due to shear forces we must

Design of offshore wind turbine blades

129

first calculate the shear deformations using the finite element method, as shown in Eq. [6.53]. 8 9 8 9 fi1 > fi1 > > > ( ) > < > < > = > = 1=3Vi1 þ 1=6Vi2 b ¼ þ E i hi l i « « > > > > > 1=6Vi1 þ 1=3Vi2 : > : > ; > ; fi2 fi2

[6.53]

where Vij ¼

Fy0 0 EIxx

y0ijm þ

Fx0 0 xijm 0 EIyy

fug ¼ ½K1 ffg with boundary condition of arbitrary node set to 0 and stiffness matrix as calculated for Eq. [6.50]. The shear flow in each element can then be calculated as the sum of the shear flow due to shear forces applied at the shear centre and the torsional moment, as shown in Eq. [6.54]. qxyi

b i Mz 2 b i Mz b i hi G G G ui1 ui2 hi þ hi þ ¼ þ rti ðui2 ui1 Þ GJ li GJ li

[6.54]

h i li þ ðVi1 þ Vi2 Þ 24 This is a computationally efficient method of calculating cross-sectional properties and stresses in each ply, which makes it ideal for use in design optimisation studies. It allows us to provide all of the information required by the DNV-GL guidelines [11], which specify the verifications for fibre failure and interfibre failure according to Puck [14].

6.4.5

Stress analysis using finite element analysis

Once the ultimate and fatigue loads for the blade design in question have been determined using a wind turbine simulation code, a more in-depth stress analysis must then be performed to check whether the blade will be able to survive the service loads. The shell-like structure of wind turbine blades described in the previous section means that analysis is usually performed using a finite element model. The models are usually built up from shell elements, although it is possible to use brick elements to represent the thicker parts of the blade and the glue. Shell elements are used when the thickness of the structure is much smaller than the other two dimensions. Typically it is not possible to calculate through-thickness stresses (although continuum shell elements do allow these stresses to be calculated at the expense of increasing the complexity of the model). This has implications

130

Offshore Wind Farms

when modelling composites because delamination (separation of the layers of the composite) occurs partly due to out-of-plane stresses. Brick elements are useful where a structure has to be modelled and it cannot be considered as a shell. In a wind turbine blade this condition may occur at the trailing edge or where there is a thick build-up of glue. The benefits of this class of element are increased model sophistication, but this comes at the expense of difficulty in modelling. A typical blade finite element model is shown in Fig. 6.10. There are many pieces of software for streamlining the time-consuming process of calculating the laminate properties and building up the blade model. Some of these are independent, such as NuMAD by Sandia National Laboratories, and some have been developed by blade companies themselves, such as LM Blades by LM Wind Power. Buckling analyses will usually be performed using finite element tools, and can either be linear or non-linear. The DNV-GL design guidelines give instructions for performing stability analyses [11].

6.4.6

Leading-edge erosion

Leading-edge erosion occurs when raindrops, hailstones or other particles impact the leading edge of the blade. The problem is worst near the tip of the blade, where the speed is highest, and it typically results in the gelcoat cracking and falling off, followed by damage to the underlying composite layers if the issue is not addressed. Currently, leading-edge erosion limits tip speeds to around 100 m/s. The problem is well known in the helicopter industry where the tip speeds are far higher than those that would ever be likely to be encountered on a wind turbine (of the order of 200 m/s). They solve it by applying aluminium sheet to the leading edge which is replaced on a regular basis, however, this is not an acceptable solution for wind turbines. The latest efforts to combat this issue have focused around applying tough pliant coatings to better absorb the impacts. These are often polyurethane elastomer coatings, which can be applied either as a repair in the form of tapes or ‘in mould’ as a preventative measure. These coatings can greatly increase the resistance to leading-edge Y X Z

Figure 6.10 Blade finite element model.


131

erosion, effectively increasing the limit on tip speed due to erosion. If the coatings are applied as tapes then they can impact energy production because of the step change on the blade surface. However, the impact is lower than would result from a damaged leading edge.

6.5

Manufacture

Manufacturing methods for wind turbine blades were originally derived from those used in the construction of yacht hulls. As the industry has grown, less labourintensive methods with better process control have been adopted. The methods that are used in the industry today are discussed in this section, and their benefits and drawbacks are discussed.

6.5.1

Wet lay-up

This processing technique was widely used in the early days of the industry. Open moulds are created for the pressure and suction sides of the blade and any internal webs. The insides of the moulds are painted with gelcoat, and then the fibre mats are placed in the mould. Once the resin and hardener have been mixed, they are poured into the mould and spread around using rollers to work the matrix material in and ensure adequate wetting of the fibres. The blade components are then left to cure and the blade is assembled using adhesives. This processing technique has fallen out of favour. It is very labour-intensive and there are health and environmental concerns around the use of open mould processes because of the release of volatile organic compounds. In addition, the resins need to have a low viscosity in order to be workable by hand (which generally results in lower material properties) and it is hard to achieve the fibre volumes that are possible with more advanced processes without voids.

6.5.2

Resin transfer moulding and vacuum infusion

Resin transfer moulding typically involves laying up the fibre mats as for a wet lay-up process. The mould is then covered by a vacuum bag, and resin and hardener are pumped into the mould whilst a vacuum pump ensures that the plies are consolidated and that all fibres are wetted. The advantage of this process is a higher level of automation and it is also easier to control the emissions from the process because it is taking place inside the vacuum bag. Careful placement of the resin injection points is essential to ensure that the resin flows through the mould correctly and wets all the fibres. One downside of this process is the fact that a lot of waste vacuum bags are produced. Most companies using this approach would then assemble the individual blade components (pressure and suction side mouldings and spar or shear webs) by bonding the components together using glue, but Siemens have developed the patented

132

Offshore Wind Farms

IntegralBlade technology, where the entire blade is made in a ‘one-shot’ process without bond lines.

6.5.3

Prepreg technology

Prepreg mats were first developed for the aircraft industry. The mats are preimpregnated with an uncured mixture of resin and hardener which is in a tacky solid state. The blade is layed up in the usual manner, and then vacuum bagged to consolidate the layers. In contrast to other vacuum bagging techniques though, the mould must then be heated to change the matrix material into a viscous liquid state and cure the material. For wind turbine laminates the temperature required is typically 80 C, which of course has implications for energy use during manufacture. The main disadvantage of prepreg materials is that they have a limited shelf life and need to be stored at low temperature to prevent them from curing (18 C is typical, at which a shelf life of 6e12 months is typical). However, the process has big advantages in terms of process control, an improved working environment and achieving high-volume fraction (and hence improved materials properties) whilst ensuring that no voids are present.

Nomenclature a 0

Axial induction factor

a

Tangential induction factor

A

Intercept parameter of fit to fatigue data

[A], [B], [D]

Laminate extensional, coupling and bending compliance matrices

[A*], [B*], [D*]

Laminate extensional, coupling and bending stiffness matrices

Aij, Bij, Dij

Laminate matrix components at row i column j

B

Number of blades, slope parameter of fit to fatigue data

CL

Lift coefficient

CD

Drag coefficient

c

Chord length

[C]

Ply stiffness matrix

CM(x0)

Moment coefficient at rotor proportion x0

CP

Power coefficient

CT

Thrust coefficient

D

PalmgreneMiner damage sum

m


133

dD

Drag force on annular ring

N/m

dFx

Axial force on annular ring

N/m

dL

Lift force on annular ring

N/m

dT

Torque on annular ring

N

E1, E2

Young’s moduli in ply 1 and 2 directions

N/m2

EF, EM

Fibre and matrix Young’s moduli

N/m2

bx E

Laminate equivalent Young’s modulus in x-direction

N/m2

EA

Beam extensional stiffness

N

EIxx, EIyy, EIxy

Beam cross-section bending stiffness about section x-axis, y-axis and product bending stiffness

Nm2

EIxx0 , EIyy0

Beam cross-section bending stiffness about principal x and y axes

Nm2

Fx, Fy, Fz, Mx, My, Mz

Shear forces in x and y directions, axial force, bending moment about x and y axes, torsional moment

N

g1, g2

Simplifying expressions

b xy G

Laminate equivalent shear modulus in xy direction

N/m2

GF, GM

Fibre and matrix shear moduli

N/m2

G12

Ply shear modulus in 1e2 direction

N/m2

GJ

Torsional stiffness

Nm2

h

Laminate thickness

m

I

Moment of inertia of stream tube

KE

Kinetic energy

k

Lift/drag ratio

{K}

Laminate midplane curvature

l

Length of cross-sectional element

m

L_

Rate of change of angular momentum

kg m rad/s2

m

Mass

kg

m_

Mass flow rate

kg/s

{M}

Laminate moment vector

N

Number of cycles to failure

n

Number of fatigue cycles

{N}

Laminate force vector

Nx, Ny, qxy

Normal, transverse and shear flows

Joules

N/m Continued

P

Power

Watts

p

Pressure

N/m2

Q

Tip loss factor

r

Radius

m

R

Rotor radius, R-value

m

[S]

Ply compliance matrix

S

Stress, strain or load in a fatigue cycle

t

Time

[Tk]

Rotation matrix for ply k

{u}

Shear deformations

m

v

Velocity

m/s

W

Resultant velocity

m/s

x

Proportion of rotor span, x-coordinate of cross-section

m

y

y-coordinate of cross-section

m

zk

Distance from laminate midplane of ply k

m

a

Angle of attack, angle of cross-sectional line element to x-axis

rad

b

Blade set angle

rad

{ε}

Ply strain vector

{ε0} ε0x ,

ε0y ,

s

Laminate midplane strain vector g0xy

Laminate midplane normal, transverse and shear strains

l

Tip-speed ratio

L

Local blade geometry parameter

yF, yM

Fibre and matrix Poisson’s ratio

n12

Ply Poisson’s ratio

f

Flow angle of resultant velocity at rotor plane, angle of principal axes

fF

Fibre volume fraction

r

Density

{s}

Ply stress vector

scx, scy, scxy

Ply stress in laminate coordinate system

N/m2

sc1, sc2, sc12

Ply stress in ply coordinate system

N/m2

q

Ply angle

rad

U

Rotor rotational speed

rad/s

u

Angular velocity of wake

rad/s

{u}

Vector of warping displacements

rad

kg/m3


135

References [1] P. Jamieson, Innovation in Wind Turbine Design, Wiley-Blackwell, Hoboken, New Jersey, 2011. [2] T. Burton, D. Sharpe, N. Jenkins, E. Bossanyi, Wind Energy Handbook, John Wiley & Sons, Ltd, Chichester, 2001. [3] P. Brøndsted, H. Lilholt, A. Lystrup, Composite materials for wind power turbine blades, Annu. Rev. Mater. Res. 35 (1) (2005) 505e538. [4] L.P. Kollar, G.S. Springer, Mechanics of Composite Structures, Cambridge University Press, Cambridge, 2003. [5] M. Mutsuiski, T. Endo, Fatigue of Metals Subjected to Varying Stress, Japan Society of Mechanical Engineers, Kyushu, Japan, 1969. [6] S.D. Downing, D.F. Socie, Simple rainflow counting algorithms, Int. J. Fatigue 4 (1) (1982) 31e40. [7] L. Gornet, O. Wesphal, C. Burtin, J.-L. Bailleul, P. Rozycki, L. Stainier, Rapid determination of the high cycle fatigue limit curve of carbon fiber epoxy matrix composite laminates by thermography methodology: tests and finite element simulations, in: Fatigue Design, CETIM, Paris, 2013. [8] A. Palmgren, Die lebensdauer von kugellagern, VDI-Zeitschrift 68 (14) (1924) 339e341. [9] M.A. Miner, Cumulative damage in fatigue, J. Appl. Mech. 12 (3) (1945) 159e164. [10] IEC, IEC 61400-1 Wind Turbines e Design Requirements, IEC, 2005. [11] DNV-GL, Guideline for the Certification of Wind Turbines, DNV-GL, 2010. [12] O.A. Bauchau, J.I. Craig, Structural Analysis e with Applications to Aerospace Structures, Springer, London, 2009. [13] R. Kindmann, M. Kraus, Steel Structures Design Using FEM, Ernst and Sohn, Berlin, 2011. [14] A. Puck, Strength Analysis of Fibre-Matrix Laminates: Models for Practical Applications, Hauser, Munich and Vienna, 1996.


Wind turbine gearbox design with drivetrain dynamic analysis

7

S. McFadden Ulster University, Magee Campus, Northern Ireland, United Kingdom B. Basu Trinity College Dublin, Dublin, Ireland

7.1

Introduction

The gearbox (GB) is an integral part of the drivetrain in large-scale wind turbine generator systems (WTGS). Gearboxes are used in the majority of WTGS with synchronous or induction generators installed for electricity generation [1]. Small wind turbine rotors may turn at speeds of the order of hundreds of revolutions per minute and may not require a gearbox. In these cases, known as direct drive, permanent magnet generators or synchronous generators with a sufficiently high number of poles may be directly coupled to the rotor. However, large wind turbines turn more slowly and a gearbox is a practical necessity. Horizontal axis wind turbines (HAWT) with power ratings in excess of 500 kW have a dedicated ISO/IEC standard for gearbox design [2]. Fig. 7.1 shows a schematic layout of the main subsystems found in a typical HAWT. The gearbox is coupled to the rotor through the low-speed shaft and is connected to the generator through the high-speed shaft. It is the convention in GB design to distinguish between the rotor side (RS) and the generator side (GS). The basic function of the gearbox is to transmit mechanical shaft power from the rotor side, running at low speeds (approximately10e18 RPM for offshore wind turbines), to the generator side, running at higher speeds (1800 RPM for 60 Hz or 1500 RPM for 50 Hz assuming a four-pole synchronous generator) [1,3]. For safety reasons a mechanical brake may be installed on the high-speed shaft. However, for installation or maintenance, a rotor lock system is also installed on the low-speed shaft (main shaft). Statistics have shown that operating and maintenance (O&M) costs for HAWT are significant. Marquez et al. [4] reported that for a 750-kW turbine over a 20-year operating life, the O&M costs are 25e30% of overall energy generation [5] or 75e90% of investment cost [6]. In addition, Marquez et al. [4] reported data from [7] that suggests that larger turbines fail more frequently. Van Bussel et al. [8] also stated that 25% of O&M costs are due to failures of main components. It is clear that all components and subsystems contribute the O&M costs; however, as highlighted by Sheng [9], the gearbox can be one of the most costly WTGS subsystems to maintain over the 20-year operating life. Data provided by [10] (reported in Igba et al. [11]) show that annual gearbox failure rates may be relatively low when compared to other


138

Offshore Wind Farms

Low-speed shaft on bearings

High-speed shaft with mechanical brake

Electrical generator

Rotor

GB Gearbox

Figure 7.1 HAWT schematic showing the main subsystems in the drivetrain (modular architecture).

subsystems, but the downtime caused by the gearbox failure is relatively high and is, therefore, a major cost driver. Given this context, we can see the importance of gearbox design and the related operational issues. Since WT gearbox is a mature technology, the design process tends to be iterative, but from the viewpoint of categorising the design process, we can discuss a generic design life cycle as shown in Fig. 7.2 (reported in Igba et al. [11]). The life cycle consists of five stages: concept, development (detailed design), production, utilisation and support, and retirement. At the concept stage, the design engineer is required to understand the technical problems at hand. A problem statement should be outlined within a design requirements document or product design specification. Several concept variants may provide a working solution to the problem. It is typical that several concept variants are evaluated against technical or economic criteria. The most appropriate concept is proposed for the development (or detailed design) stage. The detailed design stage considers all of the gearbox components and subsystems. The detailed design includes the integration of the various systems and would typically culminate in a verification or validation test at component or subsystem level. Once the

Concept stage

Development stage (detailed design)

Production stage

Utilisation and support stage

Retirement stage

Figure 7.2 Generic design life cycle. Adapted from Igba J, Alemzadeh K, Durugbo C, Henningsen K. Performance assessment of wind turbine gearboxes using in-service data: current approaches and future trends. Renewable Sustainable Energy Rev 2015;50:144e59.


139

development stage is completed and verified, the design will proceed to the production stage. The production stage may be viewed as component manufacture and integration (assembly). The utilisation and support stage is typically the longest stage of the life cycle. The retirement stage involves decommissioning of the gearbox and recycle or salvage of components. In this chapter we shall review the standard approach to the design of WTGS gearboxes and demonstrate a WTGS drivetrain dynamic analysis. As mentioned above, a designated standard exists for gearbox design in WTGS with power ratings in excess of 500 kW (ISO/IEC 61400-4). This standard was jointly prepared by IEC technical committee 88: ‘Wind turbines’ and ISO technical committee 60 ‘Gears’. This standard is part 4 of the series of standards for WTGS, IEC 61400, prepared by IEC. The standard applies to WTGS installed onshore or offshore. The gearbox standard (ISO/IEC 61400-4) applies to gears and gear elements on the main power path, auxiliary gearing (for example, power take off or yaw system drives) is excluded or covered elsewhere. This chapter will review ISO/IEC 61400-4 in the context of the design life cycle stages presented in Fig. 7.2. It should be noted that ISO/IEC 61400-4 presents a flowchart that highlights the iterative nature of GB design at the detailed stage, however the generic design life cycle still applies in a broad sense. Specifically, at concept stage (Section 7.2), we shall present the basic working principles and concept variants of typical WTGS gearboxes; at development stage (Section 7.3), we shall review the standard requirements for gears, bearings, structural elements, and lubrication and sealing in WTGS gearboxes; and at production stage (Section 7.4), we shall review the issues around component manufacture and assembly of gearboxes. Section 7.5 describes a DTD modelling tool that simulates the gearbox performance under dynamic loading events. Such tools are essential in the prediction of WTGS performance at the development and detailed design stage. Finally, some conclusions are made on the overall design process for the WTGS gearbox. Issues relating to reliability are discussed. The requirement to have fault detection and condition monitoring systems on the WTGS gearbox is briefly highlighted.

7.2

WTGS gearbox design e concept stage

Before setting out on a design exercise, the engineer must understand the problem at hand. To aid with this understanding, this section begins with an overview of the basic operation of the WTGS gearbox. The section will follow on to review some early stage design considerations and lead onto the concept variants that the designer may consider at the early stage of the design. To provide an overview, the diagram in Fig. 7.3 outlines the steps involved in the concept design of a gearbox.

7.2.1

Basic operation of the WTGS gearbox

In WTGS, the gearbox is speed-increasing, where the defining parameter is the gearbox speed ratio, u. The speed ratio relates the input speed on the rotor side, UIN, to the output speed on the generator side, UOUT.

140

Offshore Wind Farms

Steps for gearbox concept design: Step 1. Define gear ratio according to required input and output speed. Step 2. Define gearbox input torque according to WTGS power and gearbox input speed. Step 3. Design a gear stage including number of stages, gear split ratio and gear macro design. Step 4. Selection of bearing according to chosen gear stages. Step 5. Design of housing. Step 6. Efficiency check and design of lubrication/cooling system. Step 7. Finalise gearbox concept design including manufacturability, assembly.

Figure 7.3 Concept design of a gearbox.

hg ¼

UOUT UIN

[7.1]

The speed-up ratio may be cited as a ratio (X:1) or as a number (where hg > 1 in the case of speed-increasing gearboxes). Taking the example where a rotor is rotating at 50 RPM and the generator is rotating at 1500 RPM, the speed-up ratio is 30:1 or hg ¼ 30. In WTGS, the gear box speed ratio may be in the vicinity of 100:1. From a power perspective, the gearbox receives power on the rotor side, PROTOR, and delivers power on the generator side, PGB. The gearbox will experience power loss, PLOSS, as heat is generated due to friction between parts with relative motion (shaft seals, gears, bearings, etc.) and viscous losses (lubrication system). Conservation of energy will apply to give PROTOR ¼ PGB þ PLOSS :

[7.2]

However, the gearbox output power is usually related to the power from the rotor by the gearbox efficiency which is mainly affected by gear loss and oil churning loss, hM. PGB ¼ hM PROTOR

[7.3]


141

Furthermore, the relationship between the gearbox power on the generator side and the wind power, PWIND, is given by PGB ¼ hM Cp PWIND

[7.4]

where CP is known as the power coefficient for the rotor. Theoretically, the power coefficient for a rotor cannot exceed the Betz limit (equal to 16/27, approximately 0.593). Typically, the power coefficients for HAWT are in the range 0.3e0.5. Fig. 7.4(a) shows some typical wind tunnel test data for an HAWT rotor. The plot on the left shows rotor power versus rotor speed for three different wind speeds. At each wind speed, the rotor power has a peak value at a specific rotor speed. As the wind speed increases, the overall power levels show an increasing trend. The same data as Fig. 7.4(a) are used to generate Fig. 7.4(b). Here, the classic plot of power coefficient, CP, versus tip-speed ratio, l, is shown. Similar to rotor power, the power coefficient has a peak value at a specific tip-speed ratio. Torque is an important consideration in gearbox design. The torque is required to determine the gear forces at the contact points where gears mesh with each other. In addition, the torque applied to the gear train must be reacted by the structural components in the gearbox: the housing, the mounting points, flange couplings, torque arms, etc. Since the shaft power is the product of torque and rotational speed we can expand Eq. [7.3] to give TGB UOUT ¼ hM TROTOR UIN

[7.5]

where TROTOR and TGB are the torques on the rotor side (ie, gearbox input) and the generator side (ie, gearbox output), respectively. By rearranging Eq. [7.5], we get the following relationship TGB UIN ¼ hM TROTOR UOUT

[7.6]

(a)

(b)

Rotor power

Power coefficient, CP

Wind speed = 4 m/s Wind speed = 6 m/s Wind speed = 8 m/s

Wind speed = 4 m/s Wind speed = 6 m/s Wind speed = 8 m/s

Rotor speed

Tip-speed ratio, λ

Figure 7.4 Typical wind tunnel data for an HAWT turbine. Data adapted from Mathew S. Wind energy: fundamentals, resource analysis and economics. Heidelberg: Springer, Berlin; 2006.

142

Offshore Wind Farms

Recognising that the speed relationship on the right-hand side of this equation is a kinematic constraint described by Eq. [7.1], we come to the important conclusion that the effect of power loss through the gearbox is manifested as a reduction in the ideal torque output (where the ideal torque is TRotor/hg). TGB ¼ hM TROTOR hg

[7.7]

Following on from Fig. 7.4, Fig. 7.5 shows how the torqueespeed relationships change through the gearbox from the rotor side to the generator side. The wind tunnel power data from Fig. 7.4 were used to get the rotor torqueespeed curves on the left-hand side. As the wind speed increases, the torque levels increase. However, assuming a constant wind speed, as the shaft speed increases the torque decreases. To aid with the discussion, lines of constant power are superimposed onto the plot area. Each line of constant power represents the peak rotor power at the respective wind speed. These curves show that, even with constant power, the torque must decrease as the shaft speed increases. Eq. [7.7] is applied to the rotor torqueespeed curves to give the gearbox torqueespeed curves on the bottom right of the plot. It can be clearly seen that in the case of a speed-increasing gearbox, the specific speed values and the range of speed values are increased. On the other hand, the torque levels and the range of the torque values are decreased. Additionally, the torques at the gearbox GS are further decreased by the mechanical efficiency. If the gearbox had a 100% mechanical efficiency, the GB torque curves would be tangential to their respective lines of constant power, but this is not the case due to the power losses in the GB. Following this analysis, a few important observations can be made in relation to the final design. Because the rotor is running at the lowest speed, it is also delivering the Wind speed = 4 m/s Wind speed = 6 m/s Wind speed = 8 m/s

Torque

Rotor torques

Constant power

GB torques

Shaft speed

Figure 7.5 Torqueespeed on the rotor side (rotor torques) and the generator side (GB torques) for HAWT wind tunnel data.


143

highest torque. Hence, the diameter of the low-speed shaft is large compared to the diameter of the high-speed shaft. The low-speed shaft requires a large, sturdy bearings set to handle the relatively large torques and forces exerted on the rotor. It should be noted that there is a distinct advantage to having the mechanical brake on the high-speed shaft. A torque applied by the mechanical brake on the high-speed shaft is amplified through the gearbox onto the rotor by the inverse effect to that just discussed here.

7.2.2

Early stage design considerations

At the outset, the gearbox designers must develop the gearbox conceptual design with inputs such as input torque, gear ratio, interface requirement, maximum envelope, maximum weight, and manufacturability, etc., from the wind turbine manufacturer and other key suppliers in the design process. Design for reliability is a particular goal, with a design life of 20 years as a target. ISO/IEC 61400-4 recommends that all relevant parties should engage at an early stage to complete a critical system analysis. The standard suggests a list of topics for consideration in the system analysis. A Failure Modes and Effects Analysis (FMEA) is suggested [12,13]. The gearbox design must take into account the complete system architecture and, in particular, the interface requirements with other subsystems in the WTGS. The gearbox will have mechanical interfaces with other parts of the WTGS, for example, mechanical couplings to the low-speed shaft, the high-speed shaft, and the nacelle mainframe. The gearbox will have a lubrication system and will interface with external lubrication components such as reservoirs, coolers, pumps and filters. Other interfaces may exist which relate to the flow and control of sensory information or signals to and from the gearbox. All interfaces must be identified at the early stage of the design. In some instances the interfaces may overlap due to integration in the system architecture. Consider, for example, the low-speed shaft. A modular design concept (as shown in Fig. 7.1) may be adopted, whereby the low-speed shaft runs on two separate bearings mounted externally from the gearbox. An alternative modular configuration, known as a drivetrain with a three-point supporting structure (suspension), has one bearing on the low-speed shaft and another integral to the gearbox. In this case the bearing internal to the gearbox will assist with reacting the bending moments and forces on the main shaft. A fully integrated drivetrain is also possible, whereby the internal gearbox bearings on the rotor side will carry all of the low-speed shaft loads. An important aspect in engineering design of any machine is the development of relevant documentation. ISO/IEC 61400-4 gives a list of recommended documentation titles for the entire design cycle. A product design specification should be outlined at the early stage in any design process. Among the recommendations, the ISO/IEC 61400-4 standard suggests that two initial documents be prepared and issued by the wind turbine manufacturer: (1) a general specification and (2) a load specification. The general specification would outline general details on technical performance along with any supplementary information. The load specification would provide information for detailed design in respect of design loads, fatigue loads, extreme loads, etc.

144

Offshore Wind Farms

7.2.3

Concept variants for WTGS gearbox

Two basic gear layout arrangements have dominated WTGS gearbox applications: (1) gears on parallel shafts and (2) planetary gear systems (also known as an epicyclic gear system). A typical gearbox speed ratio in large WTGS may be around 30:1 to 100: 1. This represents a significant change in rotational speed. A single gear stage is limited to give a speed ratio of approximately 6:1 [3]. Hence, most WTGS gearboxes require multiple stages to the desired overall speed ratios. For example, to achieve a speed ratio of 100:1, three stages could be employed. Fig. 7.6 shows a schematic for a single-stage, parallel shaft gear arrangement. The input gear (known as the driving gear) has a number of teeth equal to NA, whereas the output gear (known as the driven gear) has a number of teeth equal to NB. The speed ratio for this particular setup is given as r ¼ hg ¼

NA : NB

[7.8]

It is clear that to obtain a speed-increasing ratio (r > 1) then the number of teeth on the input gear must be greater than the number of teeth on the output. Fig. 7.7 shows an arrangement for a two-stage, parallel shaft arrangement that will obtain a higher overall speed increase. Power, PIN, is applied at the input, power, POUT, is delivered from the output gear. In this case, an intermediate shaft is introduced to carry shaft power from the input gear to the output gear. The overall speed ratio is given as the product of each gear stage. A planetary gear set can give a higher power-to-weight ratio than a parallel shaft gear set while keeping the input and output shaft coaxial. The planetary gear system

Input gear teeth = NA

Output gear teeth = NB

Figure 7.6 Parallel gear shaft arrangement (single stage).


PIn

145

POut

Figure 7.7 Parallel gear shaft arrangement. Two stages with intermediate shaft.

is named after its likeness to an orbital planetary system. A gear in the centre is called the sun gear, the gears rotating around the sun are called planet gears, and the gear on the outside is called a ring gear. The planet gears are usually equally spaced around the sun gear and are connected to each other by planet pins onto a planet carrier. The ring gear is an internal gear with all other gears being external gears. To increase the speed, the ring gear is held in position (ie, fixed), input power applied to the planet carrier, and output power taken from the sun gear. Fig. 7.8 shows how the gears are arranged. The planetary gear set benefits from power branching, that is, the transmission forces are shared among multiple planet gears. Power branching helps to reduce the loads on individual gear teeth. The planetary gear system is well-suited to carrying high torques. The number of planet gears and the number of gear teeth are not arbitrary [14]. The pitch for the sun, planets, and ring gears must match. With an equally spaced planetary arrangement with n planet gears, the following relationship must hold: NS þ NR 2ðNS þ NP Þ ¼ Integer value ¼ n n

[7.9]

where NR is the number of teeth on the ring gear, NS is the number of teeth on the sun gear, and NP is the number of teeth on the planet gears. In addition, the following relationship must hold between gear teeth NR ¼ 2NP þ NS

[7.10]

146

Offshore Wind Farms

Ring gear (fixed)

NR = Ring teeth NP = Planet teeth NS = Sun teeth

Planet carrier Sun gear

POut

PIn

Planet gear (with bearings) Planet pin

Figure 7.8 Example of planetary gear stage with four planet gears.

The gear ratio from a fixed ring, planetary arrangement is given by the relationship r ¼ hg ¼

NR þ NS NS

[7.11]

Planetary gear stages may be used in series with parallel axis gear stages to get the overall desired ratio. As mentioned above, two or three stages may be required to get the overall ratio. Typically, due to the higher torque requirement on the rotor side, planetary gear sets are used in the first and/or second stages of the drivetrain. For large offshore wind turbines, the use of a split torque planetary gearbox is currently gaining popularity. Split torque systems have the advantage of increasing power density. These systems can transmit a very high torque in a very small space, which has made such systems attractive for offshore wind turbine applications. These gear systems use the principle of division of the transmission force between several contact areas, which in turn increases the contact ratio.

7.3

WTGS gearbox design e development stage

After the conceptual design has been selected and the major decisions have been made (modular or integrated layout, number of stages, planetary or parallel axis, etc.) the detailed design stage must begin. The FMEA, if conducted, should have led the designer to consider the major analyses required to make the design safe. ISO/IEC 61400-4 makes recommendations on determining the drivetrain operating conditions and loads. Recommendations are made regarding the use of computer simulations of drivetrain loads. In particular, ISO/IEC 61400-4 outlines the practices for determining time series, fatigue, and extreme loading scenarios. Fig. 7.9 is


147

Concept choice

Define drivetrain model and pertinent design interfaces

Define interface assumptions

Define design load cases

Critical systems analysis

Data output Run ‘loads’ simulation

Time series

Local load distribution Load revolution distribution Rain flow count Extreme loads

Gearbox design

Prototype gearbox

Gearbox workshop test Test specification Documentation input

Field test

Figure 7.9 Design process flowchart e development stage. Adapted from ISO/IEC 61400-4. Wind turbines e part 4: design requirements for wind turbines gearboxes. 2012.

148

Offshore Wind Farms

adapted from ISO/IEC 61400-4 and gives a suggested design process workflow for WT gearbox. The ISO/IEC standard makes particular recommendations in several design aspects, namely; gearbox cooling; gears; bearings, shafts, keys, housing joints, splines and fasteners; structural elements; and lubrication.

7.3.1

Gear design

Internally within the gearbox, the gears closest to the rotor side carry the highest torques. The torque gets progressively smaller as the rotational speed of the gear shafts increase through each gear stage. Hence, the gear stage closest to the rotor side is usually designed to take greater torque than the gear stage at the generator side. This aspect of operation impacts upon the design of the internal components of the gearbox. As mentioned above, planetary gear stages can carry greater torque due to power branching between planet gears and, therefore, are sometimes the preferred option at the rotor side. Otherwise, gear stages with identical ratios (assuming identical pitch, helix angle, and number of teeth) will have differing torque ratings due, primarily, to the width of the gear in the axial direction e narrow gears will carry less torque than wide gears. The increase in speed through the gear stages will impact upon the bearing selection. Gears may be manufactured as spur gears or helical gears. Spur gears are the simpler of the two gear types: the gear teeth run parallel to the gear shaft in the axial direction. Helical gears have gear teeth that run along the gear shaft at a particular angle, called the helix angle. The introduction of the helix angle allows for a smoother engagement of the gears. This is because the contact ratio e the average number of teeth in contact at any time e is typically greater for helical gears than for spur gears. However, due to the reaction of forces on the helix gears, the shaft and bearings must be designed to counteract greater axial thrust forces. Another gear configuration that is used to counteract the axial thrust is called herringbone, whereby two counteracting helices are used on a single shaft. The application of ISO/IEC 61400-4 requires that several other general gear-related standards be applied. Table 7.1 shows the typical modes of failures that can occur and the relevant standard that should be adhered to. ISO/IEC 61400-4 makes some specific recommendations related to gear rating factors. For example, the dynamic factor, KV, for WTGS gearboxes should be calculated using method B from ISO 6336-1 and be subject to a minimum value of 1.05 unless proven otherwise by measurement. Dynamic factor can significantly affect gear rating. The gear load-sharing factor or mesh load-sharing factor, Kg, applies specifically to gear stages that incorporate power branching, ie, planetary gear stages. Inaccuracies in the gears can cause deviations in the load splitting and the mesh load factor takes account of these deviations based on n, the number of planets gears. Table 7.2 provides data on the recommended mesh load factors. The gear mesh load distribution factor takes into account non-uniform load distribution over the gear face width on the surface stress (KHb) and on the tooth root stress (KFb). ISO 6336 gives procedures for calculating the face load factors and these


Table 7.1

149

Applicable gear standards for gear component design

Failure mode

Relevant details (as cited in ISO/IEC 61400-4)

Gear pitting

ISO 6336 series Minimum safety factor, SH ¼ 1.25

Bending

ISO 6336 series Minimum safety factor, SF ¼ 1.56 Life factors, ZNT and YNT, determined using 0.85 1010 cycles

Scuffing

ISO/TR 13989-1 DIN 3990-4 ANSI/AGMA 925-A02 þ ISO/TR 13989-2

Micropitting

ISO/TR15144-1

Static strength

ISO 6336 series Evaluated at extreme torque using static life factors YNT and ZNT Minimum safety factor for root bending, SF > 1.4 Minimum safety factor for surface durability, SH > 1.0

Subsurface-initiated fatigue

DNV classification note 41.2 (subclause 2.13)

Table 7.2

Mesh load factors based on number of planet gears

Number of planets, n

3

4

5

6

7

Mesh load factor, Kg

1.10

1.25

1.35

1.44

1.47

Adapted from ISO/IEC 61400e4. Wind turbines e part 4: design requirements for wind turbines gearboxes. 2012.

procedures are generally applicable for WTGS gearbox design. However, ISO/IEC 61400-4 gives some particular recommendations for calculating the face load factor for surface stress, KHb. Specifically, the gear mesh misalignment, fma, (given in millimetres) is calculated based on specific recommendations that differ from the procedures given in ISO 6336. The minimum recommended vale of KHb is 1.15.

7.4

WTGS gearbox design e production stage

There are many issues that can arise at the production stage that can affect the performance and reliability of the WTGS gearbox in service. In a broad sense, the production stage may be considered as component manufacture and assembly. A quality plan should be in place to ensure that no adverse effects occur due to production. Many processes are used in the manufacture of a single component: forming processes (casting and forging), subtractive processes (cutting and grinding), and surface

150

Offshore Wind Farms

treatment (polishing, heat treatment, etc.) to name a few. Finished components always have some deviation from the intended form. These deviations could be geometrical in nature due to the accuracy and precision limitations of the manufacturing processes. Variations also occur in the material’s physical properties due to chemical or microstructural variations in the material which, again, are related to the processing of the material. Materials may also suffer from defects within the material: flaws, voids, inclusions, etc. The designers must develop a robust design that can tolerate these minor variations. Manufacturing engineers must ensure that the components are supplied within the allowable tolerances as outlined within the specifications. These aspects should form part of the quality plan. To allow for specified variations in geometry, designers should use geometric dimensioning and tolerancing standards for all components [15]. Minimum material requirements are usually set out in accordance with a standard, for example, gear components must comply with the requirements of ISO 6336-5, steel used in bearings must meet ISO 683 requirements. Assembly processes can be permanent (for example, welding) or non-permanent (screwed fasteners, etc.). Most of the assembly processes used in gearbox manufacture are the non-permanent type. This allows for easier maintenance during the operating stage and easier disassembly at the retirement stage. Application of the correct torque to all bolted joints is essential to the assembly. In some instances, as with tapered roller bearings, precompression of the bearing is also important. This preload is achieved by applying torque to a tightening nut which compresses the bearing set. All of these torques must be applied appropriately. In the housing joints, it is recommended that split plane housing joints should have positive locating devices such as dowel pins to aid with assembly. In the case of planetary gear sets it is recommended that the ring gear joint should be capable of carrying the maximum operating load by friction only (with some safety margin). If the friction generated by bolt torque alone is insufficient, then the designer should consider the application of solid pins to carry the loads at the flange. In the case of pins, the contribution of friction should be ignored from the calculations. This feature of ISO/IEC 61400-4 represents specific, conservative design requirements that have been determined through the experience of iterative design cycles. In some cases accuracy may be lost during the assembly processes, a phenomenon known as tolerance stack up. Designers need to be aware of tolerance stack up during design of an assembly.

7.4.1

Gear manufacture and inspection

ISO/IEC 61400-4 has specific requirements related to gear manufacture. The method and processing of all gear elements should be specified as a matter of course. Grinding notches should be avoided in the gear cutting process; however, if gear notches do occur during the manufacture process, then FEA methods or the YSG factor from ISO 6336-3 should be used to determine the reduction in tooth bending stress. Rejection criteria are clearly outlined in ISO/IEC 61400-4. Otherwise, gear accuracy should be specified in accordance with ISO 1328-1. The standard ISO 1328-1 sets out 11 grades of gear accuracy. The tolerance values increase with each increase in grade level. ISO/IEC 61400-4 gives maximum accuracy grades


151

as level 6 for external gears and level 7 for internal gears (with some specific allowances to grade 8 for nitrided internal gears). Improved tolerances will ensure smoother performance but the cost factors associated with increased accuracy should be considered. It should be noted that the accuracy grades apply to assembled gears e if gears loose there accuracy during assembly, then grinding after assembly should be considered. The surface finish required on gear components is specified as Ra ¼ 0.8 mm for external gears and Ra ¼ 1.6 mm for internal gears. Improving the surface finish reduces the risk of micropitting on the gear. It should be noted that shot-peening of gear flanks is not permitted as a final operation; this is because the surface finish may be compromised. ISO/IEC 61400-4 makes specific recommendations for surface temper after grinding. In particular, a 100% sampling plan is recommended. This requirement is provided to ensure that all gears have the appropriate surface qualities, such is the importance of surface finish. Poor gear tooth surface finish can cause micropitting. Through experience, the following surface roughness, Ra, values are recommended: Ra < 0.7 mm for high-speed and intermediate pinion and gears; Ra < 0.6 mm for low-speed pinion and gear; and Ra < 0.5 mm for low-speed sun and planet. It is important that the gear surface roughness measurement compensates for the involute form of the gear tooth. A skid style stylus pick-up is often used for gears. The stylus and the skid are situated side-by-side for improved tracking of the involute form. The skid radius on a gear tooth pick-up is around 0.8 mm and is typically smaller than that found on a general-purpose pick-up. Surface crack detection is also advised whereby magnetic particle, fluorescent magnetic particle penetrant or dye penetrant inspection methods may be used as agreed by the supplier and customer (ISO 6336-5 should be referred to in this instance).

7.5

Drivetrain dynamic analysis

ISO/IEC 61400-4 recommends, as a minimum, that drivetrain analysis be performed to verify the WTGS aero-elastic response. An example of a WTGS drivetrain dynamic (DTD) model, which includes the gearbox subsystem, is discussed next.

7.5.1

Variable loading

The torque level in a wind turbine gearbox is variable. This fluctuation will normally vary between zero and rated torque as the wind speed varies. Excursions above rated torque are possible on a fixed-speed pitch-regulated machine. This can be attributed to the slow pitch response. The torque time histories will be subject to dynamic magnification and can in fact excite the drivetrain leading to resonances. Transient events such as braking can induce infrequent large-magnitude torques but of short duration. This effect of producing large-amplitude torques can be mitigated if the brake is fitted to the low-speed shaft. The normal practice to calculate the loadeduration curves is by

152

Offshore Wind Farms

combining the power curve with the distribution of instantaneous wind speeds. The computation of distribution of instantaneous wind speeds is carried out by superposing the turbulent variations around each mean speed on the Weibull distribution of hourly means.

7.5.2

Drivetrain dynamics

Rotational sampling by the blades is the source of torque fluctuations in wind turbines. Since this phenomenon is caused by the effect of rotating blades, the frequencies generated are at blade-passing frequency and multiples thereof. Hence, all wind turbines experience aerodynamic torque fluctuations at these frequencies. These fluctuating torques interact with the dynamics of the drivetrain, with the possibility of dynamic amplification and hence modifying the torque transmitted. If a fixed-speed wind turbine with an induction generator is considered, the resulting drivetrain torque variability can be assessed by dynamic analysis of a suitable drivetrain model. Such a model consists of the following elements connected in series: • • • • •

a body with rotational inertia and damping (representing the turbine rotor); a torsional spring (representing the stiffness of the gear box and the high-speed shaft coupling); a body with rotational inertia (representing the inertia of the mainshaft, rotor lock disk and generator rotor); a torsional damper (modelling the resistance produced by slip-on induction generator); the electrical grid.

These inertias, stiffness and damping must all be referred to the same shaft. For carrying out a detailed stress analysis of the drivetrain an accurate multibody model of the system is required. A detailed model provides information on deflection, stresses generated and damage induced. However, if dynamics related to the modes of vibration and development of control algorithm is the focus of attention then simplified models may be used instead. Such a simplified dynamic model is described in Section 7.5.3 which is capable of capturing the natural frequencies and modes of vibration and also provides information on the states related to torsional oscillations.

7.5.3

Two-mass drivetrain shaft model

Drivetrain components are fundamentally responsible for transmitting the generated aerodynamic torque generated on the rotor from the rotor hub to the wind turbine generator [16]. The drive torque is filtered by the drivetrain and is converted into mechanical torque, which in turn drives the generator shaft. In this section, the drivetrain is modelled by a simplified two-mass mechanical system as shown in Fig. 7.10. In comparison with higher-order multimass drivetrain models, the two-mass system has been found to provide enough accuracy for the transient stability analysis of wind turbine generation systems [17]. Muyeen et al. [18] have carried out a comparative study among different types of drivetrain modelling. Their study concluded that higher-order models (six-mass and three-mass) of drivetrains can be reduced to a


Tls

Ta

153

Tem

Ths

kls

Ωg

Ωr

Jg c ls

ηg

Jr

Figure 7.10 Torsional vibration model of the WTGS drivetrain. After Basu B, Staino A, Basu M. Role of flexible alternating current transmission systems devices in mitigating grid fault-induced vibration of wind turbines. Wind Energy 2014;17: 1017e33. http://dx.doi.org/10.1002/we.1616. Copyright © 2013 John Wiley & Sons, Ltd.

two-mass model without significant loss of accuracy with respect to response exhibited under a network disturbance. In the simplified model considered the following elements are included to represent the drivetrain dynamics: • • • • •

a mass with rotational inertia Jr (representing the turbine rotor, ie, the blades); a torsional spring and a torsional damper (representing the low-speed shaft); an ideal gear box with speed-up ratio hg ; a rigid high-speed shaft; a mass with rotational inertia Jg (representing the generator rotor).

A linear torsional spring with stiffness coefficient kls is used to represent the low-speed shaft and a torsional damper of damping coefficient cls represents the damping of the drivetrain. The equation governing the dynamics of the blades (turbine rotor) is given by Ta ðtÞ Tls ðtÞ cf Ur ðtÞ U_ r ðtÞ ¼ Jr

[7.12]

In Eq. [7.12], Ta and Tls are the aerodynamic torque generated by the wind and the torque at the low-speed shaft, respectively. A small amount of damping, represented by cf , may be also added in order to model friction dissipation effects. The dynamics of the shaft is represented by the following equation: Tls ðtÞ ¼ kls ðqr ðtÞ qls ðtÞÞ þ cls ðUr ðtÞ Uls ðtÞÞ

[7.13]

where qr is the rotor angular position; qls and Uls are the low-speed shaft angular deviation and speed, respectively. The transmission speed-up ratio hg is given by (on the assumption of an ideal gearbox): hg ¼

Tls ðtÞ Ug ðtÞ qg ðtÞ ¼ ¼ Ths ðtÞ Uls ðtÞ qls ðtÞ

[7.14]

154

Offshore Wind Farms

In Eq. [7.14], qg and Ug denote the generator shaft angular position and speed respectively, and Ths is the drive torque provided as an input to the generator. Finally, the following differential equation represents the dynamics of the mechanical part of the generator Ths ðtÞ Tem ðtÞ cg Ug ðtÞ U_ g ðtÞ ¼ Jg

[7.15]

where Tem is the electromagnetic torque developed and cg is the associated damping on the generator side. Eqs [7.12]e[7.15] describe the dynamics of the two-mass drivetrain system illustrated in Fig. 7.10. For the purpose of studying the induced vibration, the relevant degree-of-freedom describing the drivetrain dynamics associated with the torsional mode of the flexible shaft system has been considered. The equation governing the drivetrain torsional oscillation is _ ¼ Ur ðtÞ Ug ðtÞ qðtÞ hg

[7.16]

where q ¼ qr qls represents the angular difference between the two ends of the low-speed shaft. In addition to aerodynamic fluctuations, occurrence of faults in the electrical subsystem of the wind turbine may also induce torsional drivetrain oscillations [19].

7.5.4

Coupled electromechanical interaction

The importance of coupled electromechanical effects has been highlighted in the available literature [20e24]. These studies stress the necessity of a detailed mechanical model, to understand the mechanical response of wind turbines to an electrical disturbance. In addition, a detailed electrical model is also essential as the fault generating the disturbance is an electrical phenomenon. The detailed mechanical and electrical models are then needed to be coupled together. In order to do so, it is required to identify the parameters which are be to communicated or exchanged between the two detailed models (electrical and mechanical). The dynamics of the flexible rotor blades are governed by the aerodynamic load. The corresponding equations governing the variable-speed rotor dynamics are a function of the blade rotor speed and accelerations. The generated aerodynamic torque Ta ðtÞ forms an input to the drivetrain. On the electrical side, the dynamics of the generator is governed by Eq. [7.15]. For this equation, the torque from the drivetrain (high-speed shaft) serves as the input. Subsequently, the generator speed and accelerations, and the electrical torque are the output. Hence, the mechanical drivetrain and gear system form the connecting link between the rotor blades and the electrical generator. This is illustrated in Fig. 7.11. The motion of the two-mass drivetrain model with


155

Wind turbine system

Wind

Wind turbine structural and aero dynamics Tα(t)

Drivetrain and gearbox

Ths(t) FSIG

Ω g(t)

Ω˙ r(t)

Ω˙ g(t)

FACTS Device 11 kV:132 kV

Ω r(t)

690 V:11 kV

– 2 MW

PFC

R Fault jX

Figure 7.11 An electromechanically coupled system. After Basu B, Staino A, Basu M. Role of flexible alternating current transmission systems devices in mitigating grid fault-induced vibration of wind turbines. Wind Energy 2014;17: 1017e33. http://dx.doi.org/10.1002/we.1616. Copyright © 2013 John Wiley & Sons, Ltd.

the shaft is described by Eqs [7.12]e[7.14] and Eq. [7.16]. These equations lead to the following second-order linear differential equation which represents the torsional oscillation of the drivetrain with the aerodynamic torque, and the generator speed and accelerations as the input. Jr €q þ ðcls þ cf Þq_ þ kls q ¼ Ta cf

U_ g Ug Jr hg hg

[7.17]

The solution of the second-order linear differential equation provides the required quantities of interest, such as drivetrain torsional angular displacement and the angular velocity with prescribed initial conditions. However, for solving Eq. [7.17], appropriate initial conditions are required. Initial conditions that can be used for solving the equation without any loss of generality are zero values for drivetrain torsional states (ie, no initial twist) and rated generator speed with no acceleration corresponding to the rated rotor blade speed. Using Eqs [7.13] and [7.14], the high-speed shaft torque can then be computed. Also, blade rotor speed and accelerations can be computed using Eq. [7.12]. Thus, the outputs generated are the high-speed shaft torque, and the blade rotor speed and accelerations. The high-speed shaft torque is fed into the generator to update the generator dynamics, while the blade rotor speed and

156

Offshore Wind Farms

accelerations form inputs to the rotor blade dynamics equations to update the system matrices and to compute the variable-speed rotor blade dynamics. The scheme of the computation described with the outlined electromechanical coupling incorporates the effect of (1) flexible edgewise blade dynamics, (2) variable rotor blade speed, (3) two-mass drivetrain with shaft model and (4) generator dynamics.

7.5.5

Effect of variable loading on fatigue design of gear teeth

Fatigue design of gear teeth is guided by two factors. The contact stresses generated on the flanks and the bending stresses generated at the root, must both be within acceptable limits. The compression stress (for Hertzian contact) between a pair of spur gear teeth in contact at the pitch point (ie, at the point on the line joining the gear centres) is given by sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Ft E uþ1 1 sc ¼ 2 bd1 pð1 n Þ u sin a cos a

[7.18]

where Ft is the force between gear teeth at right angles to line joining the gear centres; b is the gear face width; d1 is the pinion pitch diameter of the pinion of the driving or input gear; u is the gear ratio (greater than unity); a is the pressure angle e that is, that angle at which the force acts between the gears e usually 20 degree to 25 degree, E is the elastic modulus; n is the Poisson’s ratio. The maximum bending stress at the root of gear teeth is given by sB ¼

6Ft h KS b t2

[7.19]

where h is the maximum height of single-tooth contact above the critical root section; t is the tooth thickness at the critical root section; KS is the root stress concentration factor. For the case of the gears operating at rated torque it is sufficient to show that the resultant bending stress multiplied by a suitable safety factor is less than the endurance limit stress multiplied by a number of factors (such as life factor and a number of stress modification factors). A similar procedure also has to be followed for the contact stress. For design against fatigue or calculation of fatigue damage the predicted turbine load spectrum should be used. This should also include dynamic effects. For the fatigue design calculation it is necessary to compute the design equivalent torque at the endurance limit. This computation is normally carried out with the aid of Miner’s rule for fatigue damage and calculating the infinite life torque for which the design torque spectrum yields a damage index of unity in conjunction with the prescribed SeN (load vs. number of cycle) curve for the material concerned. In this scenario, the life factor can be set to unity as it has been indirectly accounted for in the infinite life torque calculation. Codes (such as BS 436) prescribe


157

specimen torqueeendurance curves which can be used for gear tooth design. The design infinite life torque can be calculated from the load-duration spectrum according to the formula: " TN ¼

X N i i

NN

#1=m Tim

[7.20]

where Ni is the number of cycles of torque of magnitude Ti. Torques of magnitude less than TN are not considered in the computation. The number of cycles at the lower knee of the torqueeendurance curves NN is 3 106 for tooth bending. However, this is generally higher for contact stress and varies according to the type of material used. The index m, for the torqueeendurance curve-related contact stress, is half that of the contact stresseendurance curve as torque is proportional to the square of the contact stress. More details on the fatigue design of gear teeth, bearings and shafts are available in Burton et al. [16].

7.6

Conclusions

The role of the gearbox in WTGS has been highlighted. A particular focus was given to the design of wind turbine gearboxes. A generic life cycle was presented, starting at the concept stage, going through the development stage, the production stage, the operation and maintenance stage and ending with the retirement stage. As it is the most relevant standard in the area, ISO/IEC 61400-4, Design requirements for wind turbine gearboxes, was reviewed in some detail. In particular, the elements outlined in ISO/ IEC 61400-4 were categorised according to the generic life cycle. A focus was provided on the concept stage, the development stage (detailed design) and the production stage. A drivetrain dynamic analysis has been outlined. This type of analysis is a minimum requirement contained within ISO/IEC 61400-4.

References [1] Manwell JF, McGowan JG, Rogers AL. Wind energy explained: theory, design and application. 2nd ed. Wiley; 2009. [2] ISO/IEC 61400e4. Wind turbines e part 4: design requirements for wind turbines gearboxes. 2012. [3] Mathew S. Wind energy: fundamentals, resource analysis and economics. Heidelberg: Springer, Berlin; 2006. [4] García Marquez FP, Tobias AM, Pinar Pérez JM, Papaelias M. Condition monitoring of wind turbines: techniques and methods. Renewable Energy 2012;46:169e78. [5] Milborrow D. Operation and maintenance costs compared and revealed. Wind Stats 2006; 19(3):3. [6] Vachon W. Long-term O&M costs of wind turbines based on failure rates and repair. In: WINDPOWER, American Wind Energy Association annual conference; 2002. p. 2e5.

158

Offshore Wind Farms

[7] Tavner P, Spinato F, van Bussel GJW, Koutoulakos E. Reliability of Different Wind Turbine Concepts with Relevance to Offshore Application. In: European Wind Energy Conference, March 31 e April 3, Brussels: Belgium; 2008. [8] Van Bussel GJW, Boussion C, Hofemann C. A possible relation between wind conditions, advanced control and early gearbox failures in offshore wind turbines. Procedia CIRP 2013;11:301e4. [9] Sheng SS, editor. Wind turbine condition monitoring. Wind Energy May 2014;17(5): 671e2. [10] Hahn B, Durstewitz M, Rohrig K. Reliability of wind turbineseexperience of 15 years with 1500WTs. In: Proceedings of the Euromech colloquium wind energy; 2007. p. 1e4. [11] Igba J, Alemzadeh K, Durugbo C, Henningsen K. Performance assessment of wind turbine gearboxes using in-service data: current approaches and future trends. Renewable Sustainable Energy Rev 2015;50:144e59. [12] Stamatis DH. Failure mode and effect analysis: FMEA from Theory to Execution. ASQ Quality Press; 2003. [13] Arabian-Hoseynabadi H, Oraee H, Tavner PJ. Failure Modes and Effects Analysis (FMEA) for wind turbines. Int J Electr Power Energy Syst September 2010;32(7):817e24. [14] Dooner DB. Kinematic geometry of gearing. 2nd ed. Wiley; 2012. [15] Cogorno GR. Geometric dimensioning and tolerancing for mechanical design. 2nd ed. McGraw-Hill Education; 2011. [16] Burton T, Sharpe D, Jenkins N, Bossanyi E. Component design. In: Wind energy handbook. Chichester, UK: John Wiley & Sons; 2001. p. 424e38. [17] Boukhezzar B, Siguerdidjane H. Nonlinear control of a variable-speed wind turbine using a two-mass model. IEEE Trans Energy Convers 2011;26:149e62. http://dx.doi.org/ 10.1109/TEC.2010.2090155. [18] Muyeen S, Tamura J, Murata T. Wind turbine modeling. In: Stability augmentation of a grid-connected wind farm, Green Energy and Technology. London, UK: Springer London; 2009. p. 23e65. [19] Salman S, Teo A. Windmill modeling consideration and factors influencing the stability of a grid-connected wind power-based embedded generator. IEEE Trans Power Syst 2003;18: 793e802. http://dx.doi.org/10.1109/TPWRS.2003.811180. [20] Bossanyi EA. The design of closed loop controllers for wind turbines. Wind Energy 2000; 3:149e63. http://dx.doi.org/10.1002/we.34. [21] Fadaeinedjad R, Moschopoulos G, Moallem M. Investigation of voltage sag impact on wind turbine tower vibrations. Wind Energy 2008;11:351e75. http://dx.doi.org/10.1002/ we.266. [22] Bossanyi EA. Wind turbine control for load reduction. Wind Energy 2003;6:229e44. http://dx.doi.org/10.1002/we.95. [23] Jauch C. Transient and dynamic control of a variable speed wind turbine with synchronous generator. Wind Energy 2007;10:247e69. http://dx.doi.org/10.1002/we.220. [24] Ramtharan G, Jenkins N, Anaya-Lara O, Bossanyi E. Influence of rotor structural dynamics representations on the electrical transient performance of FSIG and DFIG wind turbines. Wind Energy 2007;10:293e301. http://dx.doi.org/10.1002/we.221. [25] Basu B, Staino A, Basu M. Role of flexible alternating current transmission systems devices in mitigating grid fault-induced vibration of wind turbines. Wind Energy 2014;17: 1017e33. http://dx.doi.org/10.1002/we.1616.

Design of generators for offshore wind turbines

8

A. McDonald, J. Carroll University of Strathclyde, Glasgow, United Kingdom

8.1

Introduction: key issues in generator design

The electrical generator plays a critical role in the wind turbine because of its electromechanical nature. The generator acts as an interface between the mechanical subsystems e such as the rotor and drive train upstream of the generator e and the electrical subsystems e such as the power converter and the grid. Despite being a relatively modest portion of the offshore wind farm cost, the electrical generator can have a significant influence on its cost of energy as the type and design of the generator influence turbine efficiency, operation and maintenance (O&M) costs and availability [1]. After discussing the role of the generator in wind turbines, and how the application of offshore wind turbines is different to that of conventional generation plants, this chapter will give a brief overview of electrical machines from a conceptual view point (Section 8.2), describe some of the practical engineering aspects of the generator (Section 8.3), before going on to describe the types of electrical generators that have been chosen for real offshore wind turbines and explaining the motivation for these choices (Section 8.4). At the end of the chapter, the authors will go on to survey some of the advanced types of generators that may be popular in the coming years (Section 8.5).

8.1.1

What is a generator and what should it do?

An electric generator is a device for generating electricity from a mechanical energy input. In the context of offshore wind turbines the mechanical energy is obtained by capturing the energy in the wind with a turbine rotor and e in some turbines e converting that high torque and low speed into lower torque and higher speed using a gearbox (or other torqueespeed converter). In other cases, the mechanical energy may come from many different sources. In more conventional power plants the mechanical energy input comes from prime movers powered by combustion engines, steam turbines, hydro turbines and gas turbines. Based on Faraday’s law of electromagnetic induction, electrical energy is generated through the mechanical process of passing a conductor through a magnetic field. This process creates a potential difference (voltage) between the ends of the conductor. From Ohm’s law, it is known that an electric current will flow when the ends of this


160

Offshore Wind Farms

conductor are connected to an electrical circuit with a resistance. The flow of current will interact with the magnetic field and act to produce a force resisting the motion of the conductor. In order to continue moving the conductor at the same speed, a mechanical force of the same magnitude must be provided and applied in the direction of movement. The result is that mechanical power ( force speed) is converted into electrical power (voltage current). An electrical generator performs this process of converting mechanical power into electrical power writ large. In doing so the generator produces torque (the result of force acting at some radius from the axis of the rotating machine) opposing the mechanical torque from the wind turbine rotor and drive train. From the electrical point of view, the generator is a voltage source (with some impedance) which depends on the type and design of the generator and on the operating conditions of the wind turbine and the generator. In most electrical generators, it is convenient to induce electromotive force (emf) (and hence current) in stationary conductors, and hence they are located in a part of the machine which does not rotate (called the stator, as shown in Fig. 8.2(a)). The magnetic field is moved instead, and its source is mounted on a rotating part of the generator (called the rotor). The rotor and stator are separated by a clearance termed the ‘air gap’. Fig. 8.2(b) shows a generator and its rotor and stator. In a synchronous generator the rotor magnetic field is produced either by using permanent magnets or else direct current is used to excite an electromagnet. As the excited rotor rotates it produces a rotating magnetic field in the air gap between the rotor and the stator. The stator contains a number of conductors that cut the magnetic field that is rotating in the air gap, in turn generating a current and hence electrical power output. Section 8.2 provides a more detailed explanation of the operation of the synchronous generator as well as detailing the operation of asynchronous generators.

Motion of conductor

Co

u nd

S

ct

or

N Magnetic field lines

I

u nd

ce

d

r cu

re

nt

Figure 8.1 A conductor moving through a magnetic field.


(a)

161

(b)

Stator

Airgap

Rotor

N

Rotor

S Stator

Figure 8.2 (a) Simplified cross-section of a generator showing rotor, air gap and stator. (b) Rotor and stator in a synchronous generator.

8.1.2

What makes a good generator?

Depending on design requirements an offshore wind turbine generator will be considered good if it is: • • • • •

highly efficient, both at rated power and across the power curve, reliable, with a low failure rate and a short mean time to repair, cost-effective, both to purchase and to operate and maintain, lightweight, compact and easy to install, able to operate over a large speed and torque range.

As with many engineering systems, a number of these qualities complement one another and in some cases run counter to one another. For example, a more reliable generator may have lower maintenance costs but might be less efficient and more expensive to build. As a result of this, a balance must be struck to ensure a generator with the required qualities is chosen for a particular purpose. One approach for the turbine generator designer to balance these qualities and gauge the relative importance of them is to consider how the generator design affects the levelised cost of energy [1]. This approach will be used in Section 8.4. Fig. 8.3 shows some of the large wind turbines which are currently in development or just starting to be manufactured. Generally, a good generator should be efficient, with as few losses as possible over a wide speed and torque range. Permanent magnet synchronous generators (PMGs) are usually better in this area than electrically excited machines as no current is needed to produce the rotor magnetic field and hence there are no copper losses on the generator rotor. Generators should also be reliablem with as few failure modes as possible and PMGs and squirrel cage induction machines show greater reliability due to their lack of slip rings and brushes, which are prone to failure

162

Offshore Wind Farms

Full power converter Part power converter

RE Power 6 MW No power converter

Speed and torque conversion

Sinovel 6 MW

Magnetic

Bard 5 MW

Hydraulic

Mitsubishi 7 MW Samsung 7 MW

High speed gearbox

AREVA 5 MW Gamesa 5 MW MHI Vestas 6 MW

Medium speed gearbox

Siemens 7 MW Alstom 6 MW

Enercon 7.5 MW

XEMC 5 MW

Direct drive

nc

hro

tio

ure rat pe ting em uc h-t ond Hig perc su us no PM hro nc sy us no

Sy

uc

fed ly n ub Do uctio ind n

Ind

Electrical machine type

Figure 8.3 Drivetrain choice for some large wind turbines, specified by speed and torque conversion, generator type and rating of power converter [1].

and need regular servicing. However, this improved generator reliability has a consequence for the reliability of the power converter: there is a higher failure rate for the fully rated converters required for variable speed operation of PMGs and squirrel cage induction machines than for those used in doubly fed induction generators [2]. Onshore, failures to power converters tend not to lead to long downtime as they are relatively quick to repair. Offshore e where access to the turbine is more challenging and expensive e the consequence of even small failures can be costly. Manufacturing cost effectiveness and cost planning is also an important aspect for generator and wind turbine manufacturers. In this area PMGs are not as favourable as the electrically excited machines because the rare earth materials required for the permanent magnets are more expensive and cost volatile [3].

8.1.3

What are the differences between generators in conventional power plants and those in wind turbines?

Wind turbine generators and generators from conventional power plants differ in a number of ways. The main reason for these differences is the source of the mechanical input to the generator. Generally speaking, the mechanical power of the prime mover is


163

at a higher rotational speed for conventional power plants than in wind turbines. This means that if the generator is directly connected to the prime mover (‘direct drive’) then the generator in a conventional power plant can produce more power for the same generator torque rating. The corollary of this is that to produce the same amount of power, the direct drive wind turbine generator must have a higher torque and hence will be more expensive than would be the case in a conventional power plant. As conventional power plants tend to make use of a more energy-dense medium, the turbines tend to have power ratings of 102e103 MW, whereas offshore wind turbines have power ratings in the range 2e8 MW, with some design work at 10e20 MW. Each wind turbine generator will have a lower power rating and so will be cheaper, however the total cost of electrical generators for a 1000-MW wind farm will be higher than that for a 1000-MW conventional power plant as there are fewer generator units. The generators in conventional plants tend to operate at a fixed speed and the mechanical input torque can be controlled relatively easily. Conventional plants tend to use synchronous generators which are usually directly connected to the grid. One result is that the rotational inertia of the turbine and generator naturally contributes to the stability of the electrical frequency of the electrical network. Because the excitation current can be varied, it is possible to control the power factor of the generator and provide or absorb reactive power as necessary. Although some earlier wind turbines were directly grid-connected, they used induction machines. This meant that the wind turbines were fixed speed rather than variable speed (the latter being preferable from a mechanical loading point of view) and that they consumed reactive power from the grid, this often having to be compensated for. Another difference between conventional plants and the wind turbine is the location of the electrical generator. In conventional plants the generator is placed in well-controlled conditions, the generator can be easily accessed and the total generator mass is relatively unimportant. In wind turbines, the generator is almost always placed in a nacelle on top of the wind turbine tower. Obtaining the required amount of power for driving the generator requires large wind turbine rotor blades. Due to this large blade requirement, wind turbine generators are usually located on towers that can be more than 100 m above sea level. The fact that these generators are elevated to such heights means that mass can become a constraint as additional tonnes of material can affect the tower, foundation and installation costs.

8.1.4

What are the differences between onshore and offshore wind turbine generators and the challenges for offshore wind turbine generators?

Generators also differ in a number of ways when used in onshore and offshore wind applications. The main reasons for these differences are the accessibility problems related to being offshore, larger lifting equipment costs offshore and the ability to have larger blades offshore. Offshore wind has a higher cost of energy than onshore wind, and one of the contributory factors to this is the higher operations and maintenance costs. Newer

164

Offshore Wind Farms

offshore wind farms also tend to be placed further offshore than the first wind farms of the 1990s and 2000s. Moving further offshore often means sites have higher mean wind speeds and are subject to higher mean wave heights. As maintenance and repair vessels have wind speed and wave height operational limits, sites further offshore are less accessible than near shore, and near-shore sites are less accessible than onshore. These offshore access problems are leading wind turbine manufacturers to develop different types of generators for the offshore market compared to the onshore market. Manufacturers are focusing on making offshore generators more reliable, introducing redundancy and extending time between services to ensure offshore generators do not have to be visited as often as onshore generators [4]. Another aspect influencing the design of offshore generators is the cost of replacing them when they fail. Cranes can be used onshore, but heavy lift vessels e usually rented on daily rates of over £100,000 per day e are required offshore [5]. As a means of overcoming this, manufacturers are aiming to make their offshore generators modular, to allow them to be repaired and replaced without, or with reduced, use of heavy lift vessels. Generally, offshore wind turbines have larger rotor diameters than those onshore. Partly this is to mitigate the expense of offshore foundations and cabling by increasing the power rating of each turbine and increasing each turbine’s annual energy production. Reducing the rotor power density also leads to larger swept areas e this helps improve the turbine’s capacity factors and utilisation of the transmission cabling. One advantage of offshore wind turbines is that there tend to be fewer objections from the local community than for onshore wind turbines. These objections can be sensitive to hub height, blade length and blade noise. This relaxation tends to lead to longer blades offshore as can be seen in Fig. 8.4. This larger blade size allows for greater mechanical torque to be produced offshore, resulting in generators offshore having a higher rated power. This is evident in the portfolio of major manufacturers who offer 7- and 8-MW turbines with rotor diameters up to 154 m (Siemens D7) and 164 m (Vestas V164) offshore but the largest onshore turbine is rated at 7.5 MW and has a rotor diameter of 126 m (Enercon E126). Onshore turbines. Smaller rotor size 126 m

Offshore turbines. Rotor size up to 164 m

Figure 8.4 Turbine size onshore versus turbine size offshore.


8.2 8.2.1

165

Electrical generators: types and principles of operation Types of electrical machines

Generators belong to a family of devices known as electrical machines, which share the same principles of operation. There are three types of electrical machines and wind turbines can contain all of these: • • •

Generators, rotating electrical machines which convert mechanical power into electrical power; Motors, rotating electrical machines which convert electrical power into mechanical power. These are often used in yaw drives and pitch actuators, and Transformers, stationary devices which convert from one voltage level to another voltage level. These are used to step up the voltage level from the wind turbine to medium voltage used in the wind farm electrical collection system.

Rotating electrical machines can perform the roles of either generator or motor depending on the way they are controlled, the electrical connection and the nature of the mechanical interface. Fig. 8.5 shows this. There are two major categories of rotating electrical machines, those that use/produce alternating currents (AC) and those that use/produce direct current (DC). In conventional power systems e which are based on AC motors and generators e the majority of generators are synchronous machines and most motors are asynchronous machines. In wind turbines both synchronous and asynchronous machines (also known as induction machines) are used as generators. The machine described in Section 8.1.1 is a primitive synchronous machine. In the case of the synchronous generator the rotor rotates at the same speed as the electromagnetic flux waveform in the air gap of the machine. For the wind turbine rotor to have a variable speed, the electrical frequency of the machine must also be variable in nature and this means that the generator stator must be connected to a power converter to decouple it from the grid frequency. The synchronous machine is described in more detail in Section 8.2.3.

Mechanical system T, Ω

Rotating electrical machine

Electrical system V, I Generator

Energy flow Motor

Figure 8.5 Electromechanical energy conversion in rotating electrical machines. Adapted from P.C. Sen, Principles of Electric Machines and Power Electronics, John Wiley & Sons, 2007.

166

Offshore Wind Farms

Although similar in construction to the synchronous machine, induction machines have a different principle of operation. In these machines, an alternating current is present on the rotor windings, giving rise to a rotor magnetic field which rotates at a speed slower or faster than the rotor. This field is either induced from the electromagnetic field on the stator (ie, squirrel cage induction generators) or else set up by a current from a static power converter connected to the rotor windings through brush gear and slip rings (ie, doubly fed induction generators). The principle of operation of these induction machines is explained in the Section 8.2.4.

8.2.2

Principles of electromagnetic conversion

There are two basic electromagnetic phenomena at play in rotating electrical machines: •

•

When a conductor moves relative to magnetic field, a potential difference is induced in the conductor. This was described in Section 8.1.1 and is shown in Fig. 8.1. If the conductor is connected to a closed circuit a current will flow in the direction given by Fleming’s right-hand rule (hold the right hand so that the thumb, first finger and second finger are all perpendicular to one another. The thumb is in the direction of motion; the first finger is in the direction of the field; the second finger is in the direction of the induced current). When a current-carrying conductor is present in a magnetic field it will experience a mechanical force. This is shown in Fig. 8.6. The conductor will experience a force in the direction given by Fleming’s left-hand rule (hold the left hand so that the thumb, first finger and second finger are all perpendicular to one another. The thumb is in the direction of the force; the first finger is in the direction of the field; the second finger is in the direction of the induced current).

To find the induced voltage, e, (units V), in the conductor in Fig. 8.1, one can use Eq. [8.1], e ¼ Blv

[8.1]

Co

u nd

ct

or

Magnetic field lines

S

N

Force on conductor Cu

rre

nt

Figure 8.6 Force on a current-carrying conductor in a magnetic field.


167

where B is the flux density (in T), l is the length of the conductor in the magnetic field (units of m) and v is the linear speed (in m/s). In the case of electrical generators, the linear velocity is the product of a rotational speed of a magnetic field at a radius. Good generators are able to produce significant air gap flux densities. Total induced voltage can be found by summing all the individual voltages from all of the conductors which are connected in series in a phase. Induced voltage is sometimes described as the emf. In rotating electrical machines, one often thinks of coils (a combination of pairs of antiparallel conductors distributed around the stator of the machine so that they experience flux density of the same magnitude but opposing sign). For a coil, Faraday’s law states that the induced voltage in a coil is proportional to the negative rate of change of magnetic flux. This is given in Eq. [8.2], e ¼ N

dF dl ¼ dt dt

[8.2]

where N is the number of turns in a coil, F is the magnetic flux (units Wb) and l is the flux linkage (units Wb-turns). Flux density, B, is the magnetic flux per unit area (B ¼ F/A) and flux linkage is the product of the magnetic flux passing perpendicularly through the coil and the area of the coil. To find the force, f, (units N), applied to the current-carrying conductor in Fig. 8.6, one can use Eq. [8.3], f ¼ Bli

[8.3]

where I is the current (in units of A). Going from the forces on individual conductors to the torque developed in an electrical machine requires the individual forces to be summed, and the result to be multiplied by the radius.

8.2.3

Synchronous machines

In synchronous machines, the rotor carries the field winding with the armature winding mounted on the stator. Fig. 8.7 show the rotors of three different types of synchronous machines: a cylindrical rotor, a salient pole rotor and a rotor with surface-mounted permanents. Cylindrical rotors tend to be used in higher-speed machines in conventional power plants. Salient pole rotors are often used in larger, lower-speed machines in applications such as hydro generators and some direct-drive wind turbines. The first two types require a DC supply to the rotor which can be provided through brush gear and slip rings or through brushless excitation (basically another electrical machine or set of machines mounted on the shaft). Most offshore wind turbines that have a synchronous generator use permanent magnet excitation on the rotor.

8.2.3.1

Principle of operation at no load

Fig. 8.8 shows a rotor with two poles (one pole pair). Each pole is a piece of steel with a coil wrapped around it with N number of turns. The coil is excited with a field current, If. The product, NIf is the magnetomotive force (MMF). The strength of the flux

168

Offshore Wind Farms

(a)

(b) N

N

S

S

(c) N

S

Figure 8.7 Different rotors used in synchronous generators: (a) cylindrical, (b) salient pole, (c) surface-mounted permanent magnet.

density which is produced in the air gap depends on this MMF as well as the length of the air gap, the choice of materials used in the generator rotor and stator and the dimensions of the parts of the machine which conduct the magnetic flux. In the case of this DC excited synchronous machine, the flux density can be controlled by varying If; for PM machines the MMF is fixed for a given design. In the simple machine shown in Fig. 8.8, there are three phases: a, b and c. These are shown by the coils aa0 , bb0 and cc0 (where a0 means that the conductors are in the opposite direction to the conductors a). These coils are distributed by 120 degree or 2p=3 radians relative to one another. As the rotor rotates, flux linking each phase depends on the rotor angle, q. For a given rotational speed, u, this angle is given by q ¼ ut. The starting position of the rotor in Fig. 8.8 means that coil aa0 sees 0 flux linking the coil. Looking to Fig. 8.8(b), it can be seen that this flux linkage increases to a maximum when the rotor turns by 90 degree or p=2 radians; as the rotor angle q increases further, the flux linkage drops until it is 0 when the rotor angle is 180 degree or p radians. Further rotation means that the flux linkage is negative, reaching a minimum at 270 degree or 3p=2 radians and then returning to 0 when the rotor angle is 360 degree or 2p radians. It can be shown through relatively simple trigonometry that the flux linkage


169

λ aaʹ

λ bbʹ

λ ccʹ

a

θ = ωt cʹ

b

N

bʹ

S

c

e aaʹ

e bbʹ

eccʹ

aʹ

θ = ωt

Figure 8.8 Flux linkage and induced voltage in a simple synchronous machine.

waveforms take on a sine function. The same flux linkage waveforms are applicable for the two other phases, except they are displaced by 120 degree or 2p=3 radians relative to one another. From Eq. [8.2], it can be seen that the induced voltage waveforms are found by differentiating the flux linkage functions. This gives rise to the waveforms in Fig. 8.8(c). The magnitude of the induced voltage depends on the peak flux per pole (and hence is proportional to the field current), the number of turns in a coil and also the rotational speed. It can be seen that some of these variables are fixed at the design stage (for example, number of turns), some may depend on the operational conditions (for example, rotational speed) and some can be controlled (for example, field current). For a permanent magnet generator, once the generator is designed the magnetic flux is fixed as the MMF cannot be controlled. This means that the induced voltage is speed dependent, Efu

[8.4]

The frequency of the induced voltage is dependent on the rotational speed and the number of pole pairs, p. In the case of the machine in Fig. 8.8, there are two poles and hence one pole pair (p ¼ 1). Most wind turbine generators have p > 1, and it is not uncommon for medium-speed permanent magnet generators to have p > 40 and slower directly driven permanent magnet generators to have p > 80. The frequency is given as f ¼

pum 2p

[8.5]

170

Offshore Wind Farms

In this case um is the mechanical rotational speed (rad/s). A useful concept can be the equivalent electrical angular frequency, which is defined as ue ¼ 2pf ¼ pum.

8.2.3.2

Synchronous machine equivalent circuit

Despite being an electromechanical device, it is possible to produce an equivalent circuit diagram of the synchronous machine. This can be useful for understanding the machine’s capabilities and limits, as well as relating to power engineering modelling (either for designing and modelling power converters, or understanding the role of these machines in the power system more generally). The machines are normally designed to have three balanced and identical phases. As such it is sufficient to draw the equivalent circuit for any one of the three phases. So far the discussion of magnetic flux has centred on the flux produced by the rotor field, Ff. As armature current, Ia, starts to flow in the armature winding (the electrical windings on the stator), another magnetic flux is produced, Fa. Part of this flux links with the stator winding only and is known as the leakage flux, Fal. Another part of the armature flux links with the rotor e this is known as the armature reactance flux, Far. As these fluxes are dependent on the magnitude of the armature current and lead to a voltage drop, they can be represented by reactances Xal and Xar in Fig. 8.9(a). These can be combined into the synchronous reactance, Xs, shown in Fig. 8.9(b). There is a further voltage drop due to the resistance of the stator winding Ra. Generally Xs >> Ra and in some cases the resistance can be neglected [6].

8.2.3.3

Synchronous machine phasor diagram

A phasor diagram shows the relationship between voltages and currents. One such diagram is shown in Fig. 8.10 for a non-salient pole synchronous generator. The terminal voltage, Vt, is taken as the reference phasor. The equation for the phasor diagram is, Ef ¼ Vt þ Ia Ra þ Ia jXs

[8.6]

where 4 is the phase angle between the terminal voltage and the armature current and d is the power angle (sometimes referred to as the torque angle). This reveals that for a given induced emf, increasing the armature current will lead to a drop in terminal voltage. As the current increases, the power angle will increase up to a certain limit. Xar

Xal

Xs

Ra

Ra

la

Ia Ef

Vt

Ef

Figure 8.9 Equivalent circuit of a synchronous machine.

Vt


171

Ef

δ la

ϕ

Vt

IaXs

IaRa

Figure 8.10 Phasor diagram for a synchronous generator.

8.2.3.4

Synchronous machine power and torque characteristics

If the armature resistance is neglected then it can be shown that for a three-phase cylindrical pole or surface-mounted permanent magnet machine that the power and torque are given by Eqs [8.7] and [8.8] [6]. P¼

3jVt jjEf j sin d jXs j

[8.7]

T¼

3 jVt jjEf j sin d u jXs j

[8.8]

where u is the rotational speed in rad/s. A theoretical limit of d ¼ 90 degree ¼ p=2 radians will give the peak power and torque. It can be appreciated that machines with higher power ratings tend to have higher voltage ratings or more phases. For machines with saliency e for example the machine in Fig. 8.7(b) and permanent magnet machines with buried magnets e the power and torque characteristics are modified due to a component of reluctance torque. For more details the reader is referred to [6]. In a variable speed application, such as a wind turbine, Eq. [8.4] shows that the induced emf is dependent on the rotor speed. Reactances are also dependent on electrical frequency (X ¼ ueL, where L is an inductance). This means that the power and torque capability depend on the operational speed of the turbine. In order for the synchronous machine to work at variable speed, the terminals of the generator are connected to a power converter. This converter can effectively set the generator’s terminal voltage, Vt, hence controlling the power and torque of the generator. Fig. 8.11 shows the torque speed plane for a wind turbine [7]. The axes are expressed in terms of wind turbine rotor torque (TWT ¼ kTgen, where k is the gearbox ratio and Tgen is the generator torque) and rotational speed (uWT ¼ um/k). A family of curves is defined for constant wind speed and another family represents the operation at various constant performance coefficients of the wind turbine rotor. The shape of these curves depends on the cpel characteristics and is arbitrary for this example. The filled circles represent the points of peak aerodynamic performance at different

172

Offshore Wind Farms

A power converter can modify the effective stator frequency variable speed kTgen

High wind speed

97% cp,max

flow

fhigh

99% cp,max Wind turbine cp,max 99% cp,max 97% cp,max

Low wind speed

‘Variable synchronous speed’ divided by gearbox ratio

2π flow

2π fhigh

kp

kp

ω

Figure 8.11 Torque speed plane for a variable speed wind turbine below rated, showing a synchronous generator with varying frequency.

wind speeds. Because the synchronous machine naturally has a fixed torque speed characteristic, the generator synchronous speed must be varied in order to achieve these points. This is achieved by varying the electrical frequency at the generator terminals. In this example a power converter is used to vary the frequency in the range flow f fhigh.

8.2.4 8.2.4.1

Asynchronous machines Components and principle of operation

Three-phase induction machines are the machines most commonly encountered in industry [8]. Fig. 8.12 shows the cross-section of an induction machine. The machine consists of a stationary stator, an air gap and a rotating rotor. The stator is similar to that in the synchronous machine. The rotor is different to the synchronous machine: alternating current instead of direct current excites the rotor. This alternating current can be induced on the rotor by the three-phase current on the stator or it can be provided via brush gear and slip rings. The rotor can consist of a number of short-circuited copper bars (known as a squirrel cage rotor) or a wound rotor which consists of three windings similar to the stator. Faraday’s law is again the underlying principle of the operation of an induction machine. The ladder example is a useful way of demonstrating how an induction generator works. Fig. 8.13 shows a ladder shape made from an electrically conducting


173

a

cʹ

bʹ

b

c

aʹ

Figure 8.12 Three-phase squirrel cage induction machine.

material. The rungs of the ladder are held in place and short-circuited by two bars, X and Y. When a magnet is moved relative to the rungs, a magnetic flux cuts the conducting ladder rungs. Based on Faraday’s law, a voltage is induced in the ladder because it is made of a conducting material. The induced voltage produces a current that flows through the rung that the magnet is passing over and into the short-circuiting bar before flowing back through the other rungs. This creates a set of poles, the closest of which wants to align with the permanent magnet; the ladder will experience a mechanical force trying to move it to the right. As the ladder moves to the right its conducting rungs will be cut less rapidly by the magnetic field than if it was stationary, reducing the induced voltage and consequently the current and mechanical force. The voltage, current and mechanical force would drop to zero if the ladder was moving at the same speed as the magnet (synchronous speed) due to the magnetic field not being cut by the conducting rungs on the ladder. The magnet can conceptually be replaced by an electromagnet that moves at the same speed. A further conceptual development is to recognise that a moving MMF waveform can be produced by exciting distributed coils with a time-varying current. If the ladder described in the previous paragraphs was rolled into a circle as shown in Fig. 8.14 it would function exactly as a squirrel cage rotor in a rotary induction generator. In a squirrel cage induction machine a voltage is induced in the bars of the squirrel cage making it move (rotate) as the ladder in the previous example does. When the squirrel cage is rotating at the same speed as the flux in the air gap (synchronous speed) there is no voltage induced or force applied to the rotor. When there is a difference in the angular velocity of the squirrel cage rotor and the velocity

174

Offshore Wind Farms

X

(a)

v

(b) Y

N

Rung B v v

(c)

(d) N

N

N ½i

f

i

i

B

S

½i

S v

(e) N

N

(f) N

f i

B

Figure 8.13 Ladder model of an induction machine. (a) Electrically conducting rungs and connecting bars; (b) magnet moves with velocity v; (c) inducing current in rungs; (d) produces force on rungs; (e) magnet can be replaced by moving electromagnet; (f) the moving MMF can also be produced by exciting electromagnets with a time-varying current.

of the flux in the air gap, this is known as slip. Slip, s, is expressed as a percentage of the speed of the flux in the air gap (synchronous speed) and is given by s¼

us u r n s n r ¼ us ns

[8.9]

where us is the synchronous speed and ur is the rotor speed (u is in rad/s, n is in rpm). The synchronous speed is the same rotational speed as the synchronous machine um, as described in Eq. [8.5]. Most induction motors are directly connected to the grid and so common synchronous speeds for a 50-Hz grid are 3000 rpm (p ¼ 1, two poles), 1500 rpm (p ¼ 2, four poles) and 1000 rpm (p ¼ 3, six poles). Slip allows a variation


175

Figure 8.14 Squirrel cage.

from the synchronous speed. Larger values of slip mean that a greater percentage of the rated power is lost on the rotor e leading to heat generation e and so larger machines tend to have smaller values of slip. The basic induction machine starts to work as a generator when the rotor speed exceeds the synchronous speed. This will be examined in more detail in the next subsection.

8.2.4.2

Equivalent circuit and torque speed diagrams

For the purpose of network simulation induction, generators can also be represented by an equivalent circuit. The equivalent circuit for an induction generator is similar to that of a transformer and is shown in Fig. 8.15. In Fig. 8.15, V1 is the per-phase terminal voltage, R1 is the per-phase stator winding resistance, X1 is the per-phase stator winding leakage reactance, E1 is the per-phase induced voltage in the stator winding, Xm is the per-phase stator magnetizing reactance, Rc is the per-phase stator core loss resistance, E2 is the per-phase induced voltage in the rotor at standstill, R2 is the per-phase rotor circuit resistance and X2 is the per-phase rotor leakage reactance. Rotor circuit parameters are referred to the stator using the ratio of turns on the stator and rotor, a ¼ N1/N2. Fig. 8.16 shows a torque speed diagram for an induction machine (for convenience, T is drawn as positive for generating action). The torque speed curve is given by Eq. [8.10] from [6], Tmech ¼

3 02 R02 I us 2 s

[8.10]

176

Offshore Wind Farms

(a)

l1

R1

X1

lφ

lc V1

X2ʹ = a2X2

l2ʹ = l2/a

Xm

Rc

R2 ʹ

lm E1 = aE2

R2 ʹ s

R2 ʹ s

(1-s)

=

a2 R 2 s

Pag

(b)

l1

R1

X1

X2ʹ

l2ʹ lφ E1

V1

R2 ʹ

Xm

s

Figure 8.15 Induction machine equivalent circuits. (a) Full [9]; (b) IEEE-recommended approximation [8].

Synchronous machine

T

1 0

Motor

Ind

uc

tio

nm

0

ωs

ac

hin

e

–1 Generator

s

2ω s ω

Figure 8.16 Induction machine torqueespeed curve.

It can be seen that as the slip approaches zero (s/0, u/us) the torque falls to 0. The induction machine has a linear torque speed relationship when operating close to the synchronous speed. The highest torque, Tmax, can be changed by the generator designer by adjusting the resistance of the rotor. Adjusting the rotor resistance changes the shape of the torqueespeed curve.


8.2.4.3

177

Variable speed operation

In the early grid-connected wind turbines, squirrel cage induction generators were used and directly connected to the grid. The turbines were essentially fixed speed, but the generator slip allowed some compliancy giving less severe loading than was the case for synchronous machines directly connected to the grid. The first step to introducing variable-speed operation was to use squirrel cage induction machines with two sets of windings that could be changed between; these gave rise to two different synchronous speeds and gave a crude version of variable-speed operation. Changing between the generator windings can be electrically and mechanically stressful and is not often used anymore. A further development is to use a wound rotor with variable rotor resistors. By changing the shape of the torqueespeed curve, increasing the rotor resistance can give greater slip at lower torques. This can be achieved with external resistors connected to the wound rotor with brush gear and slip rings, however this can be prone to failure. Some machines used resistors mounted onto the machine rotor with an optical communication link to signal when to switch resistance. Although able to give variable-speed operation, this is at the expense of electrical losses on the rotor. One of the next developments was to identify how to usefully extract this slip energy. In a doubly fed induction generator (DFIG) there are windings that allow for power take off on both the stator and the rotor, as shown in Fig. 8.17. The stator winding is directly connected to the grid and the rotor winding is connected to a power converter (via slip rings and brushes) before reaching the grid. The use of the power converter allows the rotor currents to be adjusted so that the frequency and direction can vary [10]. At synchronous speed, the rotor can be fed with DC and will act as if it were a synchronous machine. In subsynchronous operation, power is fed to the rotor from the grid. If the synchronous speed is 1500 rpm and the rotor speed is 1200 rpm, this implies currents of a frequency that produce a magnetic field that moves relative to the rotor at a speed of 300 rpm. In supersynchronous operation, power is taken from the rotor and fed to the grid. In the same example, a rotor speed of Generator (DFIG)

Three-stage gearbox

Transformer

G

Grid

Power converter

Figure 8.17 DFIG power train configuration.

178

Offshore Wind Farms

T

T

v3 v2 v1 0

ω

ω

0 f1 f2 f3

Figure 8.18 Varying an induction machine’s torque speed: (a) by varying terminal voltage; (b) by varying frequency.

1800 rpm would need currents of a frequency that produce a field that moves relative to the rotor at a speed of 300 rpm. The DFIG configuration allows for the rotor speed to vary around the synchronous speed by a greater amount than the squirrel cage induction generator with no power converter and also allows for the use of less power electronics than a fully rated converter. This is an advantage as power converters reduce the reliability of the power train. However, some of the reliability improvements over the standard induction generators are cancelled out due to the use of slip rings and brushes with the DFIG [2]. Fully rated converters can be used with squirrel cage induction machines to give variable-speed operation. Fig. 8.18 shows the effect on the torque speed characteristic of a machine when varying terminal voltage and when varying the frequency. These can be used in combination to meet the torque speed characteristics needed for a variable-speed wind turbine.

8.3

Practical design and manufacture of electrical generators

In the design and manufacturing of generators, designers and manufacturers must work with a number of components and specialist materials. As described in Section 8.2, a generator consists of a rotor and a stator with bearings as the mechanical interface between them (and possibly brush gear and slip rings acting as part of the electrical interface). Electrically these machines have windings, different types of insulation and electrical terminals. From a magnetic point of view the design and manufacture of flux-conducting material such as steel laminations is important and the specification of any permanent magnets has an important influence on cost and losses. As well as these items there are requirements for auxiliary subsystems such as forced air- or liquid-coolers, and mechanical parts that are able to meet load requirements and provide protection for the generator. This section gives a brief overview of generator components and their design.


8.3.1

179

Rotor

The rotor in a generator is primarily used to create a magnetic flux in the air gap but it can be used for power take off, as in a DFIG. Ordinarily, the rotor is placed inside the stator allowing for a short conduction cooling path from the hotter stator to ambient. In some cases e such as the Siemens direct-drive wind turbines e an outer rotor is used [11]. This allows for a slightly larger air gap diameter for the same outer diameter and hence better torque for the same force production. The manufacturing process of a generator rotor depends on the type of generator it is to be used in. Two of the most common types of rotors in wind energy are the permanent magnet rotor in a permanent magnet synchronous generator and the wound rotor in a DFIG. As the name suggests the permanent magnet rotor is manufactured using permanent magnets, such as neodymium (NdFeB). These have the advantages of having high maximum energy product, meaning that they can produce a strong flux density using a relatively small mass of material. They have a high coercivity and remnant flux density but a relatively modest relative permeability see, e.g., [12]. This means that they are able to resist demagnetisation but that the magnets themselves are not very good at ‘conducting’ magnetic flux and care must be taken to design the magnets to get best use of the expensive magnet material. Fig. 8.7(c) shows these magnets mounted onto the surface of the rotor. Even though they appear to be salient pole machines, this low permeability means that the air gap reluctance is fairly constant as one travels circumferentially around the air gap radius. As such they can be treated as non-salient machines. Some machines, using less energy-dense permanent magnetic materials e such as Ferrite magnets e need to use flux concentration techniques, burying magnets between steel poles. In this case these machines are salient in nature. Sometimes the magnets on the rotor are skewed relative to the teeth and slots on the stator. This helps reduce the effect of cogging torque. This effect, whereby the rotor magnets favour being in alignment with the teeth on the stator can add an unwanted ripple onto the machine torque and mean that there is a minimum starting torque to get the generator to rotate. Skewing means that the circumferential position of the magnet pole varies as one moves axially from one end of the machine to the other. Other techniques for reducing cogging torque also exist, including shaping the magnets in the radial, axial and circumferential dimensions. The wound rotor in the DFIG is manufactured using copper wire wound around steel teeth in a cylindrical core. In a squirrel cage induction machine, electrically conducting bars are embedded into the same types of slots in the rotor steel. The core of the rotor is made from steel laminations. In a DFIG the shaft of the rotor is then connected to slip rings and brushes.

8.3.2

Air gap

Termed the ‘air gap’, the gap between the rotor and stator has two interpretations. The first is the physical distance between the rotor and stator surfaces. This must have a non-zero value at all points around the air gap radius, at all points in time e otherwise

180

Offshore Wind Farms

physical damage would occur as the rotor and stator come into contact. The gap must be dimensioned to allow for manufacturing tolerances, rotor and stator eccentricity, deformation under loads, as well as the impact of loads from the wind turbine rotor itself. A starting point for design of larger, direct-drive machines is that the air gap, g, is often dimensioned according to the rule of thumb g ¼ dg/1000, where dg is the diameter at the midpoint of the air gap. Hence a 5-m diameter machine would have a 5-mm air gap. From a magnetic perspective, the air gap is important as it leads to reluctance which reduces the flux per pole. The magnetic or effective air gap is the notional dimension of air that gives rise to the same magnetic reluctance as the real machine. The effect of slotting (which tends to increase reluctance) can be factored in. The biggest difference between ‘real’ and ‘effective’ air gap values is found in surface-mounted magnet machines. Here the height of the magnet e a material which has almost the same permeability as air e is also included in the air gap. So, a 15-mm-high magnet and a physical air gap dimension of 5 mm leads to an effective magnetic air gap of 20 mm (if we neglect the effect of slotting). Reducing the physical air gap from 5 to 4 mm (20%) only reduces the magnetic air gap from 20 to 19 mm, and hence reluctance reduces by only 5%.

8.3.3

Stator

The stator is the part of the generator that remains stationary while the rotor rotates relative to it. It consists of conductors which are normally embedded into slots in the surface of the stator laminations. These conductors in the stator are usually made of copper windings but can be made from aluminium. Insulation is used to prevent short circuits between neighbouring conductors, between the conductors and the steel laminations and also between one phase and another. Insulation is normally further improved by using a technique called vacuum pressure impregnation, whereby the stator is dipped in epoxy resin and cured in a vacuum vessel. The steel laminations are thin sections coated with a thin layer of insulating material. Laminations are used instead of solid steel because they reduce eddy current losses. Special electrical steel with silicon content is used which is designed to reduce hysteresis losses. The generator frame and housing provide the structural support and protection for the internal workings of the generator. Heat can build up in a generator during the generation process leading to losses in efficiency, so cooling systems are required. Two of the basic types of cooling systems in wind energy generators are forced air and liquid cooling [13].

8.3.4

Structural integrity

High-speed machines can be thought of as standalone machines, relatively independent of the wind turbine. The major challenges from a mechanical point of view are the relatively high rotational speeds, the input loads from the high-speed shaft of the gearbox and misalignment [14].


181

Low-speed machines in direct-drive wind turbines are much larger and so tend to be significantly integrated into the wind turbine structure. As the torque rating increases, so too does the air gap diameter and axial length. Due to the magnetic forces of attraction (between the ferromagnetic parts of the rotor and the stator), there tend to be large forces trying to close the air gap. This, along with the torque loading, can mean that these generators need considerable structures in order to give sufficient stiffness [13]. For large direct-drive machines, this mass can dominate the mass of the material used to produce electrical power [15].

8.3.5

Generator losses

The major sources of losses in AC machines are copper losses and iron losses. There are also minor loss mechanisms due to bearing friction and windage (aerodynamic drag on the rotor surface). Copper losses result from current flowing through the resistance of any windings (field or armature). This can be reduced by reducing resistance, either by increasing the cross-sectional area of conductors (generally increases the size of the machine), lowering the winding temperature (reduces the resistivity) or using materials which have lower resistivity. Copper is usually used as the conductor material and increasing current densities can be achieved e without significant additional losses e with improving cooling systems. Iron loss is an important loss component in the components that conduct flux, which changes with time. They are sometimes described as ‘core losses’. The two loss mechanisms are hysteresis and eddy current losses. Both of these increase with increasing flux density in the teeth and back iron. Hysteresis losses are proportional to the electrical frequency and the eddy current losses are proportional to the square of the electrical frequency. The electrical steel used in the stator and rotor of induction machines is affected, whereas the rotors of synchronous machines are less affected as the main field is stationary relative to the rotor.

8.4

Selection of generators for offshore wind turbines

The following section provides examples of the generator types used in some of the more popular offshore wind turbines currently installed. The number of different generator types used highlights the variety in modern offshore wind turbine drivetrain configurations.

8.4.1

Siemens SWT 2.3 MW: SCIG

The Siemens SWT 2.3 uses a squirrel cage induction generator (SCIG), as shown in Fig. 8.19. It has a rated power of 2.3 MW and an output voltage of 690 V. It is cooled through an integrated heat exchanger and a separate thermostat-controlled ventilation arrangement that Siemens claim allows for cooler operation temperatures providing longer lifetime to the winding insulation [16].

182

Offshore Wind Farms

Transformer

Generator (SCIG)

Three-stage gearbox

G

Power converter

Grid

Figure 8.19 Siemens SWT 2.3-MW power-train configuration.

No slip ring or brushes are required for this generator type because the squirrel cage is excited through electromagnetic induction. This allows for increased reliability because slip rings and brushes make up w50% of the overall generator failure rate [2]. Unlike PMGs, the SCIG does not require any rare earth materials in its manufacture. This is advantageous for manufacturers as it allows for better cost planning. It is notable that Vestas also opted for the SCIG in the second version of their V112 3.0 MW [17] rather than the PMG as used in the first version of their V112 3.0 MW [18].

8.4.2

Vestas V90: DFIG

The generator in a Vestas V90 is a doubly fed induction generator, as shown in Fig. 8.17. The V90 is available with generators rated at 1.8, 2 and 3 MW. The 3-MW generator has an output voltage of 1000 V. The generator has four poles and is liquid-cooled. Slip rings and brushes are required for this generator type for power take off. The inclusion of these slip rings and brushes decreases reliability in comparison to the PMG [2], however a DFIG is cheaper than a PMG by up to w40% [19]. Like the SCIG and unlike PMGs, the DFIG does not require any rare earth materials in its manufacture. The use of the partially rated power converters with the DFIG allows for compliance to many grid codes, however the DFIG is not completely decoupled from the grid as in the PMG or SCIG, which both use fully rated converters.

8.4.3

Areva M5000: medium-speed PMG

The generator in an Areva M5000 is a medium-speed permanent magnet generator (PMG), as shown in Fig. 8.20. It has a rated power of 5 MW and an output voltage of 3.3 kV. Unlike the machines in the previous turbines it is a synchronous machine, instead of an induction machine. As it is a permanent magnet generator no brushes or slip rings are required to excite the rotor and power take off occurs at the stator. As the input to the generator is from a single-stage gearbox the generator has a lower speed range than the previous turbine types, this means higher torque and a larger generator with more windings and insulation. As a result of these extra failure modes it is expected that lower-speed generators will have more winding failures than the high-speed generators in the previous two turbines but less than in an equivalent direct-drive generator [20].


183

Transformer

Generator (PMG)

One-stage gearbox

G

Grid

Power converter

Figure 8.20 Areva M5000 power-train configuration.

The single rotor main bearing, single-stage planetary gearbox and the generator are packaged together in a compact single cast structure leading to a relatively lightweight solution.

8.4.4

Siemens SWT 6 MW: direct-drive PMG

The generator in a Siemens SWT 6 MW is a direct-drive PMG, as shown in Fig. 8.21. It has a rated power of 6 MW and an output voltage of 690 V. Like the Areva turbine, it is a synchronous machine and as it is a permanent magnet generator no brushes or slip rings are required to excite the rotor. As the input to the generator comes directly from the rotor it is a low-speed high-torque input. As the generator is direct drive it has a larger size, with an outer diameter of 6.5 m [11]. Located upwind of the tower, just behind the wind turbine rotor, these generators include structural elements to transmit loads from the wind turbine rotor back through to the turbine, as well as an access channel to get into the wind turbine hub.

8.4.5

Case study of generator selection

The following section shows a case study based on results from [2] [19], and [20] in which two turbine types are compared. One of the turbine types has a DFIG and the other has a PMG, both are compared based on cost of energy (CoE) and the results are shown in Section 8.4.5.6. The results that are compared in the following sections are based on operational data from modern multi-megawatt offshore wind turbines. Transformer

Generator (PMG)

G

Direct drive

Power converter

Figure 8.21 Siemens 6-MW power-train configuration.

Grid

184

Offshore Wind Farms

The following sections show availability and operation and maintenance (O&M) cost for sites from 10 to 100 km offshore. The CoE comparison for both turbine types is carried out for a site located 40 km offshore as in [19].

8.4.5.1

Cost of energy

For the purpose of this case study, the cost of energy has been calculated using assumptions as used in the Crown Estate’s ‘Offshore Wind Cost Reduction’ study [4]. Eq. [8.11] shows how the CoE is calculated [21], Cost of energy ¼

ðInitial capital cost Fixed charge rateÞ þ annual operating cost Annual energy production [8.11]

Capital costs include the cost of consenting/development, foundations, turbines, array cables, installation, decommissioning, insurance and project management. In this comparison the values for each of these figures, except for turbine costs, are the same and come from [4], turbine costs differ because of generator and converter costs and come from [20]. The fixed charge rate is defined so as to express the initial capital cost (and implied financing costs) fairly to a per-year basis. Annual operating costs consist of operation and maintenance costs, insurance costs, transmission charges and seabed rent. Each of these values are the same for both turbines and come from [4], except for O&M costs, which differ for the two turbine types and come from [20]. Annual energy production is different for the two turbine types as the turbines have different efficiencies and availability; in this case the data come from [20].

8.4.5.2

Generator costs

Normalised cost (%)

Fig. 8.22 shows the different manufacturing costs of a PMG and a DFIG with the same rated power. Both turbine types have other components that differ, such as converters, so Fig. 8.17 also shows the difference in cost for this when they are used with both generator types. 100 80 60 40 20 0

DFIG PRC PMG FRC

Gearbox

Generator

97.95%

60.34%

39.47%

100.00%

100.00%

100.00%

Figure 8.22 Cost comparison of PMG versus DFIG [20].

Converter


8.4.5.3

185

Efficiency and losses

Fig. 8.23 shows the electrical power curves for a DFIG turbine type and a PMG turbine type of the same rated power. The difference in power curve highlights the greater efficiency of the PMG due to the lack of copper losses on the generator rotor.

8.4.5.4

Availability

The availability of a turbine using a PMG and the availability of a turbine using a DFIG are seen in Fig. 8.24. These availability figures are based on a hypothetical offshore wind farm of 100 turbines located 10, 50 and 100 km offshore. The difference is driven by the failure rates and downtimes of both generator types and the different converters required for them to operate.

Normalised power output (%)

120 100 80 60 40 20 0

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

DFIG PRC 4.1%

8.6% 14.5% 22.6% 33.3% 45.9% 59.3% 72.1% 83.3% 91.9% 97.0% 99.0% 99.6% 99.7% 99.7% 99.8% 99.8% 99.8% 99.8% 99.8% 99.8% 99.8%

PMG FRC 4.6%

9.8% 15.7% 23.2% 33.6% 46.7% 61.1% 74.6% 85.5% 93.3% 97.9% 99.5% 99.9% 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Wind speed (m/s)

Figure 8.23 Normalised power curves for PMG and DFIG [20].

Availability (%)

95 90 85 80 75 70 DFIG PRC PMG FRC

10 km 91.63% 92.22%

50 km 90.48% 91.27%

100 km 79.34% 79.75%

Figure 8.24 Availability of PMG and DFIG turbine types at different sites [20].

186

Offshore Wind Farms

8.4.5.5

O&M costs

The O&M costs of a turbine using a PMG and a turbine using a DFIG can be seen in Fig. 8.25. Like availability, these O&M costs are based on a hypothetical offshore wind farm of 100 turbines located 10, 50 and 100 km offshore. The difference is driven by the failure rates, downtimes, repair costs, vessel costs and staff cost of both generator types and the different converters required for them to operate. Fig. 8.25 also shows the O&M cost when the cost of lost production is also included.

8.4.5.6

Analysis

O&M cost /MWh (£)

The CoE comparison for the DFIG and PMG turbine type for a hypothetical 100 turbine wind farm located 40 km offshore can be seen in Fig. 8.26. The DFIG turbine has an overall CoE of £103.43/MWh and the PMG turbine type has an overall CoE of £101.23/MWh. As mentioned in Section 8.4.5.1 it is assumed that certain parts of the capital cost and operating cost remain the same for both turbine types. The drivers for the £2.20/MWh difference between the two turbine types are the variance in turbine cost, O&M cost and energy production. It is worth noting that the turbine cost and O&M cost figures in this analysis are based on cost data from a manufacturer and O&M provider so profit margins are not considered. These profit margins would drive up the CoE for operators. To put this cost/MWh difference into perspective, an offshore wind farm of 100 turbines, each producing 12,961 MWh [22] annually for a 20-year lifetime will see a cost difference of around £50 million pounds between the two turbine types, in favour of the permanent magnet generator over the lifetime of the wind farm. Similar analysis but for medium-speed and low-speed permanent magnet generators tends to show that the cost of energy is even further reduced relative to these high-speed gearbox-driven generators [20].

70 60 50 40 30 20 10 0

10 km

50 km

100 km

17.79

20.85 17.84

DFIG PRC (lost production) (£)

17.24 14.55 29.42

PMG FRC (lost production) (£)

25.74

28.01

DFIG PRC (£) PMG FRC (£)

14.99 32.17

Figure 8.25 O&M costs of PMG and DFIG turbine types at different sites [20].

59.10 55.26


187

120 100

£/MWh

80 60 40 20 0

DFIG PRC £/MWh

Turbine Operations Support structure Array cables Installation

PMG FRC £/MWh

4.02

4.52

17.40

14.80

19

19

4

4

16

16

Contingency

14

14

Transmission

20

20

Decommisioning

5

5

Seabed rent

4

4

Figure 8.26 O&M costs of PMG and DFIG turbine types at a site 40 km offshore [19].

8.5 8.5.1

Future trends in offshore wind turbine generators Future challenges

As wind turbines get larger and move further offshore a number of challenges exist in all areas of the wind turbine and the generator is no exception to this. Larger generators located further offshore will face challenges when it comes to increased costs, reliability, accessibility and growing magnitude of loads. As generator sizes increase, the cost of materials used in generator construction will naturally increase, however the big challenge in this area will come from the larger generators requiring more rare earth material for their permanent magnets. Rare earth materials used in permanent magnets e such as Dysprosium e are in short supply and are readily mined in only a few countries. This allows the supply and hence price to be manipulated. Price volatility e which generator manufacturers try to avoid e represents a risk which turbine designers would prefer to avoid. Reliability of generators becomes even more important as turbines move further offshore and access becomes more challenging. Figs 8.24 and 8.25 show how the same technologies at different distances have lower availability and increased O&M costs as the distance to shore increases. This implies that future generators need to have lower failure rates and that the time and cost of repair need to be reduced.

188

Offshore Wind Farms

One way to reduce the cost of energy for offshore wind farms is to increase the power rating and rotor diameter size. Assuming the same tip speed limit (eg, 90 m/s) and rated power per swept area (eg, 350 W/m2) figures, one can find that the rated torque grows from 3.75 MNm for a 5-MW turbine to 30.0 MNm for a 20-MW turbine. Especially for direct drive machines, this eightfold increase in torque for a fourfold increase in power means that e unless different technologies are used e the cost and size increase in the generator will rapidly outpace improvements in energy capture.

8.5.2

Developing existing generator technologies

As a means of overcoming some of the future generator challenges, new technologies are being developed. Wind generator researchers, designers and manufacturers are investigating and developing generator technologies, like brushless DFIG generators and air-cored PM machines, as well as adapting current technologies to make them more maintainable and reliable. Brushless DFIGs are of interest to the wind energy industry because they would remove the brush and slip ring failure mode while keeping the lower failure rate of the partially rated converters [23]. The brushless DFIG operates through the use of a second stator winding e often termed the control winding e which fulfils the role of the brush gear and slip rings. The two stator windings are housed within the same generator frame but no direct coupling occurs between them. Air-cored permanent magnet machines use air-cored windings in the stator of the machine. This allows for the elimination of ferromagnetic material in the stator which leads to the elimination of the force between rotor and stator described in Section 8.3.4. This in turn leads to advantages such as weight saving of up to 30e50%, making stator repairs easier and eliminating cogging torque [24]. As a means of making generators more maintainable and reliable, researchers and designers are designing for maintainability and fault tolerance. Generators and wind turbines that include modularity and redundancy are now being seen in academia and industry [25]. Like all reliability-related issues the business case for modularity and redundancy greatly improves as turbines move further offshore. Modularity allows for a failed module of the generator to be changed without requiring the replacement of the full generator. As well as the obvious cost advantage of replacing a module instead of a full unit, there are also huge advantages in the cost savings from vessel hire due to a module not requiring heavy lift vessels that may have long waiting times and have a day rate of over £100,000 [5]. Redundancy allows for generators to continue producing energy at a full or partial percent of the rated power even after a failure occurs. Although capital costs are usually higher due to the duplication of parts of the generator or the full generator, recent studies have shown that O&M savings may justify the higher capital costs [26].

8.5.3

Developing more radical generator technologies

Along with the conventional and emerging generators discussed above, a number of other generator technologies exist or have potential to be used in wind energy in the


189

future. Technologies such as high-temperature superconducting generators and synchronous reluctance machines are being researched [27]. High-temperature superconducting machines encountered in the literature are both synchronous and asynchronous. Superconducting machines replace some of the materials used in conventional generators with superconducting materials. Superconductors offer greater efficiency across all loads due to having no DC resistance. They also allow for far stronger magnetic fields to be produced in comparison to conventional generators. This leads to higher torque density and reduced size, attractive qualities especially for direct-drive turbines. One significant challenge is that the machines have to be cryogenically cooled to around 50 K. The reliability and maintenance intervals of these cryostats and generators is one of the biggest hurdles: it can take more than a day to cool a generator to its operating temperature [27]. Synchronous reluctance machines are another form of electrical machine that are known for their low cost as they avoid the use of permanent magnets. The synchronous reluctance machine makes use of the effect of saliency, the component of torque which is different between cylindrical and salient pole synchronous machines. The lack of brush gear slip rings is attractive and a few authors have started to look at these machines in some detail [28].

8.5.4

Generators for use with advanced torque/speed conversion

Except for direct-drive wind turbines, torque and speed have been traditionally converted using a mechanical gearbox. Due to a high gearbox failure rate [29] and high direct-drive generator failure rates [30] alternative torque speed conversions for wind turbines have been investigated. These alternative methods include hydraulic and electromagnetic conversion and enable the use of other types of generators. The hydraulic torque/speed converter in a wind turbine uses a hydraulic pump, accumulator and motor, which drives a generator at a constant speed. It is the modulation of these hydraulic units that gives the turbine its variable speed capability, rather than the control of the generator. The fact that the generator can be driven at a constant speed allows for the use of a constant-speed synchronous generator. In one of the wind turbine types that uses hydraulic torque/speed conversion, brushless synchronous generators are used (in this configuration, an auxiliary electrical machine is mounted on the shaft and is used to generate and regulate the DC) [31]. Electromagnetic torque conversion and power generation have also been developed in a single unit for a wind turbine. This configuration is being called pseudo direct drive [32]. The low-speed input is converted to a higher speed through three hollow cylinder rings placed inside each other. The outer and inner rings have permanent magnets arranged in alternating northesouth patterns. The middle ring consists of steel segments that alter the magnetic field between the inner and outer ring. It is this ring which is connected to the low-speed shaft. Stator windings are then placed around the outside magnetic ring as seen in Fig. 8.27. This allows for a current to be induced in the windings as the magnetic outer ring rotates inside the stator.

190

Offshore Wind Farms

Figure 8.27 Pseudo direct drive.

8.5.5

Concluding remarks

Offshore wind turbine manufacturers have started to adopt electrical generator technologies which are somewhat different to those used on onshore, mainly because of the different challenges which are met offshore. Whereas the development of generators onshore has adapted existing low-cost, industrial electrical machine technologies to suit wind turbines, the emphasis offshore is on more tailored solutions which can deliver high efficiency and high availability. There is a wide variety of approaches currently being used in offshore wind and it is still not clear which technical solution will ultimately dominate. This means that the research and development of offshore wind turbine generators continues to be an exciting field of endeavour.

Sources of further information There are many excellent textbook on the study of electric machines. These can be useful to further understand and explore some of the basics of rotating electrical machines. Some of the better texts include: • • • •

P.C. Sen, Principles of Electric Machines and Power Electronics, John Wiley & Sons, 2007. A.E. Fitzgerald, C. Kingsley, S.D. Umans, Electric Machinery, McGraw-Hill, 2003. J. Hindmarsh, Electrical Machines and Their Applications, Butterworth-Heinemann, 1995. A. Hughes, B. Drury, Electric Motors and Drives, Elsevier, 2013.

The adaptation of electrical machines for use in renewable energy is sometimes only briefly addressed in some of these classic texts. The reader may be interested in these books which explore the wind turbine context, the interaction of the generator with some of the other subsystems as well as some of the practical aspects in more detail: • •

M. Mueller, P. Polinder (Eds.), Electrical Drives for Direct Drive Renewable Energy Systems, Woodhead Publishing (2013). V. Akhmatov, Induction Generators for Wind Power, Multi-Science Publishing, 2007.


191

Although the reader may be aware of the journals focusing on wind and other renewables, much of the up-to-date research on wind turbine electrical generators is published in journals that focus on the electrical machine itself. The following are good examples: • •

IET Electric Power Applications, Online ISSN 1751e8679, Print ISSN 1751-8660 IEEE Transactions on Energy Conversion, ISSN: 0885-8969

References [1] A.S. McDonald, O. Keysan, How can electrical machine and drive-train design influence offshore wind cost of energy? UK Magnetics Society Seminar, Electromagnetics in Renewable Energy Generation, 8th July 2015, Edinburgh, UK. [2] J. Carroll, A.S. McDonald, D. McMillian, Reliability comparison of wind turbines with DFIG and PMG drive trains, IEEE Trans. Energy Convers. 30 (June 2015) 663e670. [3] S. Eriksson, H. Bernhoff, Rotor design for PM generators reflecting the unstable neodymium price, in: Electrical Machines (ICEM), 2012 XXth International Conference on. IEEE, 2012. [4] BVG Associates, Offshore wind cost reduction pathways: technology work stream, The Crown Estate (May 2012). [5] I. Dinwoodie, et al., Development of a combined operational and strategic decision support model for offshore wind, Energy Procedia 35 (2013) 157e166. [6] P.C. Sen, Principles of Electric Machines and Power Electronics, John Wiley & Sons, 2007. [7] W.E. Leithead, S. de la Salle, D. Reardon, Role and objectives of control for wind turbines, Gener., Transm. and Distrib., IEE Proc. C 138 (2) (1991). IET. [8] I. Boldea, S. Nasar, The Induction Machine Handbook, CRC Press, 2001. [9] IEEE, IEEE Standard Test Procedure for Polyphase Induction Motors and Generators, IEEE Std 112-2004 (Revision of IEEE Std 112-1996), 2004, pp. 1e79. [10] R. Pena, J.C. Clare, G.M. Asher, Doubly fed induction generator using back-to-back PWM converters and its application to variable-speed wind-energy generation, in: IEE Proceedings-Electric Power Applications 143.3, 1996, pp. 231e241. [11] E. de Vries, Close up- siemens SWT e 6 MW 120 offshore turbine, Wind Power Mon. (June 2011). [12] Arnold Magnetics. N40M Sintered Neodymium-Iron-Boron Magnets Data Sheet, accessed online on: 18.08.15, accessed at: http://www.arnoldmagnetics.com/Neodymium_Magnets.aspx. [13] A.S. McDonald, M. Mueller, A. Zavvos, Electrical, thermal and structural generator design and systems integration for direct drive renewable energy systems, in: M. Mueller, H. Polinder (Eds.), Electrical Drives for Direct Drive Renewable Energy Systems, Woodhead Publishing, 2013. [14] M. Whittle, J. Trevelyan, W. Shin, P. Tavner, Improving wind turbine drivetrain bearing reliability through pre-misalignment, Wind Energy 17 (2014) 1217e1230. [15] A.S. McDonald, M.A. Mueller, H. Polinder, Structural mass in direct-drive permanent magnet electrical generators, in Renewable Power Generation, IET 2 (1) (March 2008) 3e15, http://dx.doi.org/10.1049/iet-rpg:20070071. [16] Siemens. Wind Turbine SWT 2.3-93, accessed online on: 17.05.15, accessed at: http:// www.energy.siemens.com/nl/en/renewable-energy/wind-power/platforms/g2-platform/windturbine-swt-2-3-93.htm.

192

Offshore Wind Farms

[17] Vestas. “General Specification V112-3.3 MW.” accessed online on: 19.05.15, accessed at: http://www.eastriding.gov.uk/publicaccessdocuments/documents/JAN2015/AAC0C504E FE411E499AFF8BC12A86C8B.pdf. [18] Vestas. “V112e3MW Offshore.” accessed online on: 19.05.15, accessed at: http://www. vestas.com/files%2Ffiler%2Fen%2Fbrochures%2Fbrochure_v112_offshore_june2010_ singlepage.pdf. [19] J. Carroll, A.S. McDonald, D. McMillian, Offshore cost of energy for DFIG PRC turbines vs PMG FRC turbines, in: IET RPG Conference, Beijing, September 2015. [20] J. Carroll, A.S. McDonald, D. McMillian, A comparison of the availability and operation & maintenance costs of offshore wind turbines with different drive train configurations, Submitted to Wind Energy J. (August 2015). [21] C. Moné, et al., 2013 Cost of Wind Energy Review, Technical Report NREL/TP5000e63267, National Renewable Energy Laboratory, February 2015. accessed on 18.08.15, accessed at: http://www.nrel.gov/docs/fy15osti/63267.pdf. [22] European Wind Energy Association. “Wind Energy Statistics and Targets.” accessed on 18.02.15, accessed at: http://www.ewea.org/uploads/pics/EWEA_Wind_energy_ factsheet.png. [23] Wind Technologies. “Technology Overview.” accessed online on: 22.05.15 accessed at: http://www.windtechnologies.com/Wind/Wind-Technologies/Show/49/65/67/Technology/ Innovative-drivetrain/Brushless-DFIG.aspx. [24] M.A. Mueller, A.S. McDonald, A lightweight low-speed permanent magnet electrical generator for direct-drive wind turbines, Wind Energ 12 (2009) 768e780, http:// dx.doi.org/10.1002/we.333. [25] NGenTec. “Design and Technologies.” accessed online on: 22.05.15, accessed at: http:// www.ngentec.com/design_and_technology.asp. [26] J. Carroll, I. Dinwoodie, A.S. McDonald, D. McMillan, Quantifying O&M savings and availability improvements from wind turbine design for maintenance techniques, in: EWEA Offshore Conference, Copenhagen, Denmark, April 2015. [27] O. Keysan, Application of high-temperature superconducting machines to direct drive renewable energy systems, in: M. Mueller, H. Polinder (Eds.), Electrical Drives for Direct Drive Renewable Energy Systems, Woodhead Publishing, 2013. [28] B. Boazzo, et al., Multipolar ferrite-assisted synchronous reluctance machines: a general design approach, Ind. Electron., IEEE Trans. on 62 (2) (February 2015) 832e845. [29] J. Carroll, A. McDonald, D. McMillan, Failure rate, repair time and unscheduled O&M cost analysis of offshore wind turbines, Wind Energ (2015), http://dx.doi.org/10.1002/ we.1887. [30] F. Spinato, P. Tavner, G.J.W. van Bussel, E. Koutoulakos, Reliability of wind turbine subassemblies, IET Renew Power Gener. 3 (4) (2009). [31] Artemis. “Wind Turbines.” accessed online on: 22.05.15, accessed at: http://www. artemisip.com/applications/wind-turbines. [32] A. Penzkofer, K. Atallah, Analytical Modeling and Optimization of Pseudo-Direct Drive Permanent Magnet Machines for Large Wind Turbines, in Magnetics, IEEE Transactions on 51 (12) (December 2015) 1e14, http://dx.doi.org/10.1109/TMAG.2015.2461175.

Modelling of power electronic components for evaluation of efficiency, power density and power-to-mass ratio of offshore wind power converters

9

R.A. Barrera-C ardenas Faculty of Pure and Applied Sciences, University of Tsukuba, Tsukuba, Japan M. Molinas Department of Engineering Cybernetics, Norwegian University of Science and Technology, Trondheim, Norway

9.1

Introduction

In an offshore environment, the design of the wind energy conversion systems (WECS) requires taking into account not only efficiency and reliability but also size and weight, as expensive platforms must be placed to support each component. A contribution of around 15% in active volume and around 10% in active weight is normally reported for power electronics converters in wind turbine applications (Blaabjerg et al., 2006). Therefore power density is also a performance index of paramount importance, especially when most of the electrical power conversion components are going to be located in the nacelle or tower of the wind turbine (WT). The efficiency (h) of an electrical system is the ratio of power output and power input (Pin) (Eq. [9.1]), and the power output is the difference between power input and the total losses from the input to output stages of the WECS, including power semiconductors and passive elements like inductors and capacitors. h¼

Pin

P

Ploss;ðiÞ 100 Pin

[9.1]

On the other hand, the power density (r) defined by Eq. [9.2] characterizes the degree of compactness of a WECS. r depends on the total converter volume and power losses of the system. The converter volume (VolTotal) is the summation of the individual volumes Vol(i) of the components, and the utilization of the VolTotal by active parts is characterized by the volume utilization factor CPV, which has typical values between 0.5 and 0.7 (Kolar et al., 2010).

Offshore Wind Farms. http://dx.doi.org/10.1016/B978-0-08-100779-2.00009-X Copyright © 2016 Elsevier Ltd. All rights reserved.

194

Offshore Wind Farms

P P Pin Ploss;ðiÞ Pin Ploss;ðiÞ r¼ ¼ 1 P VolTotal VolðiÞ C $

[9.2]

PV

The WECS should be located in the nacelle, tower or pillar of the WT, so the weight of the converter is also relevant in order to compare different designs. The power-tomass ratio (g), defined by Eq. [9.3], indicates the level of heaviness of the WECS. The total converter active mass (MassTotal) is calculated via summation of the individual masses (Mass(i)) of the components. P P Pin Ploss;ðiÞ Pin Ploss;ðiÞ g¼ ¼ P MassTotal MassðiÞ

[9.3]

This chapter describes a simple procedure to calculate h, r and g of WECS for offshore WTs via the calculation of power losses, volume and mass of the main components: the power electronics valves, the magnetic components (AC and DC filter inductors) and the capacitors (DC link capacitors and AC filter capacitors). A base topology known as the two-level voltage source converter (2L-VSC) has been considered in order to illustrate the evaluation procedure. The nominal h, r and g are obtained for a set of design parameters and constraints. Even more, the her Pareto-front and the reg Pareto-front are considered in order to compare different parameters of design.

9.2

Semiconductors and switch valves

9.2.1

Semiconductor power losses

The power semiconductor devices (PSDs) used in WECS are operated as switches, taking on two possible static states, conducting or blocking, and two possible transition states, turn-on action or turn-off action. In any of these states, one energy dissipation component is generated, which heats the semiconductor and adds to the total power dissipation of the switch. Fig. 9.1 shows the simplified device switching waveforms (voltage and current) and the power losses associated with each possible operation state of the power switch.

9.2.1.1

Conduction loss

Static loss is determined by the non-linear voltageecurrent characteristic of the PSD. A typical voltageecurrent characteristic of a PSD is shown in Fig. 9.2. The two static states or regions can be noted from Fig. 9.2, blocking state (vsw < 0) is characterized by a very small current compared with the nominal current of the device, while the conducting state characteristic (vsw > 0) is branded by small voltage (comparing with nominal blocking voltage Vbk) for currents smaller than the maximum permitted current of the device.

V,I Switching waveforms

νsw

isw t

Simplified power loss estimation

Tsw P

Eoff

Eon

Turn-off action

Turn-on action

Blocking

Transition states

Eoff

t

Conducting

Static states

Figure 9.1 Possible states of a semiconductor power switch with the simplified device switching waveforms and its power loss estimation.

Isw Maximum current

Maximum blocking voltage Vsw

Blocking state

Conducting state

Figure 9.2 Typical voltageecurrent characteristic of a semiconductor device.

196

Offshore Wind Farms

The instantaneous power loss of the semiconductor at static state can be calculated as the product of the device voltage (vsw) and current (isw). When the switch is in blocking state, the device leaks a very small current (thousands or millions of times smaller than the nominal current) for any voltage lower than the nominal Vbk of the PSD, and the blocking loss is only contributing to a minor share of the total power dissipation; therefore this loss may be neglected in PSDs (Xu, 2002). In fact, the voltageecurrent characteristic in the blocking region is usually not reported by PSD manufacturers. The conduction loss (Pcond) can be calculated as (Mirjafari and Balog, 2014): 2 Pcond ¼ Vsw0 Tj;AVG $Isw;AVG þ RC Tj;AVG $Isw;RMS

[9.4]

where Tj,AVG is the average junction temperature, Isw,AVG is the average current, and Isw,RMS is the RMS current the device is conducting in a period. The values of Isw,AVG and Isw,RMS can be calculated based on the input/output current of converter and are mainly dependent on the converter topology, modulation strategy and input/output power factor (examples of calculation are presented in Section 9.5). The parameters Vsw0 (threshold voltage) and RC (on-state resistance) can be calculated using the datasheet for each device. As an example, the characteristics describing the relationships between voltage and current of the IGBT and diode as given in the datasheet of the power module Infineon FZ1500R33HE3 (Infineon, 2013) are shown in Fig. 9.3. The voltageecurrent characteristic is dependent on the junction temperature (Tj) as shown in Fig. 9.3, where two different Tj values are considered for each device. To describe the temperature dependency of the curve the parameters Vsw0 and RC can be made temperature-dependent. The order of the approximation will depend on the availability of curve data at different operating temperatures. Normally, the datasheet includes data of two or three operating temperatures, so a linear approximation can be done: Vsw0 ðTj Þ ¼ Vsw00 $ð1 þ f Vsw0 $ðTj Tj0 ÞÞ

[9.5]

RC ðTj Þ ¼ RC0 $ð1 þ f Rc ðTj Tj0 ÞÞ

[9.6]

where fVsw0 and fRc are the temperature coefficients of Vsw0 and RC, respectively; Tj0 is a fixed reference Tj; Vsw00 and RC0 are the Vsw0 and RC at temperature Tj0.

9.2.1.2

Switching losses

On the other hand, the transition losses, also called switching losses (Psw), are calculated based on the energy dissipated during commutation, Esw,on and Esw,off for turn-on and turn-off, respectively. This commutation energy loss (Esw) mainly depends on the voltage in the PSD at the moment before turn-on action (vswb) or after turn-off action (vswa), the

Modelling of power electronic components IGBT

4

4

3.5

3.5

3

2.5

2.5

2

1.5

1.5

2000 1000 Current (A)

Diode data at 125°C Linear approx. at 125°C Diode data at 150°C Linear approx. at 150°C

3

2

1 0

Diode

4.5

IGBT data at 125°C Linear approx. at 125°C IGBT data at 150°C Linear approx. at 150°C

Voltage (V)

Voltage (V)

4.5

197

3000

1

0

1000 2000 Current (A)

3000

Figure 9.3 Currentevoltage conduction characteristic of the power module Infineon FZ1500R33HE3 for two different junction temperatures 125 and 150 C.

current through the PSD at moment after turn-on action (iswa) or before turn-off action (iswb). In order to model these dependencies, the following model is proposed to calculate Esw using the data commonly available in the datasheets of the devices: Esw;on ¼ Esw0;on $ð1 þ f Eon $ðTj Tj0 ÞÞ

[9.7]

Esw0;on ¼ vswb $ KEon0 þ KEon1 $iswa þ KEon2 $iswa 2

[9.8]

Esw;off ¼ Esw0;off $ð1 þ f Eoff $ðTj Tj0 ÞÞ

[9.9]

Esw0;off ¼ vswa $ KEoff0 þ KEoff1 $iswb þ KEoff2 $iswb 2

[9.10]

where fEon=off is the temperature coefficient of Esw,on/off; Tj0 is a fixed reference temperature; Esw0,on/off is the Esw,on/off at temperature Tj0; and KEðon=offÞ0, KEðon=offÞ1, KEðon=offÞ2 are the polynomial regression coefficients used to describe the current dependency of Esw,on/off. All these parameters can be calculated using the datasheet of the device. As an example, the characteristics describing the relationship between Esw and current of the IGBT and diode as given in the datasheet of the power module Infineon FZ1500R33HE3 (Infineon, 2013) are shown in Fig. 9.4. In the case of the diode (right side of Fig. 9.4), Esw,on is not plotted, as the manufacturer usually does not report this

198

Offshore Wind Farms

9000 8000

IGBT

2400

Eon at 150°C Eon at 125°C Eoff at 150°C Eoff at 125°C

2200 2000 Commutation energy (mJ)

Commutation energy (mJ)

7000 6000 5000 4000 3000

1800 1600 1400 1200

2000

1000

1000

800

0 0

Diode Eoff at 150°C Eoff at 125°C

500

1000 1500 2000 2500 3000 Current (A)

600 0

500

1000

1500 2000 2500 3000 Current (A)

Figure 9.4 Current-dependent commutation energy loss for turn-on and turn-off action of the IGBT and diode of the power module Infineon FZ1500R33HE3 as given in the datasheet for a blocking voltage of 1800 V and two different junction temperatures 125 and 150 C. The solid lines shows the second-order polynomial fitting curve as given in Eqs [9.8] and [9.10].

characteristic since the switching losses at turn-on of power diodes are very small and therefore neglected. Once the model of commutation energy is defined, the average switching loss for turn-on and turn-off can be expressed as, Psw;on ¼

N 1 X $ Esw;on T j¼1

[9.11]

Psw;off ¼

N 1 X $ Esw;off T j¼1

[9.12]

where N is the number of switching actions in a fundamental period T, and can be expressed as a function of the switching frequency ( fsw) or the switching period (Tsw), N¼

T ¼ T$fsw Tsw

[9.13]

Modelling of power electronic components

199

For applications with Tsw much lower than T, assuming a constant operating Tj (Tj,AVG) during T, and noting that vswb can be approximated to Vbk of the application, the following simplification is suggested: 2 Psw;on ¼ fsw $Vbk $ KEon0 þ KEon1 $Iswa;AVG þ KEon2 $Iswa;RMS [9.14] 1 þ f Eon Tj;AVG Tj0 where Iswa,AVG and Iswa,RMS are the average and RMS values of iswa, respectively. The same procedure can be applied to derive an expression for Psw,off 2 Psw;off ¼ fsw $Vbk $ KEoff0 þ KEoff1 $Iswb;AVG þ KEoff2 $Iswb;RMS [9.15] 1 þ f Eoff Tj;AVG Tj0 where Iswb,AVG and Iswb,RMS are the average and RMS values of iswb, respectively. Iswa,AVG, Iswa,RMS, Iswb,AVG and Iswb,RMS can be calculated based on the input/ output current of the converter and they are mainly dependent on the converter topology, modulation strategy and input/output power factor. Examples of calculation are presented in Section 9.5.

9.2.2

Parallel connection of power modules

Offshore WECS are required to manage high power ratings. When a low/medium voltage feeds the converter then a high current should be managed by the power switches. In this case the parallel connection of PSDs is considered to fulfil the current requirement. The number of parallel-connected PSDs (np) has no limitation (Fuji Electric Co., 2011). However, some disadvantages like current imbalance between the modules at static and transition states are inherent to the parallel connection mainly because the connected PSDs do not have identical properties. The difference in the voltageecurrent characteristic of the modules is a major cause of current imbalances. The static voltage deviation (DVsw) of two modules with the same production reference is given by small variations in the module properties from the fabrication processes or by Tj differences between modules. Fig. 9.5 shows how a difference in the static characteristic of two semiconductors connected in parallel can cause a current imbalance. Using a linear relation to model the voltageecurrent characteristic of each module and for simplicity, assuming equal threshold voltage for the modules (Vsw0,1 ¼ Vsw0,2), the average current in the parallel connection (Ip,AVG) can be expressed by Ip;AVG

isw;1 Rc;1 þ Rc;2 $ ¼ 2 Rc;2

[9.16]

where Rc,1 and Rc,2 are the RC of the two modules. When Rc,1 < Rc,2, module one has a higher current (isw,1 > isw,2) given by DVsw.

200

Offshore Wind Farms

Vsw

ΔVsw,1 Vp

ΔVsw,2

Isw Isw,2

Isw,1

Figure 9.5 Example of current imbalance given by the differences in voltageecurrent characteristic of two modules parallel-connected.

The current imbalance rate (dCI), which represents the ratio of shared current in the parallel connection, is defined as, dCI ¼

isw;1 Ip;AVG

! 1

[9.17]

Fig. 9.6 shows examples of the representative relationship between DVsw and dCI. Three IGBT power modules from Infineon, FZ750R65KE3 (6500 V 750 A), FZ2400R17HP4 (1700 V 2400 A), and FZ1500R33HE3 (3300 V 1500 A), are presented in Fig. 9.6. Then, for example, if two semiconductor modules of FZ1500R33HE3 are parallel connected and it is expected that DVsw is 1[V] at maximum current, then dCI is 15%, meaning that the current through one of the switches is 15% higher than Ip,AVG. Therefore, a decrease in total current (derating) that the parallel connection may conduct is needed in order not to exceed the maximum current of any semiconductor in the parallel connection. High-power semiconductor manufacturers recommend that the peak current of a switch does not excess 80% of the maximum current of the semiconductor (Fuji Electric Co., 2011). Following this recommendation and taking into account the dCI, when np modules are connected in parallel, the following expression must be satisfied: Ipsw 1:6$In $np $kcdp

[9.18]

Modelling of power electronic components 25

201

FZ1500R33HE3 FZ750R65KE3 FZ2400R17HP4

Current imbalance rate (%)

20

15

10

5

0

0

0.2

0.4

0.6

0.8 1 Voltage deviation (V)

1.2

1.4

1.6

Figure 9.6 Example of current imbalance ratio as a function of the voltage deviation for three different IGBT modules for Infineon manufacturer.

kcdp ¼

1 ð1 dCI Þ $ 1 þ ðnp 1Þ$ np ð1 þ dCI Þ

[9.19]

where Ipsw is the peak current of the parallel connection, In is the nominal current of one PSD (normally half of the maximum current), and kcdp is the derating factor, which is the decrease in the maximum total current that can be applied under the worst-case conditions where np 1 modules are identical and the current imbalance is concentrated on the singled-out module, whose RC is the smallest (Fuji Electric Co., 2011). An example of kcdp as a function of np for different DVsw is shown in Fig. 9.7. It can be noted that as np increases then kcdp will decrease. Also, an increase in DVsw will cause kcdp to decrease, since dCI increases. Additionally, current imbalance in the parallel connection can be taken into account for power loss calculations, since different currents in the modules will generate different losses. Then, power losses in the parallel connection can be calculated from the power losses in a module with an equivalent current and then multiplying these losses by np. It is proposed to estimate the equivalent current (Isw,eq) for semiconductor power loss calculations by Eq. [9.20], where Isw,Total is the total current of the array of IGBT modules parallel connected. Isw;eq ¼

ð1 þ dCI =2Þ $Isw;Total np

[9.20]

202

Offshore Wind Farms 100

95

Derating factor Kcdp (%)

90 ΔVsw increases

ΔVsw = 0.6(V)

85

80

75

70

65

ΔVsw = 1.6(V) 1

2

3

4

5

6

7

8

9

10

Number of modules parallel connected (np)

Figure 9.7 Example of derating factor Kcdp as a function of the number of parallel connected modules for different voltage deviations when power modules Infineon FZ1500R33HE3 are considered.

9.2.3

Series connection of power modules

If the PSDs are series-connected, then all devices carry out the same current in conduction state, but some differences in forward voltage can be presented due to non-identical voltageecurrent characteristic of the modules. For the same reason, in blocking state, voltage imbalance between modules in the series connection could exist. Since manufacturers do not normally include voltageecurrent characteristic in blocking state, it could be useless to introduce calculations for voltage imbalance. However, the considered power loss models are proportional to the forward voltage and Vbk; therefore voltage imbalance can be neglected in the power loss calculation and the average Vbk (total voltage divided by number of devices series connected, ns) can be used for calculation purposes. In order to estimate the required ns for a specific application, and considering the application notes from semiconductor manufacturers (Volke and Hornkamp, 2012; Fuji Electric Co., 2011), the following requirement must be satisfied (without considering voltage imbalance for series connection of non-identical devices): VP;max kvp $Vblock ns

and

VDC;max kvdc $Vblock ns

[9.21]

where VP,max is the maximum voltage amplitude to be blocked for the series connection array, kvp is a safety factor for peak voltage (normally between 0.75 and 0.85), VDC,max is the maximum DC voltage of the array, kvdc is a safety factor for DC


203

voltage (normally between 0.6 and 0.7) and Vblock is the rated Vbk of a single device. The safety factors kvp and kvdc are considered in order not to exceed Vblock by the repetitive overshoot voltage spikes during turn-off of the device and to guarantee that the device is switched in its safe operating area (Backlund et al., 2009).

9.2.4

Volume and mass of a switch valve

The thermal model of the PSD can be used to calculate the required thermal resistance of the cooling system for worst-case operating conditions, and then the size and weight of the power switch with the cooling system can be estimated. Two main types of cooling system are used in high-power applications: forced air cooling and liquid cooling. However, only the forced air cooling system is presented in this chapter. The volume of a power switch valve (Volvalve) is given by np or ns, the semiconductor module itself and by the module heat sink (HS): Volvalve ¼ np $ns $ðVolmod þ VolHS Þ

[9.22]

The semiconductor module volume (Volmod) can be found in the datasheet of the PSD, the HS volume (VolHS) is given by the volume of aluminium/copper structure (VolHSal) and the fan volume (Volfan). The VolHSal is inversely proportional to the thermal resistance of the HS (RthHS) for a given fan velocity (Vfan) (Wen et al., 2011), and Volfan is proportional to VolHSal. The forced air HS is designed based on a fixed Vfan, and using the definition of thermal resistance (Rth[K/W] ¼ DT/Ploss), the following model is proposed: VolHSal ¼ KHS0 $

1 RthHS

KHS1

¼ KHS0 $

Volfan ¼ Kfan0 $ðVolHSal Kfan2 ÞKfan1

Ploss;mod DTHS;max

KHS1

[9.23] [9.24]

where DTHS,max is the maximum allowable HS to ambient temperature, Ploss,mod is the total power loss of the power module, and the parameters KHS0, KHS1, Kfan0, Kfan1 and Kfan2 are proportionality regression coefficients, whose values can be found by taking data of HS and fans available in the market. Fig. 9.8 shows an example of the relationship between VolHSal and RthHS for the bonded fin HS series BF-XX from DAU. Data for three different Vfan are presented in Fig. 9.8. Parameters KHS0 and KHS1 are dependent on Vfan, since Rth of a given aluminium structure depends on Vfan, as is shown in Fig. 9.9 for four HS structures in the series. Table 9.1 shows the calculated values of parameters KHS0 and KHS1 for different Vfan. An example of the correlation between Volfan and VolHSal is presented in Fig. 9.10, where the family of axial fans in SEMIKRON series SKF-3XX and the bonded fin HSs series BF-XX from DAU are considered. The calculated regression coefficients for the example presented in Fig. 9.10 are Kfan0 ¼ 0.1992, Kfan1 ¼ 0.7467 and Kfan2 ¼ 0.1966 (dm3).

204

Offshore Wind Farms Bonded fin heat sinks, DAU series BF-XXX

10

1(m/s) 3(m/s) 5(m/s)

9

Aluminium structure volume (dm3)

8 7 6 5 4 3 2 1 0 0

0.05

0.1

0.15

0.2

0.25

0.3

Thermal resistance, Rth (K/W)

0.35

Figure 9.8 Example of relationship between aluminium structure volume and the thermal resistance of the heat sink for the bonded fin heat sinks series BF-XX from DAU manufacturer. The dash lines show the fitting curve as given in Eq. [9.23].

Bonded fin heat sinks, DAU series BF-XXX BF-102 BF-151 BF-176 BF-320

Thermal resistance (K/W)

0.25

0.2

0.15

0.1

0.05

0 1

2

3

4

5

6 Fan velocity (m/s)

7

8

9

10

11

Figure 9.9 Example of the relationship between the thermal resistance of the heat sink and the fan velocity for the bonded fin heat sinks series BF-XX from DAU manufacturer.


205

Table 9.1 Calculated regression coefficients for the proposed heat sink aluminium structure volume model at different fan velocities. Bonded fin heat sinks series BF-XX from DAU manufacturer Fan velocity (m/s)

KHS0 (dm3)

KHS1

1

56.19e-3

1.8311

3

21.72e-3

1.7415

5

16.78e-3

1.6539

10

9.322e-3

1.4321

Axial fans SEMIKRON series SKF-3XX

6

5

Fan volume (dm3)

4

3

2

1

0 0

1

2

3

4

5

6

Aluminium structure volume (dm3)

7

8

9

10

Figure 9.10 Example of relationship between fan volume and aluminium structure volume for the axial fans SEMIKRON series SKF-3XX and the bonded fin heat sinks series BF-XX from DAU manufacturer. The dash line shows the fitting curve as given in Eq. [9.24].

DTHS,max is calculated on the basis of average thermal analysis of the power module. For example, if a power module composed by IGBT and antiparallel diode is considered, the average thermal model presented in Fig. 9.11 is considered to calculate DTHS,max and the following model can be used:

DTHS;max þ Tamb

Pigbt Pdiode ¼ KSFT $Tj;max max Rth;igbt $ ; Rth;diode $ Nisxm Nisxm

[9.25]

206

Offshore Wind Farms

RthJC,T

+

Pigbt

RthCH,T

RthJC,D

+

Pdiode

ΔTHS

–

ΔTIGBT

+

–

RthCH,D

ΔTdiode

–

Tamb

Figure 9.11 Average thermal model of an IGBT power module.

Pigbt ¼ Pcond;igbt þ Psw;on;igbt þ Psw;off;igbt

[9.26]

Pdiode ¼ Pcond;diode þ Psw;off;diode

[9.27]

where KSFT is the safety factor of thermal design, Tamb is the ambient temperature, Nisxm is the number of internal IGBT/diode per module, Rth,igbt and Rth,diode are the Rth of junction-to-HS per IGBT and diode, respectively, which can be calculated by adding the junction-to-case and the case-to-HS thermal resistances (RthJC and RthCH in Fig. 9.11) of IGTB and diode, given in the datasheet of the PSD. In order to guarantee realistic HS designs, a constraint related with the minimum Rth (RthHS,min) should be taken into account, which can be defined by the maximum ratio of VolHSal to Volmod (dHS,max ¼ VolHSal/Volmod) beyond which Eq. [9.23] is not valid (normally, dHS,max 6). Then, the temperature rise of the HS with RthHS,min should be less than or equal to DTHS,max (from Eq. [9.25]), which establishes the maximum power that the module can dissipate mounting in the HS without overheating itself, and can be expressed as follows: RthHS;min $ðPigbt þ Pdiode Þ DTHS;max RthHS;min ¼

KHS0 dHS;max $Volmod

[9.28]

1 KHS1

[9.29]

Since conduction losses and switching losses are dependent on the Tj, some iteration could be needed to solve Eq. [9.25]. Alternatively, Eqs [9.4], [9.14] and [9.15] can be set up as temperature-independent for the highest acceptable Tj and design the cooling system to keep the Tj below the highest temperature assumed. This assumption gives some thermal safety margin built into the design (Drofenik and Kolar, 2005) and is considered in the examples presented in this chapter.


207

Finally, the mass of the valve can be expressed by the density and volume of each element (power semiconductor module, HS and fan): Massvalve ¼ np $ns $ðrmod $Volmod þ ral $VolHSal þ rfan $Volfan Þ

[9.30]

The density values (rmod, ral and rfan) can be calculated from the reference datasheet for each element. For example, the power module Infineon FZ1500R33HE3 has a density value rmod ¼ 1187.2 [kg/m3], and for the HS and fan considered in Fig. 9.8 and Fig. 9.10, the calculated density values are ral ¼ 1366 [kg/m3] and rfan ¼ 769.23 [kg/m3], respectively.

9.2.5

Semiconductor parameters

h and r of the converter are highly influenced by the type of PSD selected to realize the high-power switch valve. PSDs commonly used in WECS are the insulated gate bipolar transistor (IGBT), the integrated gate commutated thyristor (IGCT) and the injection enhanced gate transistor (IEGT) (Lee et al., 2014). Since analysis of each type of PSD is beyond the scope of this chapter, only switch valves based on IGBT devices are considered in the application examples of Section 9.5. However, the models presented in this section can be adapted to any of the three devices (IGBT, IGCT or IEGT). Table 9.2 presents a summary of the semiconductor parameters needed in the switch valve models presented in this chapter. Additionally, it includes parameter values of three IGBT modules with different ratings. These three modules are taken into account in Section 9.5 for evaluation of power losses, volume and mass of a 2L-VSC

9.3

Filter inductors

9.3.1

Main constraints in the inductor design

The starting point for an inductor design is the stored energy relation (Eq. [9.31]) defined by the inductance L, the inductor peak current ( IbL ), the RMS current (IL), the RMS current density (JL), the peak flux density (BL), the winding conductor fill factor (kwc), the winding window area (Aw) and the core area (Acore) (Mohan et al., 2003). L$ IbL $IL ¼ kwc $JL $BL $Aw $Acore

[9.31]

The product (Aw$Acore) appears in Eq. [9.31] and is an indication of the core size and it is called area product. For a given core material, the peak flux density is limited by the saturation flux density (Bsat). A similar situation is given for the winding conductor, which physically limits the maximum current density (Jmax) of the inductor winding. Even more, if a conductor type and core type are fixed for the inductor design, the winding conductor fill factor becomes approximately a design constant.

208

Semiconductor Parameters; IGBT modules from Infineon manufacturer for three different voltage ratings are considered.

Table 9.2

Ratings Reference

(1700Vx3600A)

(3300Vx1500A)

(6500Vx750A)

FZ3600R17KE3

FZ1500R33HE3

FZ750R65KE3

1.0108

1.0108

1.2768

1.5

1.2

1.4

3

Volume [dm ] Mass [Kg]

Diode

IGBT

Diode

IGBT

Diode

Conduction losses model

0.964

0.959

1.436

1.359

1.890

1.412

-0.886

-1.36

-0.237

-3.193

0.582

-1.821

0.401

0.249

1.130

0.804

2.326

1.862

3.468

2.542

3.257

0.623

3.178

1.339

125

125

150

150

125

125

KEon0 [mJ/V]

0.158

–

0.489

–

0.235

–

KEon1 [mJ/VA]

0.064

–

0.039

–

1.306

–

0.034

–

0.463

–

1.119

–

3.101

–

2.759

–

3.538

–

Vsw00 [V] fVsw0

[1/ C]

-3

(x10 )

RC0 [mU] fRc Tj0 Switching losses model (ON)

[1/ C]

-3

(x10 )

[ C]*

2

KEon2 [nJ/VA ] fEon

[1/ C]

-3

(x10 )

Offshore Wind Farms

IGBT

General Information

KEoff 0 [mJ/V]

0.037

0.328

0.116

0.314

0.021

0.185

KEoff 1 [mJ/VA]

0.404

0.334

0.731

0.728

1.546

1.118

0.008

-0.027

0.045

-0.135

0.014

-0.326

3.276

4.25

2.435

5.333

1.429

5.333

0.45

0.40

0.55

0.75

0.4

0.5

dCI [%] at Tj,max

18.48

28.72

19.36

26.16

12.95

21.81

RthJC [K/kW]

6.3

14

7.35

13

8.7

18.5

RthCH [K/kW]

8.7

19.5

10

11

8.8

14

2

KEoff 2 [nJ/VA ] fEoff Parallel connection Static Thermal model

[1/ C]

DVsw [V] at

-3

(x10 )

25[ C]

Nisxm Tj,max Switching times

[ C]

3

3

3

125

150

125

ton,max at Tj,max [ms]

1.05

–

1.15

–

1.2

–

toff,max at Tj,max [ms]

2.1

0.88

3.85

1.73

8.1

2.67

fsw,max** [kHz]

4.96 / 4

2.97 / 2


Switching losses model (OFF)

1.67 / 1.5

* Reference temperature for all the temperature coefficients. **Maximum switching frequency is calculated as the conduction time per switching period is higher than 98%.

209

210

Offshore Wind Farms

When a reduction in the total size of the inductor is targeted for some given electrical parameters, like inductance and current, then the designer should deal with this by pushing the peak flux density and current density as close as possible to the physical limits and thermal constraints, as the product JL$BL is inversely proportional to the winding area and the core area. Then, the minimum area product for a given set of constraints can be written as: Aw $Acore f

9.3.2

L$ IbL $IL f EL f L$IL 2 Jmax $Bsat

[9.32]

Size modelling

Geometrically, it can be shown that the area product is also related to the volume of the inductor by (Mohan et al., 2003): 4=3

Aw $Acore f VolL

[9.33]

Combining Eqs [9.33] and [9.32], the overall inductor volume (VolL) can be expressed as: 3=4 VolL f L$IL 2

[9.34]

Then, if an inductor design technology is kept (core material, conductor type, core geometry, etc.) for different inductance values and current requirements, it is proposed to predict VolL and inductor total mass (MassL) by: KVL1 VolL ¼ KVL0 $ L$IL 2

[9.35]

MassL ¼ KrL0 $ðVolL ÞKrL1

[9.36]

where KVL0, KVL1, KrL0 and KrL1 are proportionality regression coefficients found by taking data from the reference inductor technology. Fig. 9.12 presents the relationship between the inductor volume and the product L$IL 2 for three different inductor technologies from Siemens. On the other hand, Fig. 9.13 displays MassL against VolL for the same families of inductors as in Fig. 9.12. The calculated parameters of the size and mass models for the inductors considered in Figs 9.12 and 9.13 are presented in Table 9.3.

9.3.3

Winding losses

The inductor power losses (PL) are divided into winding losses (PwL) and core losses (PcoreL). Since the main use of the inductors in power converters is to filter the current in order to limit the peak-to-peak ripple current (DILh), then it is expected that the inductor current has harmonic components and these harmonics cannot be neglected


211

3AC-4EUXX-Cu 3AC-4EUXX-AI DC-4ETXX-Cu

Overall inductor volume (m3)

100

10–1

10–2

10–3 0 10

101

102

L*IL2 (J)

103

104

105

Figure 9.12 Example of inductor volume and product (L$IL 2 ) relationship for three different inductor technologies from Siemens. Three-phase reactors series 4EUXX with Cu and Al winding conductor, and DC iron core smoothing reactors series 4ETXX with Cu winding are considered. The lines shows the calculated model based on Eq. [9.35] for each family of considered inductors.

104

3AC-4EUXX-Cu 3AC-4EUXX-AI DC-4ETXX-Cu Model 3AC-4EUXX-Cu Model 3AC-4EUXX-AI Model DC-4ETXX-Cu

Total mass (kg)

103

102

101 10–2

10–1 Inductor volume (m3)

100

Figure 9.13 Inductor total mass against overall volume. Three inductor technologies from Siemens are plotted: Three-phase reactors series 4EUXX with Cu and Al winding conductor, and DC iron core smoothing reactors series 4ETXX with Cu winding.

212

Offshore Wind Farms

Parameters of Inductor model; inductor technologies from Siemens manufacturer are considered: three-phase reactors series 4EUXX with Cu and Al winding conductor, and DC iron core smoothing reactors series 4ETXX with Cu winding

Table 9.3

Parameter

3-AC Inductors

DC-Inductors

Reference

Series 4EUXX

Series 4ETXX

Conductor material

Copper

Aluminium

Copper

KVL0

3.4353e-3

2.2818e-3

0.60434e-3

KVL1

0.6865

0.82494

0.80946

KrL0

4129.2244

2276.9539

2797.6215

KrL1

1.0768

0.94879

0.99314

Krw0

9412.0118

6005.6682

10,874.8628

Krw1

0.85361

0.75117

0.82048

fLref

50

50

50

Krc0

8242.2998

8805.6895

493.0059

Krc1

0.99926

0.97691

1.0349

aL

1.1

1.1

1.1

bL

2.0

2.0

2.0

diLref

e

e

0.3

in the calculation of PwL. Fig. 9.14 shows the typical inductor current waveform in power converter applications and its decomposition into the two main components, its fundamental component (iL1) and its ripple component (iLh). To estimate winding and core losses in the inductor (Barrera-Cardenas and Molinas, 2015), it is proposed to approximate the ripple current to be a triangular ΔILh iL1 iL

≈ iLh

iLh iL = iL1 + iLh

ΔILh

≈

ΔILh

Figure 9.14 Typical inductor current waveform in power converter applications and its decomposition into the two main components, the fundamental component and the harmonic component.


213

waveform with maximum amplitude equal to the maximum current ripple in order to simplify the calculations. Additionally, the concept of loss of power density in the winding and core is used to express the winding losses as a function of the electrical parameters and reference inductor technology parameters, as follows: "

PwL

!# 2 d2iL 2 4 fsw 2 1 fL1 2 þ þ $ ¼ 1þ $ $ $Krw0 $ðVolL ÞKrw1 $ 3 p2 3 3 fL1 6 fLref [9.37]

DILh diL ¼ pffiffiffi 2$IL1

[9.38]

where Krw0 and Krw1 are proportionality regression coefficients found by taking data from reference inductor technology, diL is the ratio of peak-to-peak current ripple to maximum fundamental nominal current, fL1 is the fundamental frequency and fLref is a reference frequency for winding losses, which can be found from the datasheets. Fig. 9.15 presents the relationship between PwL and VolL for three different inductor technologies. The calculated parameters of the models for the inductors considered in Fig. 9.15 are presented in Table 9.3. It should be noted that Eq. [9.37] is valid for inductor current with fundamental component different to DC component ( fL1 > 0). When the DC current is the main

Winding losses (W)

104


103

102

10–2


100

Figure 9.15 Example of inductor winding losses and overall volume relationship for three different inductor technologies from Siemens. Three-phase reactors series 4EUXX with Cu and Al winding conductor, and DC iron core smoothing reactors series 4ETXX with Cu winding are considered. The lines shows the calculated model based on Eq. [9.37].

214

Offshore Wind Farms

component of the inductor current, assuming that the winding design is optimized for low frequencies with an effective frequency fL0 < 50 Hz, the following expression, proposed in Barrera-Cardenas and Molinas (2015) can be used: !# 2 d2iL 6 fsw $Krw0 $ðVolL ÞKrw1 ¼ 1þ 1þ 2$ $ p fL0 12 "

PwL

9.3.4

[9.39]

Core losses

The core power loss density (pcL) can be approximated using the empirical Steinmetz equation: pcL ¼

dPcoreL aL ¼ Kcore $feff $BL bL dVcL

[9.40]

where Kcore, aL and bL are the usual Steinmetz coefficients, which are related to the core material, BL is the peak flux density, and feff is the effective frequency for a non-sinusoidal current waveform (or to take into account harmonic effect in losses) (Sullivan, 1999), and it can be estimated using Eq. [9.41], where Ij is the RMS amplitude of the Fourier component at frequency wj.

2$p$feff

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

uP u j¼0.f I 2j $w2j rms dtd IðtÞ t P ¼ ¼ 2 Irms j¼0.f I j

[9.41]

The expression derived in Barrera-Cardenas and Molinas (2015) for estimation of PcoreL is considered in this chapter, which is based on the core power loss density concept: 0 B PcoreL ¼ @

6

1aL

2 2 d $f þ iLfL1 sw C A 6 þ d2iL

di bL pcL1 $ 1þ L $ $Krc0 $ðVolL ÞKrc1 2 pcL

[9.42]

where Krc0 and Krc1 are proportionality regression coefficients found by taking data from reference inductor technology and pcL* is the reference power loss density for reference inductor technology. Assuming that pcL* is given for a reference frequency ( fLref) and a reference flux density (BLref), and that the inductor design has been optimized following the optimization criterion for minimum losses and the optimum flux density method presented in Hurley et al. (1998), the reference frequency and flux density are related as: ð fLref ÞaL þ2 $ðBLref ÞbL þ2 ¼ KLopt

[9.43]


215

where KLopt* is a constant given by the inductor design parameters. Then, the reference inductor technology is used to design an inductor for fL1 higher than fLref, the fundamental flux density (BL1) should be varied according to Eq. [9.43], therefore: fLref aL þ2 BL1 bL þ2 ¼ fL1 BLref

[9.44]

then the ratio ð pcL1 =pcL Þ can be simplified as follows: pcL1 ¼ pcL

fL1 fLref

aL BL1 bL fL1 2ðaLbL Þ $ ¼ BLref fLref

[9.45]

However, if fL1 is lower than fLref, then BL1 is assumed to be constant (because of magnetic saturation), then the ratio ( pcL1/pcL*) can be simplified as follows: pcL1 ¼ pcL

fL1 fLref

aL BL1 bL fL1 aL $ ¼ BLref fLref

[9.46]

When the DC current is the main component of the inductor current, the above procedure can be modified to get an expression for PcoreL. In that case, it should be noted that the DC component does not produce core losses; therefore the reference data are given for a ratio of peak-to-peak current ripple to maximum DC current (diLref ) and a reference ripple frequency ( fLref). Then, the following expression can be derived: PcoreL ¼

pffiffiffi aL 2$ 3$fsw d i L bL $ $Krc0 $ðVolL ÞKrc1 p$fLref diLref

[9.47]

Fig. 9.16 presents the relationship between PcoreL and VolL for three different inductor technologies. The calculated parameters of this model for the inductors considered in Fig. 9.16 are presented in Table 9.3.

9.4 9.4.1

Filter capacitors Size modelling

Almost all conventional capacitors used in WECS are constructed based on plate capacitor structure. The volume of a plate capacitor (VolPC) is proportional to the area of the plates (APC) and the plate separation distance (dPC). Also, the voltage applied between plates is limited by dPC and the properties of the dielectric material (the breakdown electric strength, EBd). When an increase in the rated voltage (VCN) of the capacitor is desired, the dPC must be increased in order to avoid dielectric breakdown.

216

Offshore Wind Farms

103

Core losses (W)

102

101

100


10–2


100

Figure 9.16 Example of inductor core losses and overall volume relationship for three different inductor technologies from Siemens. Three-phase reactor series 4EUXX with Cu and Al winding conductor, and DC iron core smoothing reactors series 4ETXX with Cu winding are considered. The lines show the calculated model based on Eq. [9.42].

Taking into account the capacitance definition and the previous mentioned relations, it can be shown that the VolPC is proportional to its capacitance (C) and the square of its VCN, as follows: VolPC f APC $dPC

C$dPC C VCN 2 ¼ $ dPC f $ ε ε EBd

[9.48]

Then, if a capacitor technology is kept (dielectric material, fabrication process and geometry) for different values of C and VCN requirements, it is proposed to predict the overall capacitor volume (VolC) and capacitor total mass (MassC) by: KVC2 VolC ¼ KVC0 $ CKVC1 $VCN

[9.49]

MassC ¼ KrC0 $ðVolC ÞKrC1

[9.50]

where KVC0, KVC1, KVC2, KrC0 and KrC1 are proportionality regression coefficients found by taking data from reference capacitor technology. Fig. 9.17 presents the relationship between VolC and C for different VCN and different capacitor technologies. On the other hand, Fig. 9.18 displays MassC against VolC for the same families of capacitors considered in Fig. 9.17. The calculated parameters of the


(a)

217

DC-series MKP-B256XX, from TDK

5

(c)

700(V) 900(V) 1100(V) 1320(V) 1980(V)

1

0.5 10

2

(b)

Volume (dm3)

2

1

0.5

250(V) 330(V) 480(V)

103 Capacitance (μF)

101

DC-series LNK-M3XX1, from ICAR

(d) 3

3

2

2 700(V) 900(V) 1100(V) 1300(V) 1650(V) 1850(V)

1 102



103

AC-series MKV-E1X, from ICAR

5

3 Volume (dm )

Volume (dm3)

AC-series MKP B3236XX, from TDK

2

3

Volume (dm3)

3

900 1000 1400 1600 1800 2000

1 0.8 100


Figure 9.17 Film capacitor volume versus capacitance for different voltages ratings and capacitor technologies: (a) DC-link capacitors series MKP-B256xx from TDK; (b) DC-link capacitors series LNK-M3xx1 from ICAR; (c) AC filter capacitors series MKP B3236 from TDK; (d) AC filter capacitors series MKV-E1x from ICAR. The dash line shows the fitting curve as given in Eq. [9.49].

size and mass models for the capacitors considered in Figs 9.17 and 9.18 are presented in Table 9.4.

9.4.2

Capacitor dielectric losses

The capacitor losses (PC) are modelled as the summation of the dielectric losses (PεC) and the resistive losses (PUC). The dielectric losses can be calculated by (EPCOS, 2012): 2 PεC ¼ p$fc $C$tanðd0 Þ$VCac

[9.51]

where VCac is the maximum amplitude of the alternating voltage applied to the capacitor with effective frequency fc, and tan(d0) is the dissipation factor of the dielectric. Normally, the dielectric dissipation factor depends on the dielectric material of the capacitor and it can be considered constant for all capacitors in their normal working frequency range. For example, a typical value of the dissipation factor for polypropylene is 2e-4 ¼ 2 104 ¼ 0.0002, which is the dielectric of the capacitors considered in Figs 9.17 and 9.18. For DC capacitors, the voltage VCac is the peak value of the superimposed ripple voltage. When the ripple voltage is approximated to be a triangular waveform with

218

Offshore Wind Farms 4

3.5

3

TDK DC-series MKP-B256XX TDK AC-series MKP B3256XX ICAR DC-series LNK-M3XX1 ICAR AC-series MKV-E1X Approx. MKP-B256XX Approx. MKP-B3236XX Approx. LNK-M3XX1 Approx. MKV-E1X

Mass (kg)

2.5

2

1.5

1

0.5

0.5

1

1.5

2 Volume (dm3)

2.5

3

3.5

4

Figure 9.18 Film capacitor mass versus overall volume for the capacitor technologies considered in Fig. 9.17. The dashed line shows the fitting curve as given in Eq. [9.50].

Parameters of capacitor model. DC and AC capacitor technologies from TDK and ICAR are considered. From TDK manufacturer: the DC-link capacitors series MKP-B256xx and the AC filter capacitors series MKP B3236. From ICAR: the DC-link capacitors series LNK-M3xx1and the AC filter capacitors series MKV-E1x

Table 9.4

Parameter

DC capacitors

AC capacitors

Reference

TDK MKP-B256xx

ICAR LNK-M3xx1

TDK MKP B3236x

ICAR MKV-E1x

KVC0

2.0734e-5

5.9622e-5

67.303e-5

13.406e-5

KVC1

0.7290

0.7271

0.6770

0.5410

KVC2

1.3796

1.2473

1.0706

1.2216

KrC0

1.3428e3

0.8821e3

1.8079e3

2.7496e3

KrC1

1.0543

0.9950

1.0923

1.2060

tan(d0)

2e-4

2e-4

2e-4

2e-4

KUC0

4.069e-03

2.3056e-5

1.5711e-3

4.9369e-6

KUC1

0.3211

0.0430

0.3970

0.0783

KUC2

0.4661

0.4986

0.4539

1.0316


219

frequency equal to fsw of the converter and peak-to-peak amplitude equal to the maximum voltage ripple, the following approximation can be done based on Eq. [9.41]: PεC

pffiffiffi 3 2 $fsw $C$tanðd0 Þ$d2Vdc $VDC ¼ 2

[9.52]

where dVdc is the ratio of peak-to-peak voltage ripple to DC voltage (VDC) of the converter. For AC capacitors, the voltage VCac is the peak value of the fundamental component plus the peak voltage ripple, and the effective frequency can be approximated to the fundamental frequency, then dielectric losses can be expressed as:

PεC

dVac ¼ p$fc1 $C$tanðd0 Þ$ 1 þ 2

2

2 $Vacp

[9.53]

where dVac is the ratio of peak-to-peak voltage ripple to the peak fundamental voltage (Vacp) with fundamental frequency fc1.

9.4.3

Capacitor resistive losses

The resistive losses occur in the electrodes, in the contacting and in the inner wiring. These losses can be calculated as follows (EPCOS, 2012): PUC ¼ RsC $IC2

[9.54]

where IC is the RMS value of the capacitor current, and RsC is the series resistance at maximum hot-spot temperature, which is the sum of all resistance occurring inside the capacitor. The series resistance RsC is a typical estimated value based on average film metallization parameters (EPCOS, 2012) and its value can be found from the datasheet of the reference capacitor technology. Due to the low thickness of the metalized layer, high-frequency effects, such as skin effects, are negligible (Mirjafari and Balog, 2011). It is proposed to use a model based on C and VCN to predict the value of series resistance, as follows: KUC2 RsC ¼ KUC0 $CKUC1 $VCN

[9.55]

where KUC0, KUC1 and KUC2 are the proportionality regression coefficients found by taking data from reference capacitor technology. The value of series resistance given in the datasheets is referenced to 20 C capacitor temperature, but a conversion factor of 1.25 can be used to estimate the resistance at nominal temperature (typically, 85 C for film capacitors) (EPCOS, 2012). Fig. 9.19 presents the relationship between the series resistance per energy storage and the capacitance for different VCN and different capacitor technologies. The series resistance in Fig. 9.19 refers to the resistance at nominal temperature (85 C).

220

Offshore Wind Farms DC-series MKP-B256XX, from TDK 700 900 1100 1200 1320 1980

101

(c)

104

DC-series LNK-M3XX1, from ICAR 700 900 1100 1300 1650 1850

RsC / EC : (mΩ/J)

101

102 101


(d)


103

AC-series MKV-E1X, from ICAR

10

RsC / EC : (mΩ/J)

(b)

250 330 480

103

100 102

AC-series MKP B3236XX, from TDK

RsC / EC : (mΩ/J)

RsC / EC : (mΩ/J)

(a)

900 1000 1400 1600 1800 2000

3

100 102


102 0 10


Figure 9.19 Film capacitor series resistance per energy storage versus capacitance for different voltage ratings and capacitor technologies: (a) DC-link capacitors series MKP-B256xx from TDK; (b) DC-link capacitors series LNK-M3xx1 from ICAR; (c) AC filter capacitors series MKP B3236 from TDK; (d) AC filter capacitors series MKV-E1x from ICAR manufacturer. The dashed line shows the fitting curve as given in Eq. [9.55].

9.5

Evaluation approach and design methodology

Fully rated power converters, which are connected direct-in-line between the generator and the grid, have become the main choice of WECS in WTs for several reasons. They allow for a wide-ranging variable speed operation of the WT which increases the energy capture, as well as enabes the use of efficient permanent magnet generators. The relative ease of grid fault ride-through control is another reason. Therefore, full-scale power converters are also used increasingly in offshore wind power. Nowadays, two main topologies in commercial full-scale converters for offshore AC grid parks stand out from the literature, the two-level voltage source converter (2L-VSC) and the three-level neutral point clamped (3L-NPC) (Preindl and Bolognani, 2011; Chivite-Zabalza et al., 2013). The evaluation of h, r and g of a 2L-VSC has been chosen as the application example of the models introduced in this chapter. Fig. 9.20 shows the basic schematic of 2L-VSC with unidirectional switches implemented using IGBTs with antiparallel diodes. In what follows, the presentation will cover all the aspects needed to design the 2L-VSC and evaluate its h, r and g.


iin

221

iDC

iDC,h DC+ TC+

DB+

TB+

DA+ TA+

+

va

VDC

–

ib

vb

CDC

ic

vc

LF DC–

TC–

ia

TB–

DB–

DA– TA–

Figure 9.20 Two-level voltage source converter (2L-VSC) with IGBTs.

9.5.1

Modulation strategies

Selection of the appropriate modulation strategy has a high impact on the h of the converter and therefore at least two modulation strategies should be compared when the converter is designed. In this chapter, three modulation strategies are considered: sinusoidal PWM (SPWM), space-vector PWM (SVPWM) (also called suboptimal modulation) and a discontinuous modulation scheme known as symmetrical flat-top modulation (SFTM). Details of modulation strategy are beyond the scope of this chapter. More information about SPWM can be found in Kolar et al. (1990) and Mohan et al. (2003). Details about SVPWM and SFTM can be found in Helle (2007) and Kolar et al. (1990). In order to characterize a given modulation strategy, two main variables have to be evaluated: the modulation index (MS) and the relative turn-on time of the converter bridge legs (aa). Since there are different definitions of MS in the technical literature (Helle, 2007), in order to compare the three modulation method considered, it is convenient to define MS as: MS ¼

v A

VA;om

[9.56]

where vA is the desired inverter phase peak voltage, and VA,om is the maximum amplitude of the phase voltage reference at which the modulation approach enters the overmodulation range. This means that MS becomes unity at the overmodulation

222

Offshore Wind Farms

boundary. The value of VA,om depends on the modulation method, and for the modulation methods considered:

VA;om ¼

8 VDC > > > < 2

for SPWM

> VDC > > : pffiffiffi 3

for SVPWM and SFTM

[9.57]

From the definition in Eq. [9.56], the desired RMS inverter reference line-to-line voltage (VLL) is related to MS by: VLL ¼

Kmod

pffiffiffi 3$MS $Kmod $VDC

8 pffiffiffi > 2 > > > < 4 for SPWM ¼ > > 1 > > : pffiffiffi for SVPWM and SFTM 6

[9.58]

[9.59]

On the other hand, the phase modulation function should be defined for each modulation method in order to calculate aa, and in the case of a purely sinusoidal three-phase converter the output is postulated according to phase voltage (va ): rffiffiffi 2 va ðwtÞ ¼ $VLL $sinðwtÞ 3

[9.60]

then the phase modulation function (ma) is defined by Eqs [9.61], [9.62] and [9.63] for SPWM (Kolar et al., 1990), SVPWM (Helle, 2007) and SFTM (Helle, 2007) modulations, respectively. Once ma has been obtained, aa can be calculated according to Eq. [9.64] (Kolar et al., 1990). Fig. 9.21 shows aa for the modulation methods considered. ma;SPWM ðwtÞ ¼ MS $sinðwtÞ

[9.61]

2 ma;SVPWM ðwtÞ ¼ pffiffiffi $MS $sinðwtÞ þ v0 ; 3 8 1 va > > $ if jvc j > jva j < jvb j > > > 2 VA;OM > > > > > > jvb j < jvc j where v0 ¼ 2 V > A;OM > > > > > > > > 1 vc > > if jvb j > jvc j < jva j :2$V A;OM

[9.62]


223 α a for Ms = 1

(a)

SPWM SVPWM SFTM

1 0.8

αa

0.6 0.4 0.2 0 0

1

2

3 ωt (rad)

4

5

6

α a for Ms = 0.75

(b)

SPWM SVPWM SFTM

1 0.8

αa

0.6 0.4 0.2 0 0

1

2

3 ωt (rad)

4

5

6

Figure 9.21 The relative turn-on time of the voltage source inverter bridge legs (aa) for three different modulation methods: SPWM (sinusoidal PWM), SVPWM (space-vector PWM) and SFTM (symmetrical flat-top modulation).

2 ma;SFTM ðwtÞ ¼ pffiffiffi $MS $sinðwtÞ þ v0 ; 3 8 > va jva j > > if jvc j < jva j > jvb j > > > va VA;OM > > > > > > < jv j vb b if jva j < jvb j > jvc j where v0 ¼ > v V A;OM b > > > > > > > > vc jvc j > > > : vc VA;OM if jvb j < jvc j > jva j

[9.63]

1 aa ðwtÞ ¼ $ð1 þ ma ðwtÞÞ 2

[9.64]

9.5.2

Evaluation of the PSD currents

In order to calculate the power losses of each device (IGBT or diode), it is necessary to calculate the average and RMS values of the current that each device is conducting in a period (needed to evaluate Eq. [9.4]) and the average and RMS values of the current through the device at moment of turn-on/off action (needed to evaluate Eqs [9.8] and [9.10]). Unfortunately, there is no simple relationship between the RMS output current (Ia) and the current in the diode (Id) or IGBT (It). The current distribution between

224

Offshore Wind Farms

TA+

it+

DA+ id+ ia+

ia

ia– it–

TA–

id–

DA–

Figure 9.22 Bridge leg current definitions for the 2L-VSC.

diode (DAþ) and IGBT (TA) is a function of aa, MS and the displacement angle (4) between inverter phase voltage and current (Helle, 2007). In order to establish a relation between the output current and the diode and IGBT currents, Fig. 9.22 shows the current definition for the elements in a bridge leg of the 2L-VSC. From Fig. 9.22, the relation between diode current and IGBT current at any time is given by: ( ia ¼

itþ þ id

if

ia > 0

it þ idþ

if

ia 0

[9.65]

Since the currents itþ and id (or it and idþ) are orthogonal, it means that when the upper IGBT (TAþ) is conducting then the lower diode (DA) does not conduct and vice versa, then the average current through the IGBT and diode, respectively, can be evaluated as: It;AVG

1 ¼ 2p

Id;AVG

1 ¼ 2p

Z

pþ4

4

Z 4

ðaa ðqÞ$ia ðqÞÞ$dq

pþ4

ðð1 aa ðqÞÞ$ia ðqÞÞ$dq

[9.66]

[9.67]

Assuming a purely sinusoidal phase current with RMS value (Ia), in Kolar et al. (1990) Eqs [9.66] and [9.67] were evaluated for SPWM, and in Helle (2007) for SVPWM and SFTM. Comparing the results from Helle (2007) and Kolar et al. (1990),


225

it can be noted that the average current through the IGBT and diode is independent of the chosen modulation method and this is (Helle, 2007): pffiffiffi 2 Kmod þ $MS $cosð4Þ $Ia [9.68] It;AVG ¼ 2p 2 Id;AVG

pffiffiffi 2 Kmod $ MS $cosð4Þ $Ia ¼ 2p 2

[9.69]

In a similar way, the RMS current through the IGBT and diode can be obtained as: It;RMS

Id;RMS

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Z pþ4 1 ¼ aa ðqÞ$ðia ðqÞÞ2 $dq 2p 4

[9.70]

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Z pþ4 1 ¼ ð1 aa ðqÞÞ$ðia ðqÞÞ2 $dq 2p 4

[9.71]

Like the average current, the expression for the RMS current is also dependent of the selected modulation technique. Table 9.5 presents a closed form expression for the ratio of IGBT/diode RMS current to RMS phase current of the inverter, which are reported in Kolar et al. (1990) and Helle (2007). Expressions [9.68], [9.69] and those presented in Table 9.5 are valid whenever the active power flow is out of the inverter, as it is defined from Fig. 9.22, ie, the inverter operation mode. In the case of rectifier operation (power flows in the opposite direction), the expressions for the diode and IGBT have to be exchanged. On the other hand, in order to evaluate the switching losses, it is necessary to calculate the average and RMS values of a current waveform constructed using the current through the IGBT device sampled at the moment after each turn-on action (Ita,AVG and Ita,RMS), and the average and RMS values of current waveforms constructed using the current through the IGBT and diode devices at the moment before each turn-off action (Itb,AVG, Itb,RMS, Idb,AVG and Idb,RMS). Since the current through the upper IGBT commutates to the lower diode, it should be noted that the turn-on switching current of the upper IGBT is approximately equal to the turn-off switching current of the lower diode (itaþ y idb). Also, for switching frequencies much higher than the fundamental frequency, the current along a switching period is approximately constant, and therefore the turn-on switching current can be approximated to be the same as the turn-off switching current (itaþ y itbþ). Then, the average and RMS values of switching currents through all devices (IGBT and diode) are approximately equal: Iswa;AVG ¼ Ita;AVG ¼ Itb;AVG ¼ Idb;AVG

[9.72]

Iswa;RMS ¼ Ita;RMS ¼ Itb;RMS ¼ Idb;RMS

[9.73]

226

The ratio of RMS phase current to RMS current through the IGBT and diode in a 2L-VSC for three different modulation methods: SPWM, SVPWM and SFTM

Table 9.5 SPWM

SVPWM

SFTM

It;RMS 2 Ia Id;RMS 2 Ia It;RMS 2 Ia

Id;RMS 2 Ia It;RMS 2 Ia

3p 8$Ms $cosð4Þ 12p 8 pffiffiffi > Ms þ 3p 4$Ms $cosð4Þ2 þ 8 3$Ms $cosð4Þ p > > > for j4j < > > 12p 6 < pffiffiffi p ffiffi ffi 3 > > 3p þ 2Ms $ 2 þ sinj24j cosð4Þ2 2 sinj4j þ 2 3 cosð4Þ > > p p 2 > > < j4j < for : 6 2 12p 2 It;RMS 1 2 Ia 8 pffiffiffi pffiffiffi pffiffiffi 2 > > > ð6 8Ms Þ 3 cosð4Þ þ 3Ms ð4 þ 8 cosð4ÞÞ 8Ms sinj4j þ ð4Ms 3Þsinj24j 3 3 þ 2p þ 6j4j < 12p > > ð4M 3Þsinj24j þ 3p 3j4j s > : 6p It;RMS 2 1 2 Ia

for j4j < for

p 3

p p < j4j < 3 2

Offshore Wind Farms

Id;RMS 2 Ia

3p þ 8$Ms $cosð4Þ 12p


227

The average and RMS values of the switching current can be calculated as follows: Iswa;AVG ¼

Iswa;RMS

1 2p

Z 4

pþ4

ðaswa ðqÞ$ia ðqÞÞ$dq

[9.74]

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Z pþ4 1 ¼ aswa ðqÞ$ðia ðqÞÞ2 $dq 2p 4

[9.75]

where aswa is the switching function of the converter, which is a function of aa and given by: ( 1 0 < aa < 1 aswa ¼ [9.76] 0 aa ¼ 1 or aa ¼ 0 From Eq. [9.76], it can be noted that the switching function is constant and equal to 1 for the SPWM and SVPWM methods, but the SFTM has different values. Table 9.6 presents closed form expressions for the average and RMS values of the switching current evaluating Eqs [9.74] and [9.75] for the modulation methods considered.

9.5.3

Main guidelines of design for a 2L-VSC

In order to evaluate h and r of the converter, first the selection of the components has to be done. This selection of components is subject to the requirements of the

The average and RMS values of the switching current in a 2L-VSC for three different modulation methods: SPWM, SVPWM and SFTM

Table 9.6

Iswa;AVG Ia

SPWM and SVPWM pffiffiffi 2 p

Iswa;RMS Ia

1 pffiffiffi 2

SFTM 8 pffiffiffi > > > 2$ð2 cos 4Þ j4j < p > > > 2p 3 > > > pffiffiffi < 6 sinj4j p 2p < j4j < > 2p 3 3 > > > pffiffiffi > > > 2 $ð2 þ cos 4Þ 2p > > < j4j < p : 2p 3 s ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffi p 3 $cos 2 4 3 4 p

228

Offshore Wind Farms

application, like limitation of the current and voltage ripples or peak values of the current and voltage. In the case of 2L-VSC, three main parts are identified: the DC link, the AC filter and the switch valves. The main guidelines of design for these parts are presented below.

9.5.3.1

DC-link design and semiconductor voltage ratings requirements

The design of the DC-link in a 2L-VSC is composed by the selection of a nominal DC voltage (VDC,N) and the specification of the capacitance value (CDC) in Fig. 9.20. VDC,N is related to the nominal AC voltage by (from Eq. [9.58]): VLL;N VDC;N ¼ pffiffiffi 3$Kmod $MS;nom

[9.77]

where VLL,N is the nominal line-to-line RMS voltage and MS,nom is MS at nominal operation. Assuming that VLL,N is given by the application, then MS,nom should be determined. VDC,N itself is designed taking a safety margin on operation of the VSC to guarantee the controllability under abnormal conditions, and since the boundary of overmodulation is when MS is equal to 1, then MS,nom should be lower but close to 1. For example, a value of 0.98 for MS,nom will leave a safety margin of approximately 2% on operation, which means that the VSC can manage an increase on voltage of 2% without entering the overmodulation range. Low values of MS,nom will require higher values of VDC,N, which are mainly limited by the PSD technology if the series connection of devices is to be avoided. In 2L-VSC configuration, each valve will support the total VDC,N, and when series connection of semiconductors is not considered, the minimum Vbk (Vblock,min) required by the PSD can be calculated by (from Eq. [9.21]): 8 > > > > > >
> VDC;N $kovf $ 1 þ > > 2 > > : kvp

for

dVdc 2$

for

dVdc > 2$

kvp 1 kvdc kvp 1 kvdc

[9.78]

where dVdc is the ratio of peak-to-peak voltage ripple to VDC,N, and kovf is the overvoltage factor. For typical industrial networks, kovf ¼ 1.1 for low-voltage systems (VDC,N < 1 kV) and kovf ¼ 1.15 for medium-voltage systems (Backlund et al., 2009). The voltage device rating is selected as the next standard device voltage rating higher than Vblock,min. Standard voltage ratings for commercial IGBT power modules (IGBT with antiparallel diode) are 1.2, 1.7, 2.5, 3.3, 4.5 and 6.5 kV.


229

Fig. 9.23 shows the selection of IGBT power modules as a function of MS,nom and VLL,N for 2L-VSC, when dVdc is lower than 20% and the safety factors are kvp ¼ 0.8 and kvdc ¼ 0.65. Fig. 9.23(a) presents the selection map for SPWM and Fig. 9.23(b) is the selection map when SVPWM or SFTM is adopted. For

(a)

IGBT blocking voltage selection map for SPWM modulation 1

1.7(kV) blocking 0.99 voltage modules

Nominal modulation index (MS,nom)

0.98

3.3(kV) blocking voltage modules

0.97

0.96


0.95

0.94

0.93

Required blocking voltage higher than available module technology

0.92

0.91

0.9 400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

2600

Nominal line-to-line voltage (VLL,N) (V)

(b)

IGBT blocking voltage selection map for SVPWM and SFTM modulations 1

0.99


Nominal modulation index (MS,nom)

0.98


0.97

0.96


0.95

0.94

0.93

Required blocking voltage higher than available module technology

0.92

0.91

0.9 400

600

800

1000

1200

1400

1600

1800

2000

2200

2400

2600

Nominal line-to-line voltage (VLL,N) (V)

Figure 9.23 Selection of IGBT modules depending on required blocking voltage in a 2L-VSC. For illustration simplicity, only three standard voltage ratings are taken into account: (a) IGBT selection map for SPWM modulation; (b) IGBT selection map for SVPWM and SFTM modulations.

230

Offshore Wind Farms

illustration simplicity, only three standard voltage ratings are taken into account in Fig. 9.23 (1.7, 3.3 and 6.5 kV). From Fig. 9.23, it can be noted that for a given nominal voltage, the applicability of a given power module is limited as MS,nom is decreased in 2L-VSC. Also, under these conditions, the use of 2L-VSC in medium-voltage systems is limited by IGBT module technology, which limits the application nominal voltage to around 2.2 kV with SPWM and to around 2.5 kV with SVPWM or SFTM. On the other hand, the DC-link capacitance (CDC) can be designed according to an expression presented in Lai et al. (2008), which is derived taking into account that in one switching cycle, the input power drops to zero while the inverter keeps the maximum output power, or vice versa: CDC ¼

2 VDC;N $

PN dVdc þ 0:5$d2Vdc $fsw

[9.79]

where PN is the nominal power of the 2L-VSC. Since PN and VLL,N (which defines VDC,N) are given by the application specifications, CDC will be determine by the selection of fsw and dVdc, which are considered as design variables, and normally their values are limited by the design constraints. In the case of fsw, the maximum fsw is determined by the switch device technology, and for IGBT power modules, it is possible to reach switching frequencies of up to 4 kHz (Backlund et al., 2009) (depending of the voltage rating). On the other hand, the maximum dVdc is given by specifications related to stability and controllability requirements of the system, and it can be limited to 6%. Once VDC,N and CDC are defined, the volume, mass and capacitor dielectric losses can be calculated using models presented in Section 9.4. However, the capacitor RMS current should be evaluated in order to calculate the capacitor resistive losses (Eq. [9.54]). From Fig. 9.20, the capacitor current (iC) can be defined by the DC-link input current (iin) and the input current of the VSC (ivsc). Considering that the DC component of ivsc is supplied by the DC-link input current (it means, Iin,AVG ¼ Ivsc,AVG), then iC is defined solely by the AC components of iin and ivsc (Kolar and Round, 2006): iC ¼ ivsc;AC iin;AC

[9.80]

Since the capacitor RMS current (IC,RMS) is quite complex to derive using Eq. [9.80], it is possible to consider that the currents iin and ivsc do not contain common harmonics (which can be the case when the input current comes from a bridge diode rectifier or from a voltage source rectifier (VSR) at different fsw), and then IC,RMS can be approximated by (Kolar and Round, 2006): 2 2 2 IC;RMS ¼ Ivsc;AC;RMS þ Iin;AC;RMS

[9.81]


231

The analytical expression presented in Kolar and Round (2006) is considered in order to evaluate the contribution caused by the VSC to IC,RMS (for any PWM modulation method, and assuming purely sinusoidal phase current): 2 Ivsc;AC;RMS

pffiffiffi pffiffiffi 6$Kmod $MS 3 6$p$Kmod $Ms $ 1þ 4 ¼ $cos2 ð4Þ $Ia2 p 2 [9.82]

The contribution caused by the DC-link input current to IC,RMS is determined by the type of rectifier interfacing the 2L-VSC, and for brevity, it can be considered as a system specification parameter for the design of 2L-VSC. This value can be defined as a relative value to the DC component of the DC-link: Iin;AC;RMS ¼ dIin $Iin;AVG

[9.83]

where dIin is the ratio of RMS ripple component to the DC component of the DC-link input current. Since Iin,AVG ¼ Ivsc,AVG, and noting that the average input current of the VSC is calculated from the average current of the upper devices (TAþ, DAþ, TBþ, DBþ, TCþ and DCþ, in Fig. 9.20), the average DC-link current can be expressed by: Iin;AVG ¼ Ivsc;AVG ¼ 3$ It;AVG Id;AVG ¼ 3$Kmod $MS $cosð4Þ$Ia

[9.84]

Fig. 9.24(a) shows the dependency of DC-link capacitor current RMS value (normalized by the RMS phase current) on the scaled MS (Kmod*MS) for two displacement factor values (cos(4) ¼ 0.6 and cos(4) ¼ 1) and different values of dIin . Additionally, the three-dimensional representation of the dependency of DC-link capacitor current RMS value on MS and 4 (for dIin ¼ 30%) is presented in Fig. 9.24(b).

9.5.3.2

AC filter inductor and current ripple

The input inductance (LF) is sized according to Friedli et al. (2012), where a theoretical derivation is presented on the basis of the peak-to-peak current ripple (DILh) at fsw for a given line-to-line voltage amplitude (VLL) and a DC-link voltage (VDC): LF ¼

VLL V2 1 pffiffiffi LL $ 3 2$VDC DILh $fsw

[9.85]

Since the maximum DILh is normally specified by diL, the required inductance value to limit DILh can be calculated as follows: LF ¼

2 VLL;N $cos 4 3 1 $Kmod $MS;nom $ pffiffiffi 2 2$diL $fsw $PN

[9.86]

232

Offshore Wind Farms

(a)

Dependency of the DC-link capacitor current RMS value on the scaled modulation index

0.8

δ 0.7

0.6

(IC,RMS / Ia)

0.5

Iin

δ

= 40%

Iin

= 0%

δ = 40% Iin

δ = 40% Iin

δ

δ = 30% Iin

cos(φ ) = 1 Iin

= 0%

0.4

δ

Iin

δ 0.3

= 20%

Iin

δ

Iin

= 10% = 0%

cos(φ ) = 0.6

0.2

0.1

0 0

0.05

0.1

0.15

0.2

Kmod * Ms

0.25

0.3

Three-dimensional representation for δ

(b)

Iin

0.35

0.4

0.45

= 30%

0.7 0.6

(IC,RMS / Ia)

0.5 0.4 0.3 1 0.2 0.9 0.1

0.8

0 0.45

0.7 0.4

0.35

0.3

0.25

0.2

0.15

0.1

0.05

0

0.6

cos(φ )

Kmod * Ms

Figure 9.24 Dependency of the DC-link capacitor current RMS value on the scaled modulation index (Kmod*MS) and the displacement factor cos(4) of the fundamental phase voltage and the phase current in a 2L-VSC: (a) dependency on scaled modulation index for two displacement factor values (0.6 and 1) and five values of the ratio dIin (0%, 10%, 20%, 30% and 40%); (b) three-dimensional representation for dIin ¼ 30%.


233

Normally, the inductance value is limited by the maximum ratio of inductor voltage to the line-to-line voltage at nominal power (dVL), and then, using inductor RMS current at f1, the maximum inductance value can be approximated by: LF;max ¼

2 3$dVL $VLL;N $cos 4 rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi 2 6 þ d iL p$f1 $PN $

[9.87]

Additionally, when 2L-VSC is interfacing a machine (motor or generator), the equivalent per phase inductance of the machine (LM) acts as an additional filter, and therefore it is possible to reduce LF, and even for some conditions (when LF < LM) the AC filter inductor is not necessary. Then, when the converter interfaces a machine, LF can be calculated by: LF ¼

2 VLL;N $cos 4 3 1 $Kmod $MS;nom $ pffiffiffi LM 2 2$diL $fsw $PN

[9.88]

Finally, the nominal inductor RMS current (ILF) should be calculated in order to evaluate the inductor volume, mass and power losses using models presented in Section 9.3, and considering the approximation of Fig. 9.14, the inductor RMS current is calculated as follows:

ILF

9.5.3.3

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ! qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u u d2iL PN 2 2 t $ pffiffiffi ¼ I a þ I Lh ¼ 1þ 6 3$VLL;N $cos 4

[9.89]

Switch valve design and selection of PSDs

To design the switch valve, first the required Vbk and the maximum peak current of the valve should be calculated. Then, the selection of the PSD used in the valve is done from a set of available devices, and this selection depends on the type of preferred connection (series, parallel or neither) and the maximum rating of the available devices. Once the device is selected, the semiconductor power losses can be calculated for the worst operating condition and therefore the size of the module cooling system can be estimated using models from Section 9.2. For brevity, only the design of the switch valve for 2L-VSC with parallel connection of IGBT modules is discussed in detail, but it should be noted that the design process can easily be adapted to series connection of devices (considering Section 9.2.3). Even more, to simplify the device selection, only the IGBT modules presented in Table 9.2 are considered (which are also considered in Fig. 9.23). Since no series connection is preferred, the expression derived for Vblock,min required by the PSD can be used (Eq. [9.78] and Fig. 9.23) and VLL,N is limited to around 2.5 kV (with SVPWM) for the devices considered (maximum Vbk ¼ 6.5 kV) as discussed previously. On the other hand, the maximum peak current of the parallel

234

Offshore Wind Farms

connection (valve) is given by the phase current, its current ripple component and the overload factor (kolf) (typically 30%; then kolf ¼ 0.3), and can be calculated by: diL rffiffiffi pffiffiffi diL 2 PN $ 1 þ 2 $ð1 þ kolf Þ [9.90] Ipsw ¼ 2$Ia $ 1 þ $ $ð1 þ kolf Þ ¼ VLL;N $cos 4 3 2 Then, from Eq. [9.18], the minimum np needed to not exceed the current rating of the devices can be calculated by Eq. [9.91] (where the function x means the smallest integer not less than x). Since the switch is implemented by IGBT with antiparallel diode modules, and the IGBT and diode have different dCI (for example modules in Table 9.2), then np,min should be chosen as the maximum value from evaluation of Eq. [9.91] for IGBT and diode parameters. np;min ¼

Ipsw 1 þ dCI 1 $ þ1 1:6$In 1 dCI

[9.91]

Fig. 9.25 shows the minimum np as a function of VLL,N and PN of 2L-VSC with SVPWM modulation for the three power modules in Table 9.2. Values of cos(4) ¼ 0.85 and MS,nom ¼ 0.98 are considered in Fig. 9.25. Although Eq. [9.91] gives a minimum value of np to guarantee not to exceed the current rating of the

10

np = 7 9

Nominal power (PN) (MW)

8

np = 9 np = 5

7

np = 5

np = 7

6 5

np = 3

np = 5 np = 3

4

np = 3 3 2

np = 1

np = 1

Power module Power module Power module 1 3.3kV×1500A 1.7kV×3600A 6.5kV×750A Ref. FZ3600R17KE3 Ref. FZ1500R33HE3 Ref. FZ750R65KE3 400 600 800 1000 1200 1400 1600 1800 2000 Nominal line-to-line voltage (VLL,N) (V)

np = 1

2200

2400

Figure 9.25 Minimum number of parallel connected devices as a function of the nominal power (PN) and the nominal line-to-line voltage (VLL,N) of the 2L-VSC with SVPWM modulation for three semiconductor power modules. A displacement factor cos(4) of 0.85 and a nominal modulation index of 0.98 are considered.


235

modules parallel connected with maximum dCI, when the converter is designed to operate with a relatively high fsw, it may be necessary to increase np in order to not exceed the maximum power that the module can dissipate mounting in the HS without overheat itself (restriction from Eq. [9.28]). Fig. 9.26 shows the maximum allowable HS temperature rises (calculated from Eq. [9.25]) as a function of fsw for a 2-MW 2L-VSC with SVPWM modulation and three designs with different VLL,N for each semiconductor power module from Table 9.2 (considering one module per switch from Fig. 9.25). A converter design to be a voltage source inverter (VSI) is shown in Fig. 9.26(a), and the VSR is presented in Fig. 9.26(b). From Fig. 9.26, it can be noted that when the minimum np from Eq. [9.91] is considered, there is a limit switching frequency ( fsw,limit), which the converter can operate without exceeding the maximum power that the module can dissipate mounting in an HS with RthHS,min (RthHS,min ¼ 10 K/kW in Fig. 9.26). Beyond fsw,limit, RthHS does not exceed the maximum Tj at worst conditions and is lower than RthHS,min given by the HS technology (restriction from Eq. [9.29]). By comparing Fig. 9.26(a) with Fig. 9.26(b), it is possible to appreciate how the converter operation mode affects the thermal design of the converter, which is influenced mainly by the semiconductor parameters (IGBT and diode). Also, since power modules with higher Vbk have stronger frequency dependency (they require thicker silicon, which gives higher switching losses (Backlund et al., 2009)), fsw,limit is lower for high Vbk modules, as can be noted from Fig. 9.26. If the converter is designed to operate with an fsw higher than fsw,limit, then np should be higher than the minimum value from Eq. [9.91]. Fig. 9.27 shows fsw,limit as a function of np for a 2-MW 2L-VSC with the same design conditions as in Fig. 9.26. From Fig. 9.27, it can be noted that the fsw,limit versus np curve has an asymptotic behaviour, and fsw asymptote decreases as the required Vbk is increased. Hence, this characteristic will reduce the possible use of high Vbk devices for designs with high fsw. From Fig. 9.27(a), it can be observed how the converter operation mode influences fsw,limit. For 1.7-kV modules, VSI allows higher fsw,limit than VSR, but this difference decreases as np increases. However, for 3.3-kV and 6.5-kV modules, the influence of the operation mode is less evident, and VSR allows higher fsw,limit than VSI. Additionally, the influence of the modulation technique on fsw,limit can be observed from Fig. 9.27(b). In all cases, SFTM allows higher fsw,limit for the design parameters and constraints considered. Also, from Fig. 9.27(b), it can be noted that the influence of the modulation technique is less significant for devices with higher Vbk. Finally, a constraint related to the maximum fsw should be taken into account in order to limit the use of power modules to the range of switching frequencies that the module can manage without destruction or that the relative time that the module takes to change state (turn-on and turn-off times) does not exceed a practical limit (for example 2%). In Table 9.2, a maximum fsw,limit for each module is included, which is calculated to guarantee that the conduction time per switching period is higher than 98% of the switching period. Fig. 9.28 shows an example of power loss evaluation per module as a function of fsw for a 2-MW 2L-VSI and with SVPWM and the same design conditions as in Fig. 9.26.

236

(a)

Offshore Wind Farms Inverter mode: 2L-VSC with SVPWM modulation (PN = 2(MW); cos(φ) = 0.85; Ms = 0.99; Tamb = 40(ºC); RthHS,min = 10(K/kW), KSFT = 0.85)

60

Maximum heat sink temperature rise ( ΔTHS,max ) (ºC)

55 50 45 40

fsw,limit

35 30 25

RthHS,min × (Pigbt+Pdiode)

20

VLL = 690(V); 1.7 kV module VLL = 1250(V); 3.3 kV module

15 2 10

(b) 60

VLL = 2500(V); 6.5 kV module

Switching frequency (fsw) (Hz)

10

3

Rectifier mode: 2L-VSC with SVPWM modulation (PN = 2(MW); cos(φ) = 0.85; Ms = 0.99; Tamb = 40(ºC); RthHS,min = 10(K/kW), KSFT = 0.85) VLL = 690(V); 1.7 kV module

Maximum heat sink temperature rise ( ΔTHS,max ) (ºC)

VLL = 1250(V); 3.3 kV module VLL = 2500(V); 6.5 kV module

55 50 45

fsw,limit

40 35 30 25 20 15 2 10

RthHS,min × (Pigbt+Pdiode)


10

3

Figure 9.26 Maximum allowable heat sink temperature rises as function of the switching frequency for a 2-MW 2L-VSC with SVPWM modulation and three designs with different nominal line-to-line voltage for each semiconductor power module from Table 9.2 (one module per switch is considered): (a) inverter mode of operation; (b) rectifier mode of operation.


237

(a)

2L-VSC with SVPWM modulation (PN = 2(MW); cos(φ ) = 0.85; Ms = 0.99; Tamb = 40(ºC); RthHS,min = 10(K/kW), KSFT = 0.85) 7000 Rectifier - VLL = 690(V)

Rectifier - VLL = 1250(V)

6000

Rectifier - VLL = 2500(V) Inverter - VLL = 690(V) Inverter - VLL = 1250(V)

Limit switching frequency (fsw,limit) (Hz)

Inverter - VLL = 2500(V)

5000

4000

3000

2000

1000

0 1

2

3

4

5

6

7

8

9

10

Number of parallel connected modules (np)

(b)

Inverter mode (PN = 2(MW); cos(φ ) = 0.85; Ms= 0.99; Tamb = 40(ºC); RthHS,min = 10(K/kW), KSFT = 0.85)

7000 VLL = 690(V) - SPWM VLL = 1250(V) - SPWM

6000

VLL = 2500(V) - SPWM VLL = 690(V) - SVPWM

Limit switching frequency (fsw,limit) (Hz)

VLL = 1250(V) - SVPWM VLL = 2500(V) - SVPWM

5000

VLL = 690(V) - SFTM VLL = 1250(V) - SFTM VLL = 2500(V) - SFTM

4000

3000

2000

1000

0 1

2

3

4

5

6

7

8

9

Number of parallel connected modules (np)

Figure 9.27 Limit-switching frequency as a function of the number of parallel connected semiconductor modules for a 2-MW 2L-VSC with different nominal line-to-line voltage: (a) comparison of converter operation mode (Rectifier and inverter) with SVPWM; (b) comparison of modulation technique for inverter mode.

10

238

Offshore Wind Farms

Inverter mode: SVPWM modulation (PN = 2(MW); cos(φ ) = 0.85; Ms = 0.99; Tamb = 40(°C); RthHS,min = 10(K/kW), KSFT = 0.85) 5 VLL = 690(V) VLL = 1250(V)

Power losses per module (Pigbt + Pdiode) (kW)

4.5

VLL = 2500(V)

4

3.5

3

2.5

2

1.5 2 10

103 Switching frequency (fsw) (Hz)

Figure 9.28 Example of power losses per module as function of the switching frequency for a 2-MW 2L-VSC operated as inverter and with SVPWM. Designs with three different nominal line-to-line voltages are considered, each with different power modules and variation of the number of modules parallel connected to fulfil heat sink thermal constraints.

VLL = 1250(V) VLL = 2500(V) np for 1.7(kV) module

15

np for 3.3(kV) module

3

np for 6.5(kV) module

10

2

5

1

0 102

103

Number of module parallel connected per valve (np)

Power losses per valve (np*(Pigbt + Pdiode)) (kW)

Inverter mode: SVPWM modulation (PN = 2[MW]; cos(φ ) = 0.85; Ms = 0.99; Tamb = 40(°C); RthHS,min = 10(K/kW), KSET = 0.85) 4 20 VLL = 690(V)

0


Figure 9.29 Example of power losses per valve and number of modules parallel connected as function of the switching frequency for a 2-MW 2L-VSC operated as an inverter and with SVPWM.


239

Also, Fig. 9.29 presents an example of power loss evaluation per valve and np as a function of fsw for the same converter design of Fig. 9.28. The maximum switching frequencies considered for the 1.7-kV, 3.3-kV and 6.5-kV modules are 4, 2 and 1.5 kHz, respectively.

9.5.4

Evaluation of the power losses, volume and mass of the 2L-VSC

For a given set of design parameters, constraints and variables, the total losses of 2L-VSC are calculated by Eq. [9.92], via the summation of switch valves losses (Pvalve), AC filter inductor losses (PLF) and DC-link capacitor losses (PCDC): Plosses;VSC ¼ 6$Pvalve þ PLF þ PCDC

[9.92]

Pvalve ¼ np $ðPigbt þ Pdiode Þ

[9.93]

PLF ¼ PwLF þ PcLF

[9.94]

PCDC ¼ PεC þ PUC

[9.95]

In this case, the converter has six switch valves, each implemented by np modules (IGBT with antiparallel diode) parallel connected with IGBT and diode power losses defined in Section 9.2 by Eqs [9.26] and [9.27] at nominal current. The total VSC volume (VolVSC) is calculated by Eq. [9.96], via summation of the individual volumes of switch valves (Volvalve from Eq. [9.22]), the three-phase inductor (VolLF from Eq. [9.35]) and the DC-link capacitor (VolCDC from Eq. [9.49]). A volume utilization factor CPV of 0.6 is considered for 2L-VSC. In a similar way, the total VSC active mass (MassVSC) is obtained by Eq. [9.97]. VolVSC ¼

1 $ð6$Volvalve þ VolLF þ VolCDC Þ CPV

MassVSC ¼ 6$Massvalve þ MassLF þ MassCDC

[9.96] [9.97]

Once the total power losses, the total volume and total active mass are evaluated, then h, r and g can be calculated by evaluation of Eqs [9.1], [9.2] and [9.3], respectively.

9.6

Evaluation example of a 1-MW 2L-VSC

The design of a 1-MW 2L-VSC is considered as a design example in order to illustrate how to apply the proposed methodology. The system parameters, design constraints and reference models used in the example are indicated in Table 9.7. The three modulation strategies presented in Section 9.5.1 are compared for each operation mode of

240

Offshore Wind Farms

System parameters and design constraints for the design example 1-MW 2L-VSC

Table 9.7

Parameter

Symbol

Value

Constraint

Symbol

Value

Nominal power

PN

1 MW

Safety factor for DC voltage

kvdc

0.65

Nominal line-to-line RMS voltage

VLL,N

690 V

Safety factor for peak voltage

kvp

0.8

Power factor

cos 4

0.85

Safety factor of thermal design

KSFT

0.85

Fundamental frequency

f1

50 Hz

Max. relative inductor voltage

dVL,max

0.3

Equivalent machine inductance per phase

LM

50 mH

Max. relative heat sink structure volume

dHS,max

6

Overload factor

kolf

0.3

Max. relative AC-current ripple

dIL;max

0.2

Volume utilization factor

CPV

0.6

Max. relative DC-link ripple

dVdc,max

0.02

Relative input DC-link current ripple

dlin

0.3

Heat sink model: DAU series BF-XX with axial fan SEMIKRON series SKF-3XX at 10 m/s

Nominal modulation index

Ms,nom

0.99

Inductor model: Siemens series 4EUXX e Cu

Ambient temperature

Tamb

40 C

Capacitor model: TDK series MKP-B256xx

the 2L-VSC (rectifier or inverter). Even more, to simplify the selection of the devices, only the IGBT modules presented in Table 9.2 are considered. Despite that it has been shown in previous studies (Kolar et al., 1990; Helle, 2007; Wen et al., 2011; Preindl and Bolognani, 2011; Friedli et al., 2012; Lee et al., 2014) that the SPWM is less efficient than the SVPWM and SFTM and it is considered as a base case in order to show how the choice of the modulation strategy influences the performance indices considered in this chapter.

9.6.1

Pareto-front of the 1-MW 2L-VSC with SPWM

First, the impact of fsw and diL on the total volume, the nominal power losses and total active mass are analysed for 2L-VSR (interfacing a generator). A variation of fsw from 500 Hz to 4 kHz (or the maximum fsw that the IGBT module can be switched to) is considered, but configurations which do not correspond to the thermal and frequency requirements of the IGBT modules are not shown. Fig. 9.30 shows the nominal power losses, the total volume and the total active mass as functions of fsw for 2L-VSR with SPWM. Since SPWM with a VLL,N of 690 V is considered, then the 3.3 kV/1500 A IGBT module should be used (as it


241

(a) 7

Total Switch valves DC-link capacitor AC inductor filter

Nominal power losses (%)

6

5

Limited by LF,max

4

3

2

1

np = 1

0 500

600

700

800

900

np = 2

1000

2000

3000

Switching frequency (Hz)

(b) 0.8


0.7

Volume (m3)

0.6

0.5

0.4

0.3

Limited by IGBT module

0.2

0.1

0 500

600

700

800

900

1000

2000

3000


Figure 9.30 Design example 1-MW 690-V 2L-VSC, evaluation of power losses, volume and mass as function of the switching frequency for SPWM modulation and rectifier operation mode; (a) power losses; (b) volume; (c) mass.

242

Offshore Wind Farms

(c) 1.5

Total Switch valves DC-link capacitor AC-inductor filter

Mass (t)

1

0.5


0 500

600

700

800

900

1000


2000

3000

Figure 9.30 Continued.

can be observed from Fig. 9.23(a)) and fsw is limited to 2 kHz. In addition, the minimum fsw is limited to around 700 Hz because of the maximum relative inductor voltage constraint (which requires the maximum inductance value (LF,max) for the given diL of 20%). Fig. 9.30(a) shows how the different components contribute to the total nominal losses. It is observed that the capacitor loss contribution is very low compared with the inductor and switch valves’ losses. Also, it can be noted that increasing fsw beyond around 1.36 kHz will require connecting two modules in parallel per valve and therefore it increases in the valve losses abruptly. Fig. 9.30(b,c) shows the total volume and active mass, respectively, and the contribution of each component. It can be observed that the inductor represents the main contribution to the total mass and volume in the 2L-VSC with the characteristics indicated in Table 9.7. Also, it is noted that as fsw increases the required inductance and capacitance value decrease and therefore the inductor/capacitor volume and mass decrease. The evaluations of h, r and g as functions of fsw are presented in Fig. 9.31. It can be noted that aiming to maximize h and r (or g) are two conflicting objectives. The maximum h of 97.5% is obtained for converter design with fsw of 700 Hz, but at the same fsw the minimum r and g are obtained. On the other hand, as can be observed from Fig. 9.31, the maximum values of r and g (1.93 MW/m3 and 1.28 MW/t, respectively) are obtained when the maximum fsw (2 kHz) is considered; however a considerable reduction in nominal h (from 97.5% to 93.7%) is necessary if the converter is switched at 2 kHz.

Modelling of power electronic components 2 1.9 1.8

243 99

ρ γ η

98

1.7

3 ρ (MW/m ); γ (MW/t)

1.5 np= 1

1.4 1.3 1.2


n p= 2

96

Limited by LF,max

95

Nominal η (%)

97

1.6

1.1 1

94

0.9 93

0.8 0.7 0.6 500

600

700

800

900

1000

2000

92 3000


Figure 9.31 Design example 1-MW 690-V 2L-VSC; evaluation of the nominal efficiency h (left axis), the power density r and the power to mass ratio g (right axis) as function of the switching frequency for SPWM modulation and rectifier operation mode.

The influence of the maximum diL (diL;max ) on the nominal power losses, the total volume and total active mass are illustrated in Fig. 9.32, when an fsw of 1.25 kHz is kept constant in the converter design. It can be noted from Fig. 9.32 that the variation of diL;max only has a relevant influence on the AC inductor filter design. The inductor mass and volume decrease as the converter design allows a higher diL , but the inductor losses increases. Fig. 9.33 shows the evaluation of the nominal h, r and g as function of diL;max for fsw of 1.25 kHz. Similarly to the case of fsw variation (Fig. 9.31), it can be noted that targeting to maximize h and r (or g) are two conflicting objectives from the point of view of diL;max restriction. A reduction of 10% on diL;max restriction (from 20% to 30%) will cause a reduction of 0.75% to h and an increase of 23% and 33% to r and g of the converter, respectively. Finally, the optimized design of the 1-MW 690-V 2L-VSC with SPWM and operated as rectifier, taking into account different fsw and diL , is plotted in Fig. 9.34. The design cases from Fig. 9.31 (diL ¼ 0:2) and Fig. 9.33 ( fsw ¼ 1.25 kHz) are also included in Fig. 9.34. The relationship between h and r for the space of solution is presented in Fig. 9.34(a), which also shows the her Pareto-front for the rectifier mode 2L-VSC (black curve). The reg Pareto-front and reg relationship for the space of solution are presented in Fig. 9.34(b). It is possible to observe from Fig. 9.34(a), how variations on diL have a notorious impact on r, but little influence on h, when fsw is kept constant. Also, it can be noted from Fig. 9.34(a), how np is needed to fulfil thermal constraints and has a significant impact on the her space of solutions, which is clearly divided according to np with the

244

Offshore Wind Farms

(a) 4.5

4

Nominal power losses (%)

3.5

3

2.5

2

1.5 Total Switch valves DC-link capacitor AC inductor filter

1

0.5

0 10

12

14

16

18

20

22

24

26

28

30

Relative AC current ripple δ iL (%)

(b)0.8


0.7

0.6

3

Volume (m )

0.5

0.4

0.3

0.2

0.1

0 10

12

14

16

22 20 18 Relative AC current ripple δ iL (%)

24

26

28

30

Figure 9.32 Design example 1-MW 690-V 2L-VSC with a switching frequency of 1.25 kHz, evaluation of power losses, volume and mass as function of the relative AC current ripple for SPWM modulation and rectifier operation mode: (a) power losses; (b) volume; (c) mass.


245

(c)

1.6


1.4

1.2

Mass (t)

1

0.8

0.6

0.4

0.2

0 10

12

16

14

18 20 22 Relative AC current ripple δ (%)

24

30

28

26

Figure 9.32 Continued.

2.2

ρ

Limited by LF,max

γ

97.5

η

ρ (MW/m3); γ (MW/t)

1.8

97

1.6

96.75

1.4

96.5

1.2

96.25

1

96

0.8

0.6 10

Nominal η (%)

97.25

2

95.75

12

14

16

18

20

22

24

26

28

30

Relative AC current ripple δ (%)

Figure 9.33 Design example 1-MW 690-V 2L-VSC with a switching frequency of 1.25 kHz; evaluation of the nominal efficiency h (left axis), the power density r and the power to mass ratio g (right axis) as function of the relative AC current ripple for SPWM modulation and rectifier operation mode.

246

Offshore Wind Farms

(a)

η-ρ Pareto-front

2.5 np = 1

np = 2

fsw decreases

ρ (MW/m3)

2

1.5

Space of solutions η-ρ Pareto-front fsw= 1.25(kHz) 1 92

δ iL

δ iL= 0.2

increases 94

93

95 Nominal η (%)

1.5

Space of solutions ρ -γ Pareto-front fsw= 1.25(kHz)

δ iL= 0.2 fsw decreases

1.4

δiL decreases

1.3

γ (MW/t)

98

ρ -γ Pareto-front

(b) 1.6

97

96

1.2 1.1 1 0.9 0.8 0.7 0.6 1

1.2

1.4

1.6

1.8

ρ (MW/m3)

2

2.2

2.4

Figure 9.34 Pareto-front for design example 1-MW 690-V 2L-VSC rectifier operation mode with SPWM modulation: (a) efficiency (h) versus power density (r); (b) power density (r) versus power to mass ratio (g).


247

highest efficiencies for the solutions with one module. Additionally, when two IGBT modules are parallel connected, an increase in fsw has a small impact on r but a large influence on h. On the other hand, an increase in fsw or diL;max will improve r and g, as can be observed from Fig. 9.34(b), but it decreases the nominal h of the solution, as it is shown in Fig. 9.34(a). Also, a variation in fsw has more impact on g than r, as can be noted from Fig. 9.34(b). The reg Pareto-front (in Fig. 9.34(b)) of the solutions is short compared with the her Pareto-front (in Fig. 9.34(a)), which shows that r and g are highly correlated.

9.6.2

Modulation techniques comparison

A comparison of the three modulation strategies considered in Section 9.5.1 is performed based on the three performance indices (h, r, g) considered in this chapter. The system parameters, design constraints and reference models used in the comparison are indicated in Table 9.7. Figs 9.35 and 9.36 show a comparison of the three modulation strategies (SPWM, SVPWM and SFTM) in terms of h (top), r (middle) and g (bottom), respectively, as functions of fsw for VSR and VSI. It can be noted that solutions based on SVPWM and SFTM will leave better solutions than those based on SPWM for the range of fsw considered. Also, SVPWM and SFTM modulations allow 1.7 kV/3600 A IGBT modules to used for a VLL,N of 690 V, as was analysed from Fig. 9.23(b), and therefore frequencies higher than 2 kHz can be considered. The SFTM looks to be the best modulation in terms of the three performance indices, as can be noted from Figs 9.35 and 9.36, for VSR and VSI, respectively. Comparing Figs 9.35 and 9.36, it can be noted that VSI can achieve higher values for each performance indice than VSR. However, it should be noted that this conclusion is limited to the system parameters and design constraints considered (Table 9.7). As noted previously for the SPWM, an increase in fsw will improve r and g of the converter with a reduction of the nominal h. However it can be noted, from Figs 9.35 and 9.36, that there is an optimal fsw beyond which r or g will decrease. In the case of VSR with SFTM, the maximum r (3.55 MW/m3) and g (2.62 MW/t) are obtained for the same fsw of 3.79 kHz. In the case of VSI with SFTM, the maximum r (3.76 MW/m3) is obtained for fsw ¼ 3.79 kHz, but the maximum g (2.77 MW/t) is obtained for the maximum possible fsw (4 kHz), which indicates that the IGBT technology limits the maximum g. Fig. 9.37 shows the influence of the modulation method on the her Pareto-front for VSR (Fig. 9.37(a)) and VSI (Fig. 9.37(b)). It is clear that SFTM will give the best trade-off between h and r, and only comparable solutions are obtained with SVPWM for small values of fsw, which imply low r. The reg Pareto-front of 2L-VSC with the three modulation strategies is shown in Fig 9.38(a,b) for VSR and VSI, respectively. Again, SFTM presents the best trade-off between r and g, where the Pareto-front is obtained for the solution with the highest switching frequencies.

248 Offshore Wind Farms

Figure 9.35 Rectifier operation mode: Comparison of modulation methods in the design example 1-MW 690-V 2L-VSC. Evaluation of the nominal efficiency h (top), the power density r (middle) and the power to mass ratio g (bottom) as function of the switching frequency.

η (%)

98 96 94 92 90 500

SPWM SVPWM SFTM 600

700

800

900 1000

2000 Switching frequency (Hz)

3000

4000

5000

700

800

900 1000


3000

4000

5000

700

800

900 1000

2000

3000

4000

5000

ρ (MW/m3)

4 3

SPWM SVPWM SFTM


Inverter mode 100

2 1 500

600

3

γ (MW/t)

2.5 2

SPWM SVPWM SFTM

1.5 1 0.5 500

600

Switching frequency (Hz) 249

Figure 9.36 Inverter operation mode: Comparison of modulation methods in the design example 1-MW 690-V 2L-VSC. Evaluation of the nominal efficiency h (top), the power density r (middle) and the power to mass ratio g (bottom) as function of the switching frequency.

250

Offshore Wind Farms η - ρ Pareto-front rectifier mode

(a) 4

SPWM SVPWM SFTM

3.5

3

ρ (MW/m3)

fsw increases

2.5

2

1.5

1 91

92

93

94

95

93

97

98

99

Nominal η (%)

η - ρ Pareto-front inverter mode

(b) 4

SPWM SVPWM SFTM

3.5

ρ (MW/m3)

3

fsw

2.5

increases

2

1.5

1 91

92

93

94

95

96

97

98

99

Nominal η (%)

Figure 9.37 Influence of modulation method on her Pareto-front for design example 1-MW 690-V 2L-VSC: (a) rectifier operation mode; (b) inverter operation mode.


(a) 3

251 ρ -γ Pareto-front rectifeir mode

SPWM SVPWM SFTM

2.5

γ (MW/t)

2

1.5

fsw increases

1

0.5 1

(b) 3

1.5

2

2.5 3 ρ (MW/m )

3

3.5

4

3

3.5

4

ρ -γ Pareto-front inverter mode SPWM SVPWM SFTM

2.5

γ (MW/t)

2

1.5

fsw increases

1

0.5 1

1.5

2

2.5 3 ρ (MW/m )

Figure 9.38 Influence of modulation method on reg Pareto-front for design example 1-MW 690-V 2L-VSC: (a) rectifier operation mode; (b) inverter operation mode.

252

Offshore Wind Farms

9.6.3

Optimal selection of the switching frequency in a 2L-VSC

From Figs 9.35 and 9.36, it can be noted that h decreases with fsw, but there is one fsw that maximizes r (or also g). Then an optimal fsw can be obtained for a given set of system parameters and design constraints. The criterion to select fsw can be to maximize the following objective function: L¼

h r g þ þ hmax rmax gmax

[9.98]

where hmax, rmax and gmax are the maximum values of nominal h, r and g, respectively, for a set of design parameters and constraints. Fig. 9.39 shows the objective function as a function of fsw for the design example 1-MW 2L-VSC with SVPWM (on top) and SFTM (on the bottom) for each operative mode (VSR or VSI). Table 9.8 presents the results of the performance indices when fsw is selected to maximize Eq. [9.98]. Finally, an optimized design of a 2L-VSR is done for different converter nominal power and the parameters and constraint presented in Table 9.7. It is assumed that 2L-VSR is interfacing a wind power generator with an equivalent generator inductance per phase of 1e-4 ¼ 104 ¼ 0.0001 in per unit, which is assumed to be the same value for all the nominal power. This low inductance value has been chosen in order to acquire higher inductance values in the input filter and therefore see how the performance indices are affected when the switching frequency is changed.

(a) 3 2.8

Λ for SVPWM VSR VSI

2.6 Λ

2.4 2.2 2 1.8 1.6

1000

2000

3000

4000

5000

3000

4000

5000


(b) 3

Λ for SFTM VSR VSI

Λ

2.5

2

1.5

1000


Figure 9.39 Objective function as function of switching frequency for the design example 1-MW 690-V 2L-VSC: (a) SVPWM; (b) SFTM.


253

Results of the optimal selection of switching frequency follows the objective function (Eq. [9.98]) for the 1-MW 2L-VSC

Table 9.8

Modulation

Operation mode

Switching freq. [kHz]

Nominal efficiency [%]

Power density [MW/m3]

Power to mass ratio [MW/t]

Objective function value

SVPWM

VSR

3.107

93.68

2.938

2.141

2.927

VSI

3.437

94.22

3.36

2.409

2.956

VSR

3.807

94.01

3.547

2.615

2.950

VSI

4.000

93.64

3.754

2.773

2.950

SFTM

Fig. 9.40 shows the results of the optimized selection of fsw as a function of PN when SVPWM is selected for 2L-VSR. Fig. 9.40(a) shows the selecting of fsw based on four criteria to maximize h, r, g or L. Fig. 9.40(b) presents the number of IGBT modules in parallel connection needed when fsw is selected to maximize L. The nominal h, r and g, as functions of the nominal power, are shown in Fig 9.40(c,d,e), respectively. Each of these figures includes the maximum possible value (by selection of the corresponding fsw in Fig. 9.40(a)) and the value obtained when fsw is selected to maximize L. The results of the optimized selection of fsw as function of PN when SFTM is selected for 2L-VSR are shown in Fig. 9.41. Since SFTM is more efficient than SVPWM for the given power factor (0.85), then higher fsw can be used for SFTM and therefore higher r and g are obtained with the same nominal h.

Nomenclature 2L-VSC

Two-level voltage source converter

3L-NPC

Three-level neutral point clamped converter

IEGT

Injection enhanced gate transistor

IGBT

Insulate gate bipolar transistor

IGCT

Integrated gate commutated thyristor

PSD

Power semiconductor device

PWM

Pulse width modulation

SFTM

Symmetrical flat-top modulation

SPWM

Sinusoidal PWM

SVPWM

Space-vector PWM

WECS

Wind energy conversion system

WT

Wind turbine

fsw (kHz) np

(b)

η (%)

(c)

4 3 2 1 0

ρ (MW/m3)

Max. Λ 2

3

4

5

6

7

8

9

10

12 10 8 6 4 2 0 0

1

2

3

4

5

6

7

8

9

10

100 95 Obtained to max.Λ Maximum 0

1

2

3

4

5

6

7

8

9

10

6 4

Obtained to max. Λ Maximum 0

1

2

3

4

5

6

7

8

9

10

4 3 2 0

Obtained to max. Λ Maximum 1

2

3

4

5

6

7

8

9

10

Nominal power PN (MW)

Figure 9.40 Results for design of a 2L-VSR as function of the nominal power when SVPWM is selected: (a) switching frequency; (b) number of parallel connected IGBT modules; (c) nominal efficiency; (d) power density; (e) power-to-mass ratio.

Offshore Wind Farms

γ (MW/t)

Max. γ

Max. ρ

1

2

(e)

Max. η

0

90

(d)

254

(a)

fsw (kHz) np

(b)

Max. η

Max. ρ

Max. γ

1

2

3

4

5

6

7

8

9

10

0

1

2

3

4

5

6

7

8

9

10

10 8 6 4 2 0

100

η (%)

(c)

Max. Λ 0

95

Obtained to max. Λ Maximum

90

ρ (MW/m3)

0

(d)

1

2

3

4

5

6

7

8

9

10

6 4


2 0

(e)


4 3 2 1 0

(a)

1

2

3

4

5

6

7

8

9

10

γ (MW/t)

4 3


2 0

1

2

3

4

6 5 Nominal power PN (MW)

7

8

9

10

255

Figure 9.41 Results for design of a 2L-VSR as function of the nominal power when SFTM is selected: (a) switching frequency; (b) number of parallel connected IGBT modules; (c) nominal efficiency; (d) power density; (e) power-to-mass ratio.

256

Offshore Wind Farms

Symbols aa

Relative turn-on time of the converter bridge legs

aVsw0

Temperature coefficient of Vsw0

aRc

Temperature coefficient of RC

aEon/off

Temperature coefficient of Esw,on/off

dCI

Current imbalance rate

dHS,max

Maximum ratio of VolHSal to Volmod

diL

Ratio of peak-to-peak current ripple to maximum fundamental nominal current

dVdc

Ratio of peak-to-peak voltage ripple to DC voltage of the converter

dVac

Ratio of peak-to-peak voltage ripple to the peak fundamental voltage

DVsw

Static voltage deviation

DTHS,max

maximum allowable HS to ambient temperature

DILh

peak-to-peak ripple current

h

Efficiency

r

Power density

g

Power-to-mass ratio

Aw

Inductor winding window area

Acore

Inductor core area

APC

Capacitor plate area

BL

Inductor peak flux density

BLref

Inductor reference flux density

Bsat

Saturation flux density

C

Capacitance

CPV

Volume utilization factor

dPC

Capacitor plate separation distance

Esw

Commutation energy loss

Esw,on

Commutation energy loss at turn-on action

Esw,off

Commutation energy loss at turn-off action

Esw0,on/off

Commutation energy loss of semiconductor device at temperature Tj0

EBd

Breakdown electric strength of dielectric material

f

Fundamental frequency

fL1

Inductor fundamental frequency


257

fLref

Inductor reference frequency

feff

Effective frequency for a non-sinusoidal current waveform

fsw

Switching frequency

HS

Heat sink

iL1

Inductor fundamental current

iLh

Inductor ripple current

isw

Semiconductor device current

iswb

Current through the PSD at moment before turn-on action

iswa

Current through the PSD at moment after turn-off action

Ip,AVG

Average current in the parallel connection of PSDs

Ipsw

Peak current of the parallel connection of PSDs

Isw,AVG

Average conduction current of semiconductor device

Isw,RMS

RMS conduction current of semiconductor device

Isw,eq

Equivalent current of parallel connected PSDs for power losses calculation

Isw,Total

Total current of parallel connected PSDs

Iswa,AVG

Average current through the PSD at moment after turn-on action

Iswa,RMS

RMS current through the PSD at moment after turn-on action

Iswb,AVG

Average current through the PSD at moment before turn-off action

Iswb,RMS

RMS current through the PSD at moment before turn-off action

IbL

Inductor peak current

IL

Inductor RMS current

JL

Inductor RMS current density

Jmax

Maximum current density

kcdp

Derating factor

kvp

Safety factor of PSD for peak voltage

kvdc

Safety factor of PSD for DC voltage

kwc

Winding conductor fill factor

KEðon=offÞx

Polynomial regression coefficients for the current dependency of Esw,on/off

KHSx

Proportionality regression coefficients for heat sink volume

Kfanx

Proportionality regression coefficients for fan volume

KVLx

Proportionality regression coefficients for inductor volume

KrLx

Proportionality regression coefficients for inductor mass

Krwx

Proportionality regression coefficients for inductor winding losses Continued

258

Offshore Wind Farms

Krcx

Proportionality regression coefficients for inductor core losses

KVCx

Proportionality regression coefficients for capacitor volume

KrCx

Proportionality regression coefficients for capacitor mass

KUCx

Proportionality regression coefficients for capacitor series resistance

KSFT

Safety factor of thermal design

L

Inductance

MS

Modulation index

ma

Phase modulation function

MassTotal

Total converter mass

Massvalve

Mass of the power switch valve

MassL

Inductor total mass

MassC

Capacitor total mass

Mass(i)

Individual mass of the components

N

Number of switching actions

Nisxm

Number of internal semiconductor devices per module

np

Number of parallel connected devices

ns

Number of series connected devices

Pin

Power input

Ploss,mod

Total power losses of a PSD

PL

Inductor power losses

PC

Capacitor power losses

PεC

Capacitor dielectric losses

PUC

Capacitor resistive losses

PWL

Inductor winding losses

PcoreL

Inductor core losses

Pcond

Conduction losses

Psw

Switching losses

RC

On-state resistance of semiconductor device

RC0

On-state resistance of semiconductor device at temperature Tj0

RthHS

Thermal resistance of the HS for a given fan velocity

RthHS,min

Minimum thermal resistance of the HS for a given fan velocity

Rth,igbt

Thermal resistance of the IGBT device


259

Rth,diodet

Thermal resistance of the diode device

RthJC

Junction-to-case thermal resistance of the device

RthCH

Case-to-heat sink thermal resistance of the device

RsC

Capacitor series resistance at maximum hot-spot temperature

T

Fundamental period

Tamb

Ambient temperature

Tsw

Switching period

Tj

Junction temperature of semiconductor device

Tj0

Fixed reference junction temperature of semiconductor device

Tj,AVG

Average junction temperature of semiconductor device

Vbk

Nominal blocking voltage of semiconductor device

VCN

Capacitor-rated voltage

VCac

Maximum amplitude of the alternating voltage applied to capacitor

Vfan

Fan velocity

VDC,max

Maximum DC voltage of the series connected PSD array

VLL

Line-to-line RMS voltage

Vp,max

Maximum voltage amplitude to be blocked for the series connected PSD array

vsw

Semiconductor device voltage

vswb

Voltage in the PSD at moment before turn-on action

vswa

Voltage in the PSD at moment after turn-off action

Vsw0

Threshold voltage of semiconductor device

Vsw00

Threshold voltage of semiconductor device at temperature Tj0

VolTotal

Total converter volume

Vol(i)

Individual volume of the components

Volmod

Semiconductor module volume

Volvalve

Volume of a power switch valve

VolHS

Heat sink volume

VolHSal

Volume of aluminium/copper structure

Volfan

Fan volume

VolC

Capacitor total volume

VolL

Inductor total volume

VolPC

Volume of a plate capacitor

260

Offshore Wind Farms

References Backlund, B., Rahimo, M., Klaka, S., Siefken, J., 2009. Topologies, Voltage Ratings and State of the Art High Power Semiconductor Devices for Medium Voltage Wind Energy Conversion. IEEE, Lincoln, NE. Barrera-Cardenas, R., Molinas, M., 2015. Meta-parametrised Meta-modelling Approach for Optimal Design of Power Electronics Conversion Systems (PhD thesis red). Norwegian University of Science and Technology, Trondheim. Blaabjerg, F., Chen, Z., Teodorescu, R., Iov, F., 2006. Power electronics in wind turbine systems. Power Electronics and Motion Control Conference, 2006. IPEMC 2006, vol. 1, p. 11. Chivite-Zabalza, J., et al., 2013. Comparison of Power Conversion Topologies for a Multi-megawatt Off-shore Wind Turbine, Based on Commercial Power Electronic Building Blocks. IEEE, Vienna. Drofenik, U., Kolar, J.W., 2005. A General Scheme for Calculating Switching and Conduction Losses of Power Semiconductors in Numerical Circuit Simulations of Power Electronics Systems. Japan, IPEC, Niigata. EPCOS, 2012. Film Capacitors for Industrial Applications, s.l.: TDK. Friedli, T., Kolar, J.W., Rodriguez, J., Wheeler, P.W., 2012. Comparative evaluation of three-phase AC-AC matrix converter and voltage DC-link back-to-back converter systems. IEEE Transaction on Industrial Electronics 59 (12), 4487e4510. Fuji Electric Co, 2011. Fuji IGBT Modules Application Manual [Internett]. Available at: http:// www.fujielectric.com [Funnet 02.12.2014]. Helle, L., 2007. Modeling and Comparison of Power Converters for Double Fed Induction Generators in Wind Turbines. First red. Aalborg: Aalborg University, Denmark. Hurley, W.G., Wolfle, W.H., Breslin, J.G., 1998. Optimized transformer design: inclusive of high-frequency effects. IEEE Transactions on Power Electronics 13 (4), 651e659. Infineon, 2013. Technical Information IGBT-Modules FZ1500R33HE3 [Internett]. Available at: http://www.infineon.com/ [Funnet 05.12.2014]. Kolar, J., Ertl, H., Zach, F., 1990. Influence of the modulation method on the conduction and switching losses of a PWM converter system. IEEE Industrial Application Society Annual Meeting 1 (1), 502e512. Kolar, J.W., et al., 2010. Performance Trends and Limitations of Power Electronic Systems. IEEE, Nuremberg. Kolar, J.W., Round, S.D., 2006. Analytical calculation of the RMS current stress on the DC-link capacitor of voltage-PWM converter systems. Electric Power Applications, IEE Proceedings 153 (4), 535e543. Lai, R., et al., 2008. A systematic topology evaluation methodology for high-density three-phase PWM ac-ac converters. IEEE Transantions on Power Electronics 23 (6), 2665e2680. Lee, K., et al., 2014. Comparison of High Power Semiconductor Devices Losses in 5MW PMSG MV Wind Turbines. IEEE, Fort Worth, TX, pp. 2511e2518. Mirjafari, M., Balog, R., 2011. Multi-objective design optimization of renewable energy system inverters using a descriptive language for the components. s.l.: IEEE Applied Power Electronics Conference Exposition (APEC). Mirjafari, M., Balog, R., 2014. Survey of modelling techniques used in optimisation of power electronic components. IET Power Electronics 7 (5), 1192e1203. Mohan, N., Undeland, T., Robbins, W., 2003. Power Electronics: Converters, Applications, and Design, third ed. red. s.l.: John Wiley & Sons, Inc.


261

Preindl, M., Bolognani, S., 2011. Optimized design of two and three level full-scale voltage source converters for multi-MW wind power plants at different voltage levels. s.l.: IEEE e 7th Annual Conference IEEE Industrial Electronics Society IECON. Sullivan, C., 1999. Optimal choice for number of strands in a litz-wire transformer winding. IEEE Transaction on Power Electronics, 14(2), pp. 283e291. Volke, A., Hornkamp, M., 2012. IGBT Modules: Technologies, Driver and Application, Second red. Infineon technologies A.G., Munich. Wen, B., Boroyevich, D., Mattavelli, A.P., 2011. Investigation of tradeoffs between efficiency, power density and switching frequency in three-phase two-level PWM boost rectifier. Birmingham, IEEE; Proceedings of the 2011e14th European Conference on Power Electronics and Applications (EPE 2011). Xu, D., Lu, H., Huang, L., Azuma, S., Kimata, M., Uchida, R., 2002. Power loss and junction temperature analysis of power semiconductor devices. IEEE Transactions on Industry Applications 38 (5), 1426e1431.


Design of offshore wind turbine towers

10

R.R. Damiani RRD Engineering, Arvada, CO, United States

10.1

Introduction

Following the IEC 61400-3 classification [1], the tower is part of the support structure (SSt), which, for offshore applications, includes the substructure (SbS) and the actual foundation embedded in the soil. The tower is a relatively simple component when compared to the moving parts in the rotor nacelle assembly (RNA). Yet, the SSt is responsible for approximately 16% of the installed cost on land and almost 20% offshore [4]. Hence, there exists ample margin to effectively reduce overall project costs by optimizing the tower and SbS configurations. Furthermore, the continuous growth of turbine RNAs and the increasingly more challenging siting at sea, as for instance in tropical cyclone regions, push the boundaries of the experience gained in tower engineering. For these reasons, the topic of SSt design has become more central in the ongoing research and development efforts toward a lower levelized cost of energy (LCOE). Fundamentally, the designer is left with the problem of finding an appropriate distribution, along the tower length, of mass and stiffness properties that ensure safe turbine operation under all prescribed external conditions, including actions from the environment and from the interaction with the grid and the control system. The turbine hub height dictates the necessary length of the tower, and ideally one would strike a balance between gains in energy capture at higher altitudes and the costs of a taller tower. Historically, tower lengths were set at approximately one rotor diameter. Installations at low wind sites and at sea no longer follow this rule of thumb. For wind development to be profitable in less windy, forested areas, for example, higher hub heights that take advantage of higher wind speeds and less turbulent atmospheric layers are required. At sea, on the other hand, lesser wind shear values and the SbS interface terminating at several meters above the still-water level (SWL) favor shorter towers than on land for a given hub height. Nonetheless, solving the design problem in either case is non-trivial. Offshore installations, given the inherent high balance of station (BOS) costs, promote the largest turbines and are generally characterized by more significant tower-head masses and ultimate thrust values than on land (eg, 350þ t and 1800 kN for a typical 6-MW offshore machine). These extreme loads are also to be combined with a particularly corrosive environment, the possible presence of other sources of loading (eg, floating ice), and extraordinary fatigue loads coming from some 109 cycles due to rotor aerodynamics and some 108 cycles due to wave loading alone. For these reasons, the designer must ensure that the overall system simultaneously meets Offshore Wind Farms. http://dx.doi.org/10.1016/B978-0-08-100779-2.00010-6 Copyright © 2016 Elsevier Ltd. All rights reserved.

264

Offshore Wind Farms

several structural criteria: with regards to the response to external and internal excitation, the system must achieve the prescribed modal behavior, avoiding risk of instabilities or resonance; strength and deflection limit states must be verified while also ensuring an economically optimized load distribution and material utilization throughout the tower (and SbS below); manufacturability constraints, for instance on the weldable wall thickness of steel cans and the rollability of plates must be verified; finally, transportability and installation loads and processes must be examined and quantified. As far as detail design, important aspects must be covered such as: the interface with the RNA and the transition piece (TP), also flange and weld design, access door and manholes, the housing of power electronics, hoists and manlifts, and the needed coating protections. Loads, generally determined through aeroelastic simulations, are to be applied to finite element models to assess the three-dimensional (3D) stress state. The same models are used to confirm modal characteristics. In this case, an additional difficulty comes from the assessment of the effects associated with the soilestructure interaction (SSI) and of the overall stiffness offered by the SbS and pile subsystem. It is clear that even confining one’s view to the tower alone, multiple disciplines (eg, civil and mechanical engineering, structural dynamics, metallurgy) must be invoked to effectively and comprehensively tackle the design problem. This chapter offers an overview of the configurations and layouts currently available for wind towers, of their design process, and of the key engineering aspects that need to be addressed for a reliable and effective design. No individual reference or computer software can replace the experience and good judgment of a well-versed tower engineer, and this chapter wants to highlight the importance of accounting for the effects on system dynamics and installed costs of the SSt design choices. The chapter is organized as follows. Section 10.2 presents a gallery of tower configurations and discusses the limitations for land-based systems that are pushing toward new materials and designs that can also be utilized offshore. Specific offshore requirements that make offshore wind turbine (OWT) SSts uniquely challenging, but also prone to a number of options and innovations, are also mentioned in this section. The main standards of reference for design and certification are presented in Section 10.3, which also discusses the importance of quantifying the reliability of offshore systems in tropical cyclone regions and to strike a balance between capital investment costs and those of repairs. In Section 10.4, the typical processes used to determine structural loads are presented, together with an overview of the sources of loading and the distinction between coupled and uncoupled load analyses and systems. The importance of controlling the system eigenfrequency is shown as being accompanied by the complex ramifications in terms of aerodynamic damping, fatigue loads from aerodynamics and hydrodynamics, and the effects on these aspects of turbine lifetime availability and site conditions. Section 10.5 discusses a possible approach to the preliminary sizing of the primary steel of a tubular tower. The major factors of the detailed design of flanges and welds are also presented. Secondary steel design, including damping devices and corrosion protection strategies, is discussed in Section 10.6. Finally, in Section 10.7, key facets of systems engineering and optimization are discussed, which emphasize the importance of a multidisciplinary


265

optimization of the wind plant system over that of individual components, in order to achieve the ultimate goal of a minimum LCOE. Concluding remarks are presented in Section 10.8.

10.2

Function and types of towers

Since the inception of the wind industry, the towers have been considered as one of the key components for wind turbines as they perform two fundamental functions: (1) provide access to a favorable wind resource by supporting the rotor at a sufficiently high hub height; and (2) provide a safe and reliable load path from the turbine to the foundation. In the following subsections, an overview of layout options for both land-based and offshore SSts is given. These alternatives differ in geometric layout and load paths, utilized materials, or both. Regardless of these choices, the towers must guarantee that the system modal characteristics are kept within the acceptable bounds and away from the turbine excitation frequency bands. Special attention must also be paid to the fabrication costs, and the transportation and installation procedures, which may make one alternative more or less favorable for a specific site and turbine configuration. While the one-size-fits-all model is attractive, rarely one particular design can be applicable to many different environments and applications. In fact, it is becoming progressively more crucial to optimize the entire turbine from a system perspective, where all project aspects, from the structural performance to the BOS costs, are considered.

10.2.1 Full lattice, tubular, and wooden towers Several configurations have been explored throughout the history of the wind power industry, and several different criteria have led to the widespread use of the tubular design. Lattice structures, quite common in the first installations, see Fig. 10.1, have mostly been abandoned on the account of aesthetics and environmental concerns. From an environmental perspective, these towers offer roosting/nesting places for birds and thus may raise the probability of bird strikes. Tubular towers are normally made up of cylindrical or conical frustum segments of 20e30 m in length, which are transported separately and bolted together at their ends (flanges) on site. As such, tubular towers allow for an enclosed space in which to locate power electronics and switchboards, and protect maintenance personnel from the environment. A lattice tower, however, allows for a stiff, yet very light structure by placing the main steel away from the neutral axis. The first lattice towers were designed following the electrical transmission line experience, where angle iron components were bolted or welded together to form the lattice. Larger turbines, as the 160-m hub height, Fuhrlander FL 2500 shown in Fig. 10.2, require larger members, and the bolted connection at the joints becomes prohibitive especially for numerous applications such as those in large wind farms. Checking bolt torque levels, clearing

266

Offshore Wind Farms

Figure 10.1 Example of free-standing lattice tower used for old wind turbines at Altamont, California (a) and Tehachapi, California (b) wind farms. Photos courtesy of NREL Pix Library, numbers 17329 and 18457.

the towers of debris or ice, and even safe access for maintenance staff can be too expensive or difficult to achieve. Also to note is the need for a TP to house the yaw bearing. The design of this component is critical as it must allow for the transition from polygonal to circular cross-section and for the safe transfer of the distributed load at the yaw bearing to the concentrated loads in the main lattice legs. Slender masts with guy wires have been used for turbines of submegawatt size, as seen for example in Fig. 10.3. Guyed towers are light, easy to transport, and allow for the entire structure to be brought down to the ground for maintenance. Unfortunately, for large turbine sizes, the guys would require a very large footprint, making the installation more cumbersome, and given the number and wire diameter sizes would also affect the quality of the wind, not to mention the impediments associated with environmental issues. Fiber-reinforced polymer towers (see Fig. 10.5) and wooden towers have also been used for water-pumping windmills in the past, and at present for small wind turbines. Today, there are proponents of engineered wood for utility scale turbines, and one such


267

Figure 10.2 The Fuhrlander 2.5 MW Laasow (FL 2500) was touted as the world’s tallest wind turbine until a few years ago. The lattice tower is 160 m high and the blade tip reaches 205 m. € The new MHI-Vestas V164e8.0-MW turbine is being tested in Osterlid, Denmark atop a w140-m tower, and the blade tip reaches 220 m. Copyright http://en.wikipedia.org/wiki/Fuhrl%C3%A4nder_Wind_Turbine_Laasow.

€ and deployed in Germany by Timber-Tower, example has been certified by TUV GmbH (see Fig. 10.4). Wood offers some economical advantages thanks to the ease of laminate manufacturing, competitive strength/weight ratio, and stable costs. Transportation issues are also relaxed and direct local sourcing is a possibility. However, wooden towers have not found a great deal of application in the industry. Composite and hybrid towers are currently being investigated (eg, Refs [5e8]). In principle, composite materials may render better fatigue characteristics, increased structural damping, reduced logistics costs, and possibility of on-site construction. However, it still has to be proven that these advantages offset manufacturing and material costs to make this option viable.

10.2.2 Manufacturing and installation challenges on land The steady decrease in prime real estate with high wind resource at low altitude, grid availability, and no environmental concerns or public interaction, translates into the need for higher hub heights. In many sites where the wind resource at 80 m

268

Offshore Wind Farms

Figure 10.3 The GEV MP C series consists of two-bladed, downwind wind turbines mounted on tilt-up, guyed towers. Rotor diameters reach 32 m and the hub height reach 60 m. Vergnet Eolien http://www.vergnet.com/en/gevmpc.php.

Figure 10.4 Example of a 100-m wooden tower for a Vensys 1.5 MW turbine in Germany. TimberTower GmbH, Hannover, Germany. http://www.timbertower.de/tower-construction/.

(the standard hub height for most utility scale wind turbines in the U.S.) is only marginal, the presence of a significant wind shear means that the resource at 120e150 m hub height becomes economically feasible for development. For example, it is calculated that 1800 GW of wind power potential across 237,000 square-miles of the


269

Figure 10.5 Examples of GFRP towers: (a) small wind turbine application (Photo courtesy of NREL); (b) utility scale turbine tower getting ready for structural testing. http://www.cres.gr/megawind/COMPOSITE_MATERIAL_TOWER.htm visited on 2/20/2015.

U.S., or an area roughly the size of Texas [9], could be unlocked by wind turbines with hub heights of up to 140 m. Furthermore, the ever-growing size of turbine rotors necessary to guarantee economical capacity factors in less windy areas, demands taller and bigger towers. Finally, repowering old wind farms with new, larger turbines, and the possibility of adopting staggered siting with different hub heights to reduce wake losses in a wind array, call for taller towers too. Tubular steel towers have reached a limit of applicability on land: larger turbine heads require large base diameters to meet safe frequency-band criteria, but transportation limitations (eg, bridge clearance) impose a difficult limit on the maximum hauling girth (4.3 m in the U.S.). Increasing wall thickness may partially mitigate the problem, but it can only be achieved through expensive fabrication processes (because of plate rolling and welding constraints). Beside guaranteeing permits for blade tips approaching 200 þ m above ground level (AGL), the challenge for large towers on land is thus how to address the transportation challenges, and also how to facilitate turbine erection to those heights. In the U.S., for instance, there are only some 10 crawler cranes in the class (1250 and 1600 t) needed for erections of 3- and 5-MW turbine nacelles and towers [9]. Future designs will necessarily be “designs for production” accounting for new and on-site manufacturing techniques, different materials, and/or innovative transportation and installation strategies.

10.2.3 The promise of concrete Two alternatives that have been successful on land and that can at least partially bypass the above logistical constraints are the reinforced concrete tower and the hybrid concreteesteel tower. In the latter, the concrete base (a pedestal) supports the upper, more conventional, steel portion. There are several advantages to utilizing concrete.

270

Offshore Wind Farms

Concrete is an inherently durable material that can retain the required engineering properties under various environmental conditions; it offers corrosion protection to the reinforcing material (steel or composite), and presents high fatigue resistance and damping characteristics; it can be tailored both in inner constituents (for example to achieve different degrees of strength or to achieve crack-resistance and self-healing properties) and in the shape of the casting and arrangement of the reinforcement [10]. Concrete can be sourced locally and is also easily recyclable as aggregate material. Quality control on precast panels can guarantee tight tolerances and finishes, whereas established formwork solutions and mobile mixing plants can enable on-site construction without transportation impediments. The versatility of concrete can be used by the designer to tailor structural properties along the height of the tower, to take advantage of modularity in precast panels, and, through the employment of prestress, to achieve high hub heights with minimum steel quantities. The possibility of tuning the system eigenfrequencies by changing the amount of prestress is very attractive. Furthermore, by adapting the level of prestress via post-tensioned tendons, installations could be repowered for larger machines as long as the foundation is designed to cope with larger stresses. It is argued that some form of prestressed reinforced concrete (pre- or post-tensioned, and steel or composite fiber-reinforced) is the only technology that can unlock the highest hub heights while bypassing the land transportation restrictions. The 7.5-MW E-126 Enercon (see Fig. 10.6) features a prefab concrete tower with a 14.5 m base diameter and a 135 m hub height. Obviously, even though transportation logistics are overcome by interlocking concrete panels, the tower and turbine erection remains a formidable challenge in the quest for low project costs.

Figure 10.6 Example of a precast reinforced concrete tower: Enercon E126 100-m 7.5-MW turbine tower installed in Germany (a); zoom of the tower base (b). Photos courtesy of Patrick Fullenkamp, GLWN.


271

10.2.4 From onshore to offshore The vast experience gained onshore with wind turbine generators (WTGs) can only be partially applied to OWTs, as important differences exist between onshore and offshore. The variety of SSts and foundations, the addition of hydrodynamic loads, extreme corrosive environments, the vital need to minimize maintenance more than on land due to the extremely expensive operation and maintenance (O&M) are just a few of these differences. Oil and gas (O&G) experience can certainly be tapped, but differences in the systems between wind and O&G, such as larger wind loading, structural dynamics, and structural non-linearities still make the engineering of an OWT very unique. For this reason, the so-called fully coupled approach, where the entire system (from the WTG to the foundation, including effects of aerodynamics, hydrodynamics, servo-mechanics, and structural dynamics) is analyzed with ad hoc tools is the preferred and most rigorous design method. From a loading standpoint, important differences can be listed. Offshore sites tend to have steeper shear and less turbulent wind profiles due to reduced roughness; this, in addition to affordably lower hub heights, decreases the aerodynamic periodic loads. However, offshore climates are generally characterized by Weibull wind distributions with larger scale and shape factors, which lead to higher mean wind speeds and higher overall probabilities of large wind speeds and gusts. Additionally, it can be expected that array wake effects on turbine inflow are more important offshore, as the wake structures decay more slowly in a less turbulent environment. Therefore, one may expect a greater power output but also higher mean load levels when going offshore. Moreover, the absence of significant damping when the rotor is parked or idling can be an important design situation for offshore towers. In some situations, for instance with large hub heights and deep water sites, the hydrodynamic loads may dominate the fatigue damage equivalent loads (DELs). If a turbine is shutdown in high winds, the associated severe sea state may induce vibrations in the SSt, that are significantly underdamped. A similar situation may occur during operational conditions if wind and waves are misaligned, in which case the side-to-side oscillations are also lightly damped. This aspect requires more investigation than on land, where idling is normally a benign situation; hence, ad hoc dynamic control strategies and damping devices may need to be employed offshore. Further aspects specific to OWT SSt include scour, marine growth, and potential vessel impact. Scouring may reduce the natural frequencies of the OWT system with possible resonance-band impingement. Measurements have revealed scouring depths of up to 1.3 times the diameter of the embedded pile. Marine growth will increase mass and hydrodynamic loads in the submerged part of the SSt. Marine growth thickness and depth extents are normally given by site-specific design basis documents or relevant standards (eg, Ref. [11]). Vessel impact is intended to be resisted by the plastic deformation of boat fenders, and accidental impacts elsewhere should not exceed yield strength. For offshore applications, the transportation constraints are largely relaxed if manufacturing has water (quay-side) access. For this reason and because of the experience gained with systems onshore, mostly tubular steel towers have been used for offshore wind turbines. The towers are connected to the SbSs, which transfer wind

272

Offshore Wind Farms

and inertial loads from the tower base, and wave and current hydrodynamic loads to the foundation at the seabed. The connection between tower and SbS is ensured by the TP, a component that can take various shapes, but that is essential in guaranteeing the safe and reliable operation of the wind turbine. The simplest way to put a wind turbine at sea is to consider the typical tubular tower and to extend it all the way below the seabed via the so-called monopile configuration. Since the pile must be driven into the soil via hammers or vibrohammers, it is convenient to separate the pile from the tower above and to connect the two through a TP, normally grouted to the pile. More recent innovations incorporate the TP within the monopile component, which expedites the installation, but requires larger thickness at the top of the pile to guarantee that the action of pile-driving does not push the tower-SbS connecting elements out of tolerance. Deeper waters and larger RNA masses associated with larger turbines have stringent modal requirements that necessitate very large monopiles, which make them uneconomical. Other fixed-bottom configurations include lattice SbSs (tripod or jackets from the O&G experience) and gravity-based foundations (GBFs) (see Fig. 10.7). The TP in the case of jackets must transfer loads to the primary members (legs) of the SbS and promote the transition from multiple-member to single-tube shapes. Concrete towers

Figure 10.7 Examples of fixed-bottom systems: (a) from left to right, monopile, tripile, tripod, jacket, and gravity foundation. Examples of real offshore installations of turbines on fixed-bottom SbSs: (b) monopile; (c) jacket; (d) tripile; (e) shows GBFs under construction at the quayside.


273

have not seen application offshore as of yet, but reinforced concrete has been already used in GBFs (eg, Middelgrunden and Nysted wind farms) and TP grouted connections may be considered special cases of reinforced concrete structures. It is conceivable that prestressed, post-tensioned concrete towers will find applicability, especially married to GBF SbSs and high hub heights, given the advantages mentioned in Section 10.2.3. In particular, prefabricated panels can be produced on a dry-dock onshore and easily transported to the site where post-tensioning can be applied during installation. The aggressive corrosive environment must be counteracted with adequate protection of the steel reinforcements and tendons. This translates into constructive measures (minimum protective distance of concrete outer surface from rebars) and concrete formulation (strength and density of the concrete). For GBFs, transportation via towing operations and controlled sinking further require floating stability checks. Floating SbSs, or platforms, include semisubmersible (stability guaranteed by buoyancy and restoring mooring), spars (stability guaranteed by gravity), and tension leg platform (TLP) (stability guaranteed by excess buoyancy) (see Fig. 10.8). These systems promise economic advantages, especially in deeper waters, where material and installation costs make fixed-bottom structures economically unfeasible. Of course, floating configurations have their own unique challenges, and again are quite different in terms of dynamic behavior from the O&G counterparts. The towers that are mounted on floating platforms are still designed according to the same design procedures as those of fixed SbSs, but wave excitation and hydrodynamic loading become even more important. With floating OWTs, controls are key for dampening dangerous oscillations throughout the system and the tower in particular. Currently installed examples include turbines on a spar and a semisubmersible (Statoil’s Hywind [12] and Principle Power’s WindFloat [13]). A hybrid configuration which promises certain advantages is the hinged, floating tower [11]. Guy wires help the stabilization of the tower from above the mean sea level to the seafloor. An example of a floating tower is shown in Fig. 10.9.

Figure 10.8 Examples of floating OWT SbSs: spar, semisubmersible, and TLP. Illustration by Joshua Bauer, NREL.

274

Offshore Wind Farms

Figure 10.9 Example of hinged, floating tower. SWAY’s system consists of a downwind turbine supported by a floating tower/spar anchored to the seabed via a tension-torsion leg, which is equipped with a passive, subsea yaw-swivel. The entire structure can yaw around this point. The guy wires are connected to that point and reduce the stress on the tower and spar which can then be made lighter. http://www.sway.no.

Finally, another alternative design is an all-lattice structure, from the seabed to the hub (see Ref. [14] and Fig. 10.2). In this case, there is no distinction between the tower and the SbS, but a TP still needs to be devised to be located at the base of the nacelle to accommodate the yaw bearing and RNA slewing ring. Given the importance of stiffness requirements, the individual members must be of generous proportion to guarantee adequate shear transfer, and the connections at the joints rely on welds that must be carefully verified and inspected. Therefore, fabrication of lattices, including jacket and tripod SbSs, is less amenable to automation and the potential savings in mass can be partially offset by labor costs. Cast joint-cans, where only circular welds have to be made, can mitigate the issue and are also intrinsically more resistant to fatigue. One important aspect to underline is that the SSt, encompassing both tower and SbS, is site-dependent. Whereas the turbine itself is designed against the specifications in the standards (eg, class IB from [1]) and therefore can be deployed at a number of different sites that feature wind conditions within the turbine class envelope, the SSt will be optimized based on bathymetry, wave and current regimes, modal requirements, and wind statistics.

10.3

Standards of reference

Wind power generators and their SSts are dynamically connected and need to be analyzed as a combined system. For this reason, one should seek harmonization between the design of the power generation equipment and that of the SSt. As a result,


275

for the designers of an OWT and its components, navigating the number of codes and standards may be a daunting undertaking. A few guidelines can be used to help in the selection of the design standards. So far, the majority of offshore wind installations has taken place in Europe, where International Standardization Organization (ISO) and European-specific standards (eg, Ref. [15]), expressed in load resistance factor design (LRFD) format, are widely adopted. In other regions, the choice of design practices and material references should align with locally based guidelines, construction permitting processes, authorities’ protocols, material specification, and construction yard practice. Steel properties for shell, flanges, and bolts are commonly given by Refs [16,17]. In the U.S., the American Wind Energy Association (AWEA) issued Ref. [18], which provides overall guidance for offshore wind installations in U.S. waters. For example, Ref. [18] specifies that a 1000-year maximum wave crest level, as also indicated in Ref. [19], should be considered when determining the deck height in the U.S. waters prone to tropical cyclones, which in turn yields the elevation of the tower-base flange. AWEA [18] recognizes that a number of design aspects of OWTs are not adequately covered in existing codified guidance. Where local codes and standards may not suffice, the designers need to refer to internationally recognized codes and regulations that, subject to the owner’s engineer’s approval, may lead to a safe and reliable design. For instance, standards such as Refs [19,22] are examples of normative references cited by both Refs [18,1]. Offshore towers are fundamentally ruled by Refs [23,1,19], for the pure structural compliance, and by a number of different codes for corrosion protection and manufacturing (eg, Refs [24e26]). In Europe, Refs [3,15] have also been used in place or in conjunction with Ref. [1]; for the SbS and for interface component fatigue limit state (FLS) design (eg, TP and grouted connections), Ref. [11] may also be used. In order to determine the metocean conditions, such as wind and wave regimes per Ref. [1], and to translate them into design parameters, Ref. [27], now aligned with Ref. [28], can be employed. Structural design checks normally are to the ISO 19900 series [3,2]; the latter two are used for global and local/shell buckling verifications. Refs [15,29] also provide guidance for resonance avoidance and SSt frequency placement (see Section 10.4.1.2). Table 10.1 provides a possible selection and prioritization of codes and standards that can be used for design and certification preparation of the SSt; note, however, that many other codes could and should be used in the design of structural details, and local ordinances may require a different hierarchy than the one proposed here. Finally, all materials, fabrication, tolerances, workmanship, etc. shall conform to the codes adopted for design. The SSt designer shall comply with all necessary technical and procedural requirements advised by the owner’s engineer, the WTG manufacturer and the certification and verification agency (CVA). In particular, welding specifications and detailing for the SSt should adhere to Refs [30,39]; locally accepted codes may also be used as long as they meet the approval of the CVA, because they represent an integral part of the WTG certification.

276

Offshore Wind Farms

Table 10.1 Possible prioritization in the codes and standards for the design of an offshore tower References

Title

Comments

[18]

AWEA offshore compliance recommended practices

Broad design intent and guidance.

[1]

IEC 61400-3 wind turbinesdPart 3: Design requirements for offshore wind turbines

Arrangement of DLCs. Appurtenances and secondary steel guidance.

[23]

IEC 61400-1. Wind turbinesdPart 1: Design requirements

Arrangement of DLCs, OWT component design general guidance.

[21]

ISO 19902:2007dpetroleum and natural gas industriesdfixed steel offshore structures

General guidance and structural checks for SbS, transition piece, secondary appurtenances. Fatigue curves.

[3,15]

GL IVdPart 2 guideline for the certification of offshore wind turbines

DLCs, structural checks for global tower buckling, and modal requirements.

[2]

Eurocode 3: Design of steel structuresdPart 1e6: General rulesdsupplementary rules for the shell structures

Shell buckling checks.

[30]

Eurocode 3: Design of steel structuresdParts 1e9: Fatigue

Fatigue design of welds, detail categories, SeN curves.

[31]

ANSI/AISC 360-10dspecification for structural steel buildings

General guidance on steel construction, weldments, stability, etc.

[20]

ISO 19901-3:2014dpetroleum and natural gas industriesdspecific requirements for offshore structuresdPart 3: Topsides structure

Deck primary and secondary steel structural design including crane pedestals, masts, towers, booms, helicopter landing pads.

[19]

API RP 2A:2014dplanning, designing and constructing fixed offshore platformsdworking stress design

General guidance on DLC combination, member and joint structural checks, RSR checks.

[11]

DNV-OS-J101: Design of offshore wind turbine structures

Grouted connections. Safety factors. Secondary appurtenances.


Table 10.1

277

Continued

References

Title

Comments

Reinforced concrete structures codes and standards [32]

ACI 318-14dbuilding code requirements for structural concrete and commentary

Reinforced concrete SSt.

[33]

ACI 357R-84dguide for the design and construction of fixed offshore concrete structures

Gravity based foundations.

[22]

ISO 19903:2006dpetroleum and natural gas industriesdfixed concrete offshore structures

General guidance for offshore RC structures.

[34]

Eurocode 4: Design of composite steel and concrete structures

[35]

Eurocode 2: Design of concrete structures

To be used in conjunction with Ref. [34]

[36]

Model code for concrete structures 2010

High load cycle number fatigue design of pre-stressed concrete.

[37]

NS 3473dconcrete structuresddesign rules

High load cycle number fatigue design of pre-stressed concrete.

Floating wind turbine design standards [11]

DNV-OS-J103: Design of floating wind turbine structures

Materials and PSFs. Station keeping.

[38]

Guide for building and classingdfloating offshore wind turbine installations

Floating wind turbine design.

Additional reference standards [39]

AWS structural welding codedsteel

Welding specifications and detailing.

[40]

VDI 2230 Part 1: Systematic calculation of highly stressed bolted joints

Bolted connection specifications and detailing.

[41]

DNV-RP-C203: Fatigue strength analysis of offshore steel structures

Guidance on fatigue analysis of offshore structures.

[42]

ABS guide for building and classingdbottom-founded offshore wind turbine installations

Special DLCs for hurricane-prone regions. Continued

278

Table 10.1

Offshore Wind Farms

Continued

References

Title

Comments

[43]

ABS guide for buckling and ultimate strength assessment for offshore structures

Member design, buckling assessment, material PSFs.

[44]

ABS guide for fatigue assessment of offshore structures

Member design, fatigue design, SeN curves.

[27]

API RP 2MET: Derivation of metocean design and operating conditions

General guidance on metocean conditions derivation.

[45]

API RP 2SIM: Structural integrity management of fixed offshore structures

[46]

API RP 2GEO: Geotechnical and foundation design considerations

General guidance on foundations design.

[47]

NORSOK classification note no. 30.4: Foundations

For foundation stiffness, lateral capacity etc.

[48]

IEC 60364dElectrical installations of buildings Part 5e54: Selection and erection of electrical equipment earthing arrangements, protective conductors and protective bonding conductors

Grounding protection.

[49,50]

ISO 12944 paints and varnishesdcorrosion protection of steel structures by protective paint systems

Corrosion protection and coatings.

[24]

NORSOK M-501: Surface preparation and protective coating

Coatings.

[16]

EN 10025: 2004dEuropean structural steel standard

Structural steel specifications.

[17]

EN 14399: High-strength structural bolting assemblies for preloading

Bolt assemblies properties.

10.3.1

Advanced standards development

In the U.S., Ref. [18] defines OWTs as L2 structures following Ref. [51] (see Table 10.2), associated to medium consequences of failures in the event of structural collapse. The consequences of failure are mostly linked to lost revenues and potentially adversary national policy consequences for an industry that is still very young. The exposure category has direct impact on the reliability level assumed (or 1 probability of failure).


279

Table 10.2 Exposure levels based on life safety and consequence categories Consequence category Life-safety category

C1 high consequence

C2 medium consequence

C3 low consequence

S1

Manned non-evacuated

L1

L1

L1

S2

Manned evacuated

L1

L2

L2

S3

Unmanned

L1

L2

L3

From ISO 19900: 2013-Petroleum and Natural Gas Industries e General Requirements for Offshore Structures. Revision to 2002 edition, 2013.

There is still ongoing debate on what the appropriate level of reliability in U.S. waters should be Refs [42,52,53]. Especially in hurricane-prone waters, in fact, it can be argued that the level of reliability achievable following Ref. [1], which was calibrated to European conditions, might be less than that in the North Sea [42]. A counterargument states that the reliability index of northern Europe might be overly conservative, as it was originally derived from the O&G experience in the Gulf of Mexico. These aspects will be revisited by the standards’ technical committees in the upcoming years, and the new edition of Ref. [1], for instance, will have informational content on design in tropical cyclone regions and on adequate approaches to guarantee an appropriate level of safety (eg, robustness and reserve strength ratio verifications in line with Ref. [19]). One method to calibrate the reliability of the SSt design makes use of the so-called hazard curves. The curves help identify the risk of failure via an assessment of the overload hazard as a function of the metocean event return periods (RPs) [53] (see also Fig. 10.11 for a few examples of hazard curves). In Fig. 10.11(a), the ratio of the overturning moment at the base of the SSt (or tower) calculated as the response to a ultimate limit state (ULS) event at a given RP, to the same load for a 50-year RP, is graphed. The latter is the basic RP indicated, for example, by Ref. [1] for maximum gusts and wave loads. In regions where tropical cyclones (hurricanes and typhoons) are present, the hazard of overloading the structure could be higher than in extratropical storm regions (see also Fig. 10.10). This is indicated by the steeper slope of the “Gulf of Mexico” curve in Fig. 10.11(a) as opposed to the “New England” curve. A set of hazard curves can also be used to evaluate the necessary load factor (LF) to achieve a certain reliability level. While the topic is more complex than this chapter may cover, the concept may be shown by the curves in Fig. 10.11(b). In that figure, the reliability level associated with a 500-year RSR (from Ref. [19]) is used to determine the necessary LF for an SSt (four-legged jacket and tower) in a tropical-cyloneprone region. As can be seen, the necessary PSF is larger than the 1.35 recommended by Ref. [1].

280

Offshore Wind Farms

Figure 10.10 Tower collapse following typhoon event. Courtesy of Dr. Xao Chen, Chinese Academy of Science.

(a) 2 1.8 1.6 Load ratio

1.4 1.2 1 0.8 0.6

New England

0.4

Gulf of Mexico

0.2 0 10

Load ratio

(b)

100 Return period (years)

2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

1000

DLC 1.6a DLC 1.6b Required PSF = 1.56 IEC PSF = 1.35 50

500 Return period (years)

5000

Figure 10.11 Examples of hazard curves for a typical tower and four-legged jacket SSt: (a) for two regions in the U.S., more (Gulf of Mexico) and less (New England) prone to tropical cyclone overload hazard, respectively; (b) an example showing the required LF for an installation in a hurricane-prone region in order to meet the 500-year RSR recommended by Ref. [19].


281

2.5

Normalized cost

2

Cost of failure increases

Installation cos increases

1.5

Lower LF and/or nominal return period

1

0.5

Higher LF and/or nominal return period

Min cost

0 0

0.5

1 Normalized load

1.5

2

Figure 10.12 Costebenefit analysis schematics.

Over the years, the O&G industry has conducted several rigorous reliability studies that have also been validated by hundreds of installations [19,54e60]. The offshore wind industry could potentially rely on that experience, but OWTs are significantly different from O&G platforms, both from a structural and an economic point of view, and will need to be studied appropriately to arrive at an optimum reliability and cost balance. Hazard curves and reliability studies are tools that can in the end lead to a more balanced design where the overall life cost of the project (comprising both installed and lifecycle cost) is minimized as graphically shown in Fig. 10.12. As a general rule, the early involvement of the project CVA is encouraged as that may further aid in the selection and prioritization of the standards to be used and allow the designer to fine-tune the PSFs to achieve the most economical design with an optimum level of reliability.

10.4

Design spiral process and loads’ analysis

The conventional SSt design process starts with the compilation of the so-called “design basis” document, which details turbine parameters (loads, dimensions, project lifetime, PSFs, etc.), environmental site-specific conditions, and guidance on the choice of reference standards. Based on these specifications and on an initial conceptual layout, a preliminary design is achieved, which verifies simple structural checks (modal performance and ULS criteria, as discussed in the folowing sections). These first few steps in the design spiral process (see Fig. 10.13) heavily rely on the experience of the engineering team. Then, the spiral process itself tends to guide the choices based on the structural and functional requirements that must be met by the SSt. At every successive turn of the spiral, a more focused and detailed analysis

282

Offshore Wind Farms

Turbine and substructure data

FEM and optimization

Update and refine design each cycle Door and internals

Draft tower geometry

Final design Flange and interface design

Modal analysis

Deflection and blade tip clearance checks

Buckling/ strength checks

Figure 10.13 Typical design cycle for the tower.

cycle is accompanied by: modifications to the geometry, new and additional loads analyses, and refined limit state assessments. After each cycle, more design criteria are satisfied and the tower layout converges toward a finalized geometry. Computer-aided design/manufacturing (CAD/CAM) models can help in this process because they allow for parametric storing and updating of the data, and for an efficient exchange of geometry, material, and infrastructure information. In parallel to this process, the interaction among the designers of all OWT components should be encouraged to arrive at a fully integrated design and to verify that choices in one of the parts do not penalize load states in another. As any other component of an OWT, the tower must be designed to sustain the loads it will encounter throughout the lifetime of the system with a sufficient safety margin and based on a target value for the structural reliability. The tower response must thus be analyzed under all possible scenarios representative of load situations encountered in real life and based on the L2 exposure level. The loading scenarios must include operational and parked conditions. Operational DLCs include normal operation and power production as well as start-up, shutdown, and fault cases. DLCs are constructed from a combination of relevant design situations and external conditions. To this end, the DLCs are normally provided by the design and certification standards to be used in conjunction with site-specific data. When combined with that data, standards such as Refs [1,11,15,23] describe the appropriate profiles,


283

spectra, and parameters for factors such as wind, wave, current, and ice, which must be considered in the loads analyses. The DLCs account for a minimum number of combinations of design situations and external conditions such as: • • • •

normal operation and normal environmental conditions, normal operation and extreme environmental conditions, fault situations and appropriate external conditions, and transportation, installation, maintenance and appropriate external conditions.

Table 10.3, for example, summarizes the DLCs and environment scenarios prescribed by Ref. [1]. In general, tower loads are determined via numerical aero-hydro-servo-elastic (AHSE) simulations carried out via dedicated computer-aided engineering (CAE) tools. These tools simultaneously model aerodynamics, hydrodynamics, structural dynamics, and control system dynamics within each DLC of interest. Simplified calculations may be used in the very first phases of design, but the complexity of offshore systems outright bans any attempt at completing the system design without integrated approaches. During the initial design iterations, however, the designer’s experience may help in the selection of a few key CAE simulations and (design driving) DLCs to rapidly narrow down the pool of candidate tower layouts. The total number of simulations for a more complete loads’ analysis reaches the few thousands. This is because various combinations of wind speed and wave/current/ice conditions must be examined, and more realizations are necessary to reach statistical significance within turbulent wind regimes and stochastic wave simulations. Therefore, the full suite of loads’ simulations is normally carried out on just one or two select final configurations. Within the LRFD design, the loads’ analysis results are then used to verify ULS, FLS, and service limit states (SLS). Fatigue loads have historically been validated by testing performed during type certification. ULS loads, on the other hand, cannot be easily verified with field measurements and are therefore left to numerical calculations. Experience with previous models and verifications against multiple approaches and codes will increase confidence in the results and reduce project risk. As the design spiral converges toward the final tower layout, the number of DLCs is progressively increased and more attention is paid to the component details. While the design of an isolated tower is relatively simple, the presence of a turbine generator and SbS still in the design loop, actually renders the process more complicated than one would initially anticipate. At the end of the design process, the CVA will need to approve of the design calculations and reports, which must demonstrate compliance with the codes, standards, and CVA’s internal review protocols. Through this thirdparty verification, the risks are further reduced. While the process described above is fairly complex, critical identification and understanding of the relative role of each loading source, and of the main structural and performance requirements, will help the engineer complete the design work successively. The following sections discuss principal loading sources and dynamic criteria that must be satisfied to initialize the tower design.

Table 10.3

Main DLCs and loading scenarios Sea conditions

Type of analysis

DLC

Design situation

Wind conditions

Waves

Currents

Water

Other conditions

1.1e1.6

Power production

NTMa, ETM, ECD, EWS

NSS, SSS, SWH

NCM

MSL, NWLR

(COD, UNI) & (COD, MUL) & MIS

ULS & FLS

2.1e2.4

Power production plus faults

NTM, EOG

NSS

NCM

MSL, NWLR

(COD, UNI)dcontrol system fault/grid loss

ULS & FLS

3.1e3.3

Start-up

NWP, EOG, EDC

NSS

NCM

MSL, NWLR

(COD, UNI)

ULS & FLS

4.1e4.2

Normal shut down

NWP, EOG

NSS

NCM

MSL, NWLR

(COD, UNI)

FLS & ULS

5.1

Emergency shut down

NTM

NSS

NCM

MSL

(COD, UNI)

ULS

6.1e6.4

Parked/idling

EWM, RWM, NTM

ESS, RWH, EWH, NSS

ECM

EWLR

(MIS, MUL) & (COD, MUL)dgrid loss/ extreme yaw misalignment

ULS & FLS

7.1e7.2

Parked/idling plus faults

EWM, RWM, NTM

ESS, RWH, EWH, NSS

ECM

NWLR

(MIS, MUL) & (COD, MUL)

ULS & FLS

8.1e8.3

Transport/ installation/ maintenance

EWM, RWM, NTM

ESS, RWH, EWH, NSS

ECM

NWLR

(COD, UNI) & (COD, MUL)dgrid loss

ULS & FLS

E1eE5

Power production

NTM

loads from temperature fluctuations, arch effect, moving ice floe, fast ice

NWLR

e

ULS & FLS

E6eE7

Parked/idling

EWM, NTM

Hummocked ice and ice ridges, moving ice flo

NWLR

e

ULS & FLS

a For an explanation of the acronyms see the glossary at the end of the chapter and Ref. [1]. From IEC 61400e3 Wind Turbines e Part 3: Design Requirements for Offshore Wind Turbines, 2009.


285

10.4.1 Sources of loading The tower, as previously mentioned, has the fundamental role of transferring loads to the SbS, or the foundation. From a civil engineering perspective, tower loads can be categorized into permanent actions (dead loads) and live loads. Dead loads are the gravitational loads associated with the self-weight of the structure, and the weights of the tower internals (see Section 10.6.1), as well as appurtenances connected to the deck (transformers, cranes, etc.) and monopile (eg, cathodic protection, platforms, boat landings, etc.). It has to be well understood that due to the vibrational and deflection characteristics of the OWT, even these so-called dead loads have very important dynamic effects, as for example the PeD effect associated with the displacement of the RNA center of mass (CM). Live loads include: • • •

• • • • • •

aerodynamic loads from the RNA, ie, forces and moments originating at the rotor and routed through the drive train and bedplate, drag loads from direct action of the wind on the tower, and potential vortex shedding loads, inertial loads associated with the vibrational modes of the system (eg, due to accelerations of the RNA mass) excited by the turbulent and sheared wind environment, by the inertial and aerostructural properties of the rotating rotor, and by the oscillations of the SbS (especially in cases of floating SbSs and seismic DLCs), loads derived from installation methods (hoisting, upending, etc.) and maintenance actions (including potential aircraft landing), hydrodynamic loads on the monopile portion: wave, current, and ice loads, seismic loads, loads derived from the actions on the SbS and reactions at the tower base, loads from impacts (eg, from boat landing operations, crane operations, aircraft landing), and loads from actuation of operation and control devices (yaw, pitch, brake, torque control mechanisms).

In Fig. 10.14, the approximate location of the application points of loads from RNA and hydrodynamics is shown together with the main coordinate system adopted by this chapter (as well as by the main standards of reference). Only rough approximations of the above loads can be hand-calculated. For instance, in order to calculate the direct wind action on the tower, thus the associated shear and bending moments, one may use basic aerodynamics principles. Through consultations of the standards, wind shear values (eg, Ref. [1]) and dynamic amplification factors (DAFs) (or gust factors from Refs [61e64]) may be obtained to calculate and integrate the drag force along the tower span. The determination of most of the other load components, however, demands a more rigorous CAE tool and coupled numerical simulations. Whereas only a detailed loads’ analysis can quantify the coupled effects of all load sources, a few general considerations can be made. The primary loading source for the tower proper comes from the aerodynamic loads induced by the rotor. Nonetheless, gravitational and inertial loads associated with the RNA mass should not be overlooked when calculating the stability and buckling resistance of the tower. This is especially true for ULS loads. In general, fatigue loading tends to dominate the design of the flanges and welds, while the tower shell may be driven by modal and

286

Offshore Wind Farms

Figure 10.14 Main tower reference system used in the standards [1,23] and this chapter. Also, principal sources of loading and their general areas of application are shown. Modified from an illustration by Joshua Bauer, NREL.

buckling-strength requirements. Moreover, depending on the water depth and the size of the OWT, the hydrodynamics may play a very important role in FLS. For shallow-water sites, tower DELs are dominated by aerodynamics. As water depths increase, hydrodynamic loads become progressively more important. Due to the aerodynamic damping behavior (discussed below), if wind and waves are misaligned or if the machine is idling (or parked), the DELs from wave action may increase substantially. Therefore, it is important to have a good understanding of the expected OWT availability and account for its positive or negative feedback on the FLS of the entire system.

10.4.1.1 Turbine loads (RNA loads) From a steady-state point of view, RNA loads reduce to three forces and three moments along the main coordinate axes; they are generated by the rotor aerodynamics under asymmetric conditions due to wind shear and rotor orientation (eg, non-zero yaw and tilt angles). Further periodic loads are generated by structural imbalance,


287

tower shadow effects, and turbulence sampling as discussed below in Section 10.4.1.2. In addition to the normal operational loads, transients, such as shutdowns, and blade/ rotor/yaw faults can give rise to important ULS loads for the tower and cannot be underestimated. Other important load contributions derive from the gyroscopic effects when the turbine yaws with rotor spinning. The thrust is the largest responsible for the bending moment distribution along the tower. If the turbine is a downwind turbine, the additional effect of the gravitational load due to a downwind offset of the RNA center of mass (CM) from the tower centerline may significantly increase the tower utilization. In an upwind turbine, the RNA mass contribution to the bending moment is minimal and any upwind offset of the RNA CM can be conservatively ignored, besides, the PeD effect would tend to reduce this effect anyway. From an FLS standpoint, the aerodynamic loads tend to dominate the design, with the exception of cases in deep-water sites where hydrodynamics excitation and low damping situations (see Section 10.4.1.2) can also be important. RNA loads can, at first approximation, be calculated pseudo-statically, considering a rigid tower, but DAFs should be used to start off the tower design. Thrust can increase by a factor of 1.5 or more due to inertial and gust effects. Short of high-fidelity computational fluid dynamics (CFD) simulations, rotor loads can be calculated via either beam element momentum theory (BEMT), generalized dynamic wake (GDW), or free wake vortex methods (FWVM) [65]. BEMT is fairly efficient and generally accurate, but tends to miss effects associated with the dynamic variations of the inflow and with the time-dependent evolution of the rotor wake. These effects are better captured by the GDW theory, which solves the linearized, inviscid equations of motion [66,67]. A higher-fidelity option, the FWVM methods, can track the vorticity in the rotor wake and the bound vorticity at the blade lifting lines (or surfaces depending on the level of fidelity). By using the Biot-Savart law, the flow field of interest, and the induced velocity at the rotor, can be resolved. Any of the solution methods will need to account for unsteady aerodynamics and dynamic-stall effects that can be present especially under skewed flow conditions (eg, with yaw errors > 0 degree). Among other features to model are tower flow-damming (or shadowing) effects, blade root and tip losses, and fully turbulent, 3D inflow-field, possibly accounting for wake effects within a park array. For a review of aerodynamics theory and requirements for turbine aeroelastic codes, see Ref. [67]. Dedicated turbine AHSE tools normally let the user choose the aerodynamics module options and the aerodynamic solver. They can account for the motion of the blades due to elastic and rigid motions of the hub, as well as yawed flow situations in fairly accurate fashion, returning good estimates of the loads.

10.4.1.2 1PenP forcing, resonance avoidance, and modal requirements From what is stated above, major loads for a turbine tower derive from the aerodynamics and inertial actions of the RNA. Rotor aerodynamics is complicated by turbulence and unsteady phenomena, which result in effects that cannot be modeled via a quasistatic approach. Additionally, turbulence and wind shear sampling by the rotating blades

288

Offshore Wind Farms

produce significant periodic excitations at a frequency equal to n times the rotor rotational frequency, or nP (as well as higher harmonics), where n is the number of rotor blades. Moreover, the tower itself affects the flow field (tower shadow or damming effect), which, when sampled by the rotor, further yields an nP forcing. Finally, rotor imbalances (both aerodynamics and structural, also quantified by the standards [3,23]) give rise to 1P excitations. For modern OWTs, the natural frequencies of the SSt and blades, and the main rotor excitation frequencies are in a comparable range. This leads to a potential coupling in the vibrational modes among the various components. At a crude level of approximation, an OWT can be regarded as a series of damped harmonic oscillators (see also Fig. 10.31 for the one-degree-of-freedom oscillator). It can be easily shown that the equation of motion for the generic j-th modal degree of freedom (DOF) can be written as: €xj þ 2xj u0; j x_j þ u20; j xj ¼

Fj mj

[10.1]

where for the j-th eigenmode, xj describes the DOF variable; mj is the generalized modal mass; u0,j is the eigenfrequency; Fj is the generalized dynamic forcing function; and xj is the damping ratio. Coincidence of structural eigenfrequencies with wind turbine dynamic forcing, known as resonance condition, can lead to large amplitude stresses and increased damage rates. Simplistically, this is illustrated by the DAF trend visible in Fig. 10.15, where the DAF is the ratio of the maximum amplitude of the dynamic response to the static response, ie, 1 DAF ¼ sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 2 u u 2 þ 2xj u0;jj 1 u0;jj

[10.2]

where uj is the forcing frequency associated with the j-th mode of vibration. For the above reason, the wind turbine rotor blades and SSt are designed to avoid resonance potential. In particular, the current practice is to design the wind turbine SSt such that the tower fundamental resonance frequency does not coincide with either the rotor (1P) or blade passing (nP) frequencies. Depending on the placement of the natural frequency with respect to the operational 1P and nP ranges, the SSt is defined as stiffestiff, softestiff, or softesoft (see Fig. 10.16). The stiffestiff option may lead to extensive use of structural steel, and it is generally avoided for pure cost reasons. The softesoft configuration may be very advantageous from an economical standpoint, but system frequencies may be dangerously close to the wave spectrum band with high-energy content. Thus far, the softestiff approach, where the support first bending eigenfrequency is placed between the 1P and nP ranges, has been the preferred choice. In any case, this approach has significant consequences for the structural design of OWTs, and may very well be the first driver for the total mass of towers, piles (as in the case of monopiles (MPs)), and SbSs.


289

Frequency response – dynamic amplification factor

20

DAF

15

ξ = 0.01

10

ξ = 0.05 ξ = 0.10

5

ξ = 0.20 0 0.0

0.5

ξ = 0.50 ξ = 1.00 1.0 1.5 2.0 Frequency ratio ω /ω 0

2.5

Figure 10.15 Typical response of a second-order mechanical system (damped harmonic oscillator) in terms of DAF as a function of the ratio of the forcing frequency to the system natural frequency, and for various x values (see text for more details).

Campbell diagram 2nd global fore-aft

1.2

6P

0.8 Stiff-stiff

1st global fore-aft

0.4 Soft-stiff

Rated

0.6 Cut-in

Frequency (Hz)

1.0 Blade 1st flapwise

3P

0.2 1P

Wave excitation

0.0 0

Soft-soft

2

4

6 RPM

8

10

12

Figure 10.16 Typical Campbell diagram for a 5-MW turbine on a monopile SSt.

290

Offshore Wind Farms

Obviously, for floating systems, an additional set of resonance conditions associated with the rigid body sea-keeping modes of the floating platform must be assessed. It is therefore not surprising that the first natural frequency of the SSt (hereafter, f0) plays a critical role within the overall system dynamics, and it is the first and foremost structural parameter to be assessed in tower design. During the front-end engineering design (FEED) (ie, when within the inner arms of the spiral of Fig. 10.13), f0 shall be accurately calculated via modal analysis in 3D finite element analysis (FEA) commercial software. In the preliminary phases of design, however, approximated expressions can be used, as for example Eq. [10.3]: 1 f0 x 2p

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3EJxx ð0:23mtwr þ mRNA ÞL3

[10.3]

Eq. [10.3], shown in Ref. [68], is strictly valid for an untapered and rigidly cantilevered tower, and shows dependency on: L, the tower length, or SSt length; mtwr and mRNA, the tower mass or SSt mass, and RNA mass; E, the Young’s modulus; Jxx, the cross-sectional area moment of inertia. Extensions of Eq. [10.3] have also been offered in the literature, such as Eq. [10.4] from Ref. [69]: 1 f0 x 2p

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3EJxx gK ðgM þ mRNA Þmtwr L3

[10.4]

where the stiffness (gK) and mass (gM) correction factors must be calculated as a function of the structure stiffness, applied load, and soil-pile (foundation) stiffness (see Ref. [69] for details). Energy-based methods, such as the RayleigheRitz method, can be effectively used to arrive at a good approximation of the first modeshape and eigenvalue of the SSt. Using a RayleigheRitz approach, and a trigonometric (or modal) representation of the tower node displacements (y(z, t)), f0 may also be written as: vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u Z L u 00 u EðzÞJxx ðzÞby ðzÞ2 dz 1 u 0 uZ f0 x 2p u t L rðzÞAðzÞby ðxÞ2 dz þ mRNA by ðLÞ2 0

[10.5]

with: yðz; tÞ ¼

X

fj ðtÞyj ðzÞ

j

by ðzÞ ¼

X yj ðzÞ j


291

where the integrals are written as functions of the span coordinate z, but can also be easily discretized for numerical quadrature. In Eq. [10.5], fj(t) is the j-th modal coefficient, or periodic function of time; r(z)A is the distributed mass; y is the lateral deflection of the tower along its span described by a linear combination of functions satisfying the boundary conditions (for instance known mode-shapes for a similarly constrained configuration); and the prime symbols denote derivatives with respect to z. By including the effect of the SSI stiffness (assuming krot, klat the equivalent rotational and lateral elastic constants respectively) the same method yields: vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi uR L 00 0 y ðzÞ2 dz þ klat by ð0Þ2 þ krot by ð0Þ2 1 u xx ðzÞb t 0 EðzÞJ f0 x RL 2p y ðxÞ2 dz þ mRNA by ðLÞ2 0 rðzÞAðzÞb

[10.6]

As can be seen from the above equations, the RNA mass and hub height are two main parameters affecting f0. While the RNA mass may be more difficult to change even in a fully integrated OWT design approach, the hub height can be manipulated to fine-tune the tower response, often more efficiently than changes in shell diameter and thickness. The problem of resonance avoidance is nonetheless complicated by the limited accuracy in the predicted forcing amplitudes and system resonant frequencies. On the one hand, rotor aerodynamic load amplitudes depend on stochastic inflow characteristics. On the other hand, the resonant frequencies are also a function of the soil and foundation physical properties, which can only be known with a certain level of uncertainty. Soilestructure interaction characteristics are also time-dependent (eg, due to scouring effects). Scouring and reduction in foundation integrity over time are especially problematic; by reducing f0, these effects may push the system resonance to the low frequencies, at which much of the broadband wave and gust energy is contained, or align it more closely with the 1P band. To an extreme, actual f0 can be up to 10e20% off the calculated ones because of miscalculated soil effects. It is crucial to account for this possibility and to verify the actual turbine performance in the commissioning phase. The design standards, for example, offer specific guidance toward resonance avoidance. Ref. [29] recommends avoiding operating in a frequency interval defined as the tower (SSt) eigenfrequency 10%, whereas Ref. [15] recommends a minimum distance of 5% from the fundamental system eigenfrequency. Adopting an extra 10% margin on top of what the standards suggest is a good strategy for the first steps in the design spiral. During commissioning, the control system and operating regimes could be altered to reach acceptable parameters in case the resulting modal performance is not as expected. Alternatively, costly retrofits to the tower configuration, including the addition of dampers, could be envisaged. An example of the response of a wind system with f0 inside and outside the 1P frequency band is given in Fig. 10.17, which shows the power spectral density (PSD) of the mudline overturning moment (OTM) for a typical 5-MW OWT. The graph is representative of the undesirable shift in f0 that may occur due to either unexpected soil

292

Offshore Wind Farms

1015

Spectral density, My (–)

No resonance With resonance

1P 6P

10

3P

10

105

0

0.2

0.4 0.6 Frequency (1/s)

0.8

1

Figure 10.17 Typical power spectral density response for the SSt FA bending moment at mudline during operational conditions for a 5-MW turbine. The two lines correspond to an SSt with f0 outside (without resonance) and inside the 1P frequency band. From T. Fischer, W. de Vries, Final Report Task 4.1-Deliverable d 4.1.5-(wp4: Offshore Foundations and Support Structures), Upwind Project 4.1, Universit at Stuttgart, Allmandring 5B, 70569 Suttgart, Germany, 2011 Contract No.:019945 (SES6).

conditions, degradation, or installation issues. It can be observed how the SSt response is dangerously amplified with direct consequences on fatigue loads. Besides the frequency content of the structural response, the other key factor in the resulting loads is the system damping. If sufficient damping can be guaranteed, say via adoption of tuned mass dampers (TMDs) or active mass dampers (AMDs), proximity to resonance conditions may no longer be a limiting factor. The damping is usually given for each principal (j-th) mode of vibration in terms of cj, or more often in terms of either dj or xj with respect to the cc,j damping ratio, see Eqs. [10.7] and [10.8]. 1 xj ¼ rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi

[10.7]

cj cc; j pffiffiffiffiffiffiffiffi ¼ 2 m j kj

[10.8]

1þ

2p dj

2

xj ¼ cc; j


293

Damping derives from, in order of contribution entity, aerodynamic effects, structural features (materials and additional dampers), soil, and hydrodynamics. Aerodynamic damping is primarily related to the aerodynamics of the rotor under a fore-aft (FA) oscillation of the tower top. As the tower head moves upwind, the relative wind speed increases, thus the blade airfoils’ angles of attack increase, generally leading to an increase in lift, drag, and resulting thrust force. The opposite occurs when the tower head moves downwind. Additionally, the drag on the tower also participates in a similar fashion to increase damping, but this effect is generally small, as the effective wind speed change is insignificant, and the main effect is through the mentioned lift/thrust mechanism with changes in the rotating blade angle of attacks (AOAs). Because of the foregoing, tower side-to-side oscillations have little or no damping, as nearly no aerodynamic damping exists. Furthermore, under idling or parked conditions, aerodynamic damping is minimal as well. This is particularly important for those loading scenarios where waves and wind present a substantial misalignment and wave forcing may be falling onto the natural frequencies of the system. In sites with large misalignments between wind and waves, the side-to-side bending moment can become a design driver in FLS. Structural damping is primarily due to internal friction in the materials of the SSt. Most of the energy dissipation occurs at the joints, and in particular at grouted connections. Soil damping mostly originates from ground deformation due to the action of the structure piling. Additionally, an adequate control system, for example, acting on generator torque and blade pitch, may further decrease vibrational amplitudes in towers and SSt. For floating turbines, tower/SSt damping along roll, pitch, and yaw may be an absolute necessity to achieve with dedicated collective and cyclic pitch of the rotor blades. Different control strategies, as for example via either the so-called “rotational speed window” (or frequency skipping) that acts on generator torque, or via tower feedback control and independent pitch control (IPC), acting on blade pitch, have been proposed (see for instance Ref. [70] for an excellent review on the subject). These strategies work well under operational conditions. As discussed above, for deeper waters where hydrodynamics may dominate the FLS loading state, idling cases with severe sea states can be design drivers. In those cases, standard torque and pitch controllers are ineffective. A “soft-cutout strategy” may be employed, however, which de facto extends the cutout wind speed and derates the turbine, thus reducing idling time spent in high wind/wave conditions. These strategies are very promising and may also bring forth larger energy capture, but need to be carefully devised. Alternatively, one could incorporate a TMD into the SSt design from the beginning (see also Section 10.6.1). Aeroelastic tools directly account for aero-hydrodynamics and TMD damping; soil and structural (internal) damping can be input by the user. It is usually good norm not to overestimate the damping. Recent studies have shown that for 3.6-MW wind turbines a logarithmic decrement of some 12% is achievable [71].

294

Offshore Wind Farms

10.4.1.3 Direct action from the wind Direct wind loading on the tower is due to aerodynamic drag on the structural shells. The wind profile can be taken as in Eq. [10.9], where the as is given by the standards (eg, Ref. [1]). jUðzÞj ¼ jUhub j

z

as

zhub

[10.9]

For a typical tubular tower, Eq. [10.10] can be used to calculate the pseudostatic loads on the tower due to wind action on its cylindrical or conical segments. f a ¼ 0:5ra pDsh Cd Gf UjUj

[10.10]

fa is the force per unit length due to wind aerodynamic drag, with ra the air density, Dsh the outer diameter (OD) of the tower, U the wind velocity, and Cd the drag coefficient (x 0.6e0.7); all are potentially functions of height MSL. By integration along the span of the tower, including the thrust of the RNA, one can attain the total shear and bending moment along the tower segments. Normally the thrust of the rotor under operational DLCs is the prime responsible for the bending stresses in the tower, and for FLS verification the direct wind drag is less important; yet, parked cases under ULS conditions can give rise to important drag loads that cannot be underestimated, especially for high hub heights and smaller rotors. The gust factor in Eq. [10.10] accounts for the effect on wind actions from the non-simultaneous occurrence of peak wind pressures on the structure surface together with the effect of the vibrations of the structure due to turbulence. Various methods to calculate Gf are offered by the standards (eg, Refs [61e64]), and more recent treatments with a focus on wind turbine towers can be found in the literature [65,72]. In Ref. [63], structures are defined as dynamically sensitive (or flexible) if their first natural frequency is f0 < 1 Hz (more detailed criteria can be found in Ref. [62], Parts 1e4). When this criterion is met, Gf can be significantly larger than 1. Besides ULS cases, turbulent wind actions can be important in the vortex-induced vibrations, which can be observed when the rotor is either not operating or altogether absent during installation and maintenance. The frequency of vortex shedding is tied to the Strouhal number and a critical wind speed can be calculated as: Ucr ¼

f0 Dsh St

[10.11]

where Dsh can be taken as the tower-top OD. St, the Strouhal number, varies around 0.15e0.20 depending on the tower taper ratio. It should be verified that critical wind speeds are within the operational regime, where the effect of the spinning rotor would tend to disrupt the mechanism of vortex formation behind the tower. Note that the cross-wind oscillations are insidious due to the reduced damping offered by the rotor, but again they are unlikely in normal operation conditions. During installation and/or


295

maintenance, the tower has higher natural frequencies and the probability of cross-wind resonance excitation at low and mid-wind speeds should be determined.

10.4.1.4 Hydrodynamic and ice loads in the case of a MP tower The effects of wave and current kinematics on the loading of a monopile tower can be calculated via the Morison equation for slender members: f w ¼ 0:25rw pD2sh Cm U_ w þ 0:5rw pDsh Cd Uw jUw j

[10.12]

where fw is the force per unit length due to wave and current kinematics; rw is the sea water density; Dsh is the tower OD; Uw is the wave and current velocity (vector sum); and Cd and Cm are the drag coefficient and the added mass coefficient, respectively. Following [1], two components of sea current velocity shall be taken into account: • •

wind-generated, near surface currents, and subsurface currents generated by tidal motion, storm surges, and atmospheric pressure variations.

The subsurface current velocity can follow a powerelaw profile as a function of water depth. For near-surface currents, it is common to assume a linear distribution of water speed with depth, with surface speed proportional to the hourly wind speed at 10 m MSL, and with effects vanishing below 20 m depths. Wave and current velocities should be summed vectorially to render Uw, which may be augmented to also contain the structure deflection velocity. The wave particle kinematic velocity can be calculated via Airy and Wheeler stretching theories [28,73]. The coefficients Cd and Cm should account for the presence of marine growth, and depend on Reynolds and KeuleganeCarpenter numbers, thus on the specific DLC under consideration (see also Ref. [28]). Corrections to the Morison equation have been proposed for diffraction effects (eg, the MacCamy-Fuchs correction [74]), and for wave non-linearities. Additionally, one should verify the probability of breaking wave occurrence (normally when Hw/dw > 0.78) and associated slamming loads on the SSts; Ref. [11] prescribes a simplified equation to account for these additional loads. Clearance (for example an air-gap of approximately 1.5 m) between the wave crest and the components unable to withstand these loads should be envisioned. Other concentrated loads may derive from the presence of secondary steel structures, such as boat landings, and J-tubes, which may attract more wave loads. These loads are normally of lesser entity, yet the welded connections to the principal steel can be a source of crack nucleation and corrosion. Efforts should be undertaken to verify that cracks are not able to propagate in the main load-bearing structure. Beside direct hydrodynamic loads, an MP tower may need to withstand sea-ice loading. In waters with winter climates, ice loading may be severe, and several situations are to be investigated as, for example [1]: • •

horizontal loads due to temperature fluctuations in fast-ice covers, horizontal and vertical loads from fast-ice cover subject to water-level fluctuations,

296

Offshore Wind Farms

Figure 10.18 Example of an ice-cone installed at the base of the towers at the Nysted wind farm, Denmark. • •

horizontal loads from moving ice floes, and pressure from hummocked ice and ice ridges.

A particularly important aspect to assess is the possibility of dynamic locking of the ice-breaking frequency to the wind turbine eigenfrequencies with consequences on both FLS and ULS. Specific CAE tools exist to investigate various types of ice loading configurations, and the standards provide good guidance (eg, Refs [75,11]), but Arctic engineering experience should be sought after during preliminary design in order to avoid reliability risks and expensive retrofits at later stages. A typical feature used with monopiles is the so-called ice-cone (see also Fig. 10.18), which induces bending stresses in the ice sheets that can then more easily break off, thereby reducing pressure on the SSt. Other aspects to consider are those associated with seabed movements, including sand waves, shoals, and scouring. These effects can result in consequences to soile structure interaction, removal or vertical and lateral support to the monopile and change of dynamic properties of the entire SSt. If these aspects are expected to be significant, the foundation design may need to consider them via embedment length modifications and possible scour protection.

10.4.1.5 Gravitational, inertial, and impact loads Inertial and gravitational loads are both static and dynamic loads resulting from vibration, rotation, gravity, and seismic accelerations. The self-weight of the OWT is simply


297

given by the mass schedule of the various components. For the design of the tower, the RNA mass must be known fairly accurately. Ref. [3] assumes gf x 1.1 on gravitational load (to compare to 1.35 for aerodynamic loads) to account for uncertainties in the as-built masses. For transportation and erection cases, gf x 1.25. Other inertial forces, when calculated separately (ie, not via AHSE coupled approaches), should also be factored (w1.2 or 1.35 for normal and extreme DLCs, respectively, following Ref. [3]). For FLS and SLS, unity load factors may be assumed. An earthquake analysis is normally not required for areas where the ground acceleration is less than 0.05 g. However, for certain sites, seismic activity may be combined with increased environmental loading (large wave amplitudes up to tsunami-type waves) and that should be investigated as an additional DLC, or it should be ensured that the basic load envelope from the standards’ DLCs encompasses these loading scenarios. Seismic loads may be assessed in accordance with Refs [3,23,63,64,76]. The analysis of the dynamic response to an earthquake can be performed via AHSE simulations with a given time history of the ground acceleration. Response spectrum analysis can also be used (see for example Refs [63,64,76]). The standards provide a way to combine seismic loads with other environmental and turbine state loads under normal conditions (eg, Refs [3,23]): the ground acceleration shall be evaluated with a 475-year return period, and the larger of either the average operational loads over the OWT lifetime, or the loads associated with an emergency shutdown should be superimposed to the seismic loads. For seismic loads, the PSF is taken as gf ¼ 1. Recent studies [71,77,78] demonstrate the importance of seismic loads for the larger hub heights and tower-top masses. Impacts from crane operations, aircraft landings, and accidental boat collisions should all be investigated. The most likely events among the above-mentioned ones are boat collisions. Dedicated boat-landing fenders are designed to plastically deform under impact, but other sections of the SSt shall withstand general impacts, without exceeding yield strength. Ref. [3] prescribes a calculation method for the horizontal load (Fsi) from boat impact: pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Fsi ¼ vsi csi asi msi

[10.13]

where msi is the displacement mass of impact vessel; vsi is the vessel impact speed; csi is the stiffness of the impacting part of the vessel; and asi is the added mass coefficient during collision (1.4e1.6 sideway impact, 1.1 bow or stern collision). Finally, loads from faults and emergency stop events can create dangerous impulses and shock load spectra on the entire OWT, and can culminate in very large deflections and loads. As discussed in Section 10.4, they need to be investigated following the prescriptions in the standards.

10.4.2 Aero-hydro-servo-elastic simulations Whereas analytical relationships and approximate equations that make use of pseudo static approaches can be used for the preliminary design of OWT towers, more extensive

298

Offshore Wind Farms

loads’ simulations are inevitable in their final design and optimization. The reason for this necessity is multifold. On the one hand, the complex structural dynamics of rotating machinery and stochastically varying environmental loads make simplified approaches too coarse. One might risk overconservatism (hence large costs) by accompanying larger safety factors to simple equations, or, more dangerously, might miss important dynamic coupling phenomena. As already stated, in fact, excitation frequencies and major component eigenfrequencies are in close proximity, thereby raising the possibility of dynamic coupling among vibrational modes of different parts of the OWT. Moreover, aeroelastic non-linearities and significant transient loads from start-ups, shutdowns, and fault events are difficult to model via pseudostatic approaches and their analytical treatment is normally out of reach. For all these reasons, the design of OWT towers and other components must be performed in integrated fashion, where fully coupled loads’ simulations can be undertaken. A large variety of CAE tools exist, with various levels of detail applied to the modeling of the different components. Rotor aerodynamics, hydrodynamics, control dynamics, soil and foundation dynamics, as well as structural dynamics can all be accounted for via a combination of multibody formulations, modal reduced systems, and finite element frameworks (an example of such a framework is given in Fig. 10.19). ElastoDyn Structural dynamics BeamDyn Nonlinear finite element blade dynamics AeroDyn Aerodynamics ServoDyn Control and electrical drive dynamics Fast 8 Driver

HydroDyn Hydrodynamics SubDyn Multimember substructure dynamics MAP Mooring statics and dynamics FEAMooring Mooring dynamics IceFloe Ice dynamics from DNV-GL IceDyn Ice dynamics from UMich

Figure 10.19 Example of CAE tool framework: FAST 8. Courtesy of NREL.

4 2 0 –2 –4 –6 500

510

520

530 Time (s)

540

550

560

5100

10000

5000

5000

Upwind pile force (kN)

× 10

520

530 Time (s)

540

550

0 560 × 105 2

4

3 2

1

1 0 500 4

× 10

510

520

530 Time (s)

540

550

0 560

× 105 2

4

3 1

2 1 0 500

OTM at mudline (kN m)

4

510

510

520

530 Time (s)

540

550

0 560

Downwind pile force (kN)

4900 500

OTM at tower base (kN m)

Generator power (kW)

299

Blade flapwise bending (kN m)

Wave elevation (m)


Figure 10.20 Example of time-series outputs from CAE tool loads’ analysis simulations of a 5-MW turbine on the OC4 jacket. From top to bottom, time series of: wave elevation; generated power and blade root flapwise bending moment; OTM (fore-aft) moment at the base of the tower and at the mudline; axial force magnitude for two pile heads, one located upwind and one downwind.

Example of time-series plots from fully coupled tools are shown in Figs. 10.20 and 10.21. Engineers should always question their models’ results, and it is important to have a knowledge of the efforts undertaken for the verification and validation of the various tools before accepting their outputs. Given the lack of maturity of the offshore wind industry, validation data are still in their infancy. Moreover, validation is challenging because of difficulties associated with quality control of the measured data at sea, difficulty in selecting more standard load cases from a real-life, long-term measurement campaign, and testing costs. Code-to-code comparisons offer a great way to substantiate the validity of new codes, when compared to more established and extensively validated tools. Examples of these efforts are the offshore code comparison collaboration (OC3) and OC4 projects [79e81] (see also Fig. 10.22). In general, it is important to have the CVA on board to vet and approve the use of the CAE tools of choice.

300

Offshore Wind Farms

Wave elevation (m)

1.5 1 0.5 0

–0.5 –1 350

360

370 Time (s)

380

390

6200

400

Generator power (kW)

15,000

6100

10,000

6000

5000

10

× 10

350

360

370 Time (s)

380

390

4

4

5

0 340

0 400

3

350

360

370 Time (s)

380

390

Platform pitch (degree)

OTM at tower base (kN m)

5900 340

Blade flapwise bending (kN m)

–1.5 340

2 400

Figure 10.21 Example of loads simulation output for a 6-MW turbine mounted on a floating spar. From top to bottom, time series of: wave elevation; generated power and blade root flapwise bending moment; OTM (fore-aft) moment at the base of the tower and platform heel angle.

The output of AHSE loads’ simulations is then post-processed to verify ULS and FLS limit states, where the structural integrity is checked following guidance from codes and standards.

10.5

Shell and flange sizing

In the case of a steel tubular tower, the main output of the design exercise is the sizing of the shell segments and of the welded and bolted connections. In the following two sections, an overview of the limit state verifications and of the structural criteria to be satisfied for these components is provided. The physical principles are easily extended to other tower configurations, with the addition of specific verifications. For lattice towers, the size of the individual members and the qualification of the joints are the principal design outcome. They can be sized by performing member and joint checks


301

Figure 10.22 Examples of code-to-code verification efforts for the development of FAST v8: (a) comparison of time-series of jacket OTM at the mudline among various codes; (b) the data in terms of PSD.

as specified by appropriate standards (eg, Refs [19,31]). In the case of concrete towers, the size and arrangement of the reinforcement, of the prestressing tendons, and the characteristics of the concrete (eg, density and compression strength) are additional quantities that need to be determined following specific standard guidance (eg, Refs [22,34,36]).

10.5.1 Load resistance factored design: FLS, ULS, and SLS verifications From what has been stated thus far, it should not be surprising that the tower must satisfy a large number of structural and modal performance criteria; Ref. [1], for example, prescribes both operational, parked, and fault-loading situations under a number of turbulent and steady wind models and sea state conditions. These DLCs should be translated into stress and strain fields that must be verified against ULS, FLS, and SLS limit states within the LRFD approach. In general, different segments of the tower may be driven by different limit states: for example, the upper segments may be mostly driven by FLS requirements, whereas the bottom portion may be mostly ULS-driven. These factors may change depending on the site conditions, water depths, SbS configuration, and turbine sizes. Accounting for additional effects, such as corrosion and secondary steel (see Section 10.6.1), makes the verification phase all the more complicated.

302

Offshore Wind Farms

10.5.1.1 Safety factors The generic ULS limit state verification may be expressed as follows (cf. Ref. [23]): gn SðFd Þ Rðfd Þ

[10.14a]

or explicitly gf Fk

fk gn gm

[10.14b]

where S(Fd) is the probability distribution of the generic, design (factored) load within the LRFD approach; R(fd) is the analogous function for the material factored resistance; Fd (Fk) is the factored (unfactored) characteristic load; fd (fk) is the material factored (unfactored) resistance; gn is the consequence of failure PSF (or “importance” factor); gf is the generic load PSF; and gm is the material PSF. Because many DLCs involve response to turbulent inflow under a range of mean wind speeds, and wave spectrum forcing, the exceedance probability for the characteristic load must be computed based on the expected wind and wave distributions. Guidance on the determination of the characteristic loads through the extrapolation of the results of limited-duration AHSE simulations is given in the standards (cf. Refs [15,23]). The standards also provide recommended values for gf (see also Table 10.4) and gn [1,11,15,21,23]. The load factors represent the uncertainty in the load stochastic distribution and in the load assessment. In general, design standards allow for the adoption of lower than prescribed load PSFs in those cases where the load magnitudes are established by measurements, yielding a high degree of confidence. The important factors are based on the redundancy and fail-safe characteristics of the various components. Following the classification in Ref. [23], towers are considered components of class 2 (see Table 10.5), and gn ¼ 1 may be assumed. Material PSFs can be either taken from specific, recognized design codes (eg, Refs [43,82e84]), or minimum values may be taken per the main design standards such as Ref. [23] (see Table 10.6). For SLS verifications, which include critical deflection analyses such as blade-totower clearance, ULS load factors may be used in combination with PSFs for consequence of failure and material resistance as indicated by the appropriate standards, or as shown in Tables 10.5 and 10.6. Table 10.4

Examples of ULS, SLS, and FLS load PSFs Unfavorable loads

Limit state

Favorable loads

a

Normal abnormal transport and installation

All DLCs

ULS/SLS

1.35

1.1

1.5

0.9

FLS

1.0

1.0

1.0

1.0

a Normal or abnormal attributes for the various DLCs are given by the standards (eg, Ref. [1]). From IEC 61400e1. Wind Turbines e Part 1: Design Requirements, 2005.


Table 10.5

303

Examples of minimum gn as a function of component class gn

Component class

ULS

FLS

SLS

Comment

1

0.9

1.0

1.0

“Fail-safe” structural components whose failure does not result in the failure of a major part of a wind turbine (eg, replaceable bearings)

2

1.0

1.15

1.0

“Non fail-safe” structural components whose failures may lead to the failure of a major part of a wind turbine

3

1.3

1.3

1.3

“Non fail-safe” mechanical components that link actuators and brakes to main structural components for the purpose of implementing non-redundant turbine protection functions

From IEC 61400e1. Wind Turbines e Part 1: Design Requirements, 2005.

Table 10.6

Examples of minimum gm as a function of failure mode gm

Failure mode

ULS

FLS

SLS

Yielding of ductile materials Global buckling of curved shells Rupture from exceeding tensile or compression strength

1.1 1.2

1.1 (welded and structural steel) to 1.7 (composites)

1.0 (if elastic properties proven by full-scale testing); 1.0 (otherwise)

1.3

From IEC 61400e1. Wind Turbines e Part 1: Design Requirements, 2005.

The generic FLS verification may be expressed as follows (cf. Ref. [23]): Dfat ¼

X ni 1 b i Ni

[10.15]

with b i ¼ Ni sa;i gf gn gm N

[10.16]

where Dfat is the fatigue damage; ni is the number of cycles at the i-th load range; and b i is the “factored” Ni, ie, the number of cycles at failure corresponding to the N PSF-augmented i-th load range, (gf gngmsa,i). Ni values can be calculated from sN curves for the material and structural element under consideration as described in Section 10.5.2. The FLS PSFs are given by the same design and certification standards, as for example shown in Tables 10.4e10.6

304

Offshore Wind Farms

Finally, per Ref. [18], one should design the OWT and its components with the overall exposure category L2 (see Section 10.3.1), which also requires that a 500-year robustness check be carried out with unity PSFs.

10.5.2

Approximate derivation of structural loads and shell design

Analytically, one could account for the various contributions to the tower loads as in Eqs. [10.17]e[10.20]. The normal (axial, along the z axis) force can be written as: ZL Nd ðzÞ ¼ gf FzRNA gfg mRNA g gfg

rAgdz

[10.17]

z

where Nd is the design (factored) normal load at the tower station of interest; FzRNA is the aerodynamic force from the RNA along the z axis; gf is the generic load PSF; mRNA is the RNA mass; g is the gravity acceleration; gfg is the gravitational load PSF; r is the material density; L is the tower length, or SSt length; A is the cross-sectional area; and z is the coordinate along the tower span. The shear components along x and y (reference system is that of Fig. 10.14) can be taken as: ZL Tx ðzÞ ¼ gf FxRNA þ gfa

f a $bidz

[10.18a]

f a $bjdz

[10.18b]

z

ZL Ty ðzÞ ¼ gf FyRNA þ gfa z

where FxRNA is the force from the RNA along the x axis; FyRNA is the force from the RNA along the y-axis; fa is given in Eq. [10.10]; gfa is the aerodynamic load PSF; bi is the unit vector along the x axis; bj is the unit vector along the y axis. The bending moment components can be written as:

Z L b Mx ðzÞ ¼ MxRNA FyRNA ðzRNA zÞ f a $ jz þ rAgðyðzÞ yðzÞÞ dz z

[10.19a] ZL My ðzÞ ¼ MyRNA þ FxRNA ðzRNA zÞ þ

f a $biz þ rAgðxðzÞ xðzÞÞ dz

z

[10.19b]


305

where Mx is the component of the bending moment along the x axis at the station of interest; My is the component of the bending moment load along the y axis at the station of interest; MxRNA is the RNA aerodynamic moment along the x axis; MyRNA is the RNA aerodynamic moment along the y-axis; z is the dummy coordinate along the z axis. Finally the torque about the z axis is given by: Mz ðzÞ ¼ MzRNA FxRNA ðyðzRNA Þ yðzÞÞ þ FyRNA ðxðzRNA Þ xðzÞÞ

[10.20]

where Mz is the torsion moment load along the z axis at the station of interest. Note that the second order PD effect was accounted for in Eqs. [10.19a]e[10.20], and that potentially different PSFs for the various loading components were employed following standards such as Ref. [3]. The RNA forces must be calculated following what was stated in Section 10.4.1.1. The shear and moment components can then be combined to arrive at characteristic design values: Td ðzÞ ¼

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Tx ðzÞ2 þ Ty ðzÞ2

[10.21a]

Md ðzÞ ¼

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Mx ðzÞ2 þ My ðzÞ2

[10.21b]

Finally, the normal stresses (sz,Ed along the meridional, and sq,Ed circumferential directions), the shear stress (szq,Ed), and the Von-Mises equivalent stress (svm) for the generic cross-section of a tubular tower can conservatively be written as: sz;Ed ¼

Nd Md Dsh þ A 2Jxx

sq;Ed ¼ gf ðkw 1Þqmax szq;Ed ¼ 2Td =A þ svm ¼

[10.22a] Dsh ts 2ts

Mz 2Amid ts

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi s2z;Ed þ s2q;Ed sz;Ed sq;Ed þ 3s2zq;Ed

[10.22b] [10.22c] [10.22d]

where the argument z was dropped without losing generality; kw is the dynamic pressure factor to calculate hoop stressesdit is a function of cylinder dimensions and external pressure buckling factor per Ref. [2]; ts is the shell thickness; Amid is the area inscribed by the mid-thickness line; and all the other symbols were above introduced except for qmax, which is the qmax expressed as in Eq. [10.23]: qmax ¼ 0:5 ra jUj2

[10.23]

306

Offshore Wind Farms

with ra being the air density, and U the wind velocity, which may also include structural motion components (normally negligible). While the above treatment is sufficient for conceptual and preliminary assessments, only rigorous loads and FEA analyses can fully support a more detailed design. In particular, fully coupled loads’ analyses via AHSE simulations are the only way to capture important interactions and vibrational dynamics and to perform FLS and ULS verifications for the certification of OWT SSts. The FLS verification must determine the accumulated damage over the SSt design lifetime (usually of 20 years), accounting for the appropriate operational and non-operational DLCs. From what was stated earlier, the focus should be on winde wave misalignment and idling DLCs where the vibrational damping is negligible. Based on the PalmgreneMiner rule [85,86], the total damage requirement can be expressed as: Dfat ¼

X ni Ni

i

1

[10.24]

where Dfat is the fatigue damage; ni is the number of cycles at the i-th load range; and Ni is the number of cycles at failure for the i-th load range level. Ni values can be calculated from seN curves for the material and element under consideration (eg, weld butt, steel sheet, flange, bolts; see also Fig. 10.23). The typical seN curve is derived from material coupon testing and it can, at first approximation, be expressed by: 1=m

sa;i ¼ CNi

[10.25]

where sa,i is the i-th bin stress range; C is the constant in the seN curve, approximately equal to the material ultimate strength; and m is the inverse exponent in the seN curve. Material standard specification provide values for m; for steels m ranges between 3 and 5. It is common practice to factor sa,i in Eq. [10.25] for the load, material resistance, b i ) as shown and consequence of failure PSFs to arrive at a “factored” Ni value ( N in Eq. [10.16]. Hence, Eq. [10.24] can be replaced by Eq. [10.15] and Eq. [10.25] rewrites as: 1=m

sa;i

bi CN ¼ gf gn gm

[10.25a]

Note that in Eq. [10.25a], a load factor gf potentially different from 1 is included. This means that the calculated fatigue loads that will complement the analysis are considered unfactored. To account for the presence of non-zero mean loads (hence stresses), Eq. [10.26] (known as linear Goodman’s correction) can be used: seq;i ¼ sa;i

su su sm;i

[10.26]


307

Direct stress range Δσ R (N/mm2)

1000

1

100

1 m=3

160 140 125 112 100 90 80 71 63 56 50 45 40 36

2

3

1 Detail category ΔσC m=5 10 1.0E+04

1.0E+05

2 5 1.0E+06 1.0E+07 1.0E+08 Endurance, number of cycles N

2 Constant amplitude fatigue limit ΔσD 3 Cut-off limit Δσ L 1.0E+09

Figure 10.23 seN curves for various detail categories. From European Committee for Standardisation, Eurocode 3: Design of Steel StructuresdPart 1e9: Fatigue, 2005.

where seq,i is the i-th bin equivalent stress range after Goodman’s correction, and that can replace sa,i in Eq. [10.25a]; su is the ultimate strength; and sm,i is the i-th bin stress-range mean. In order to compute each ni, the loads output by the AHSE simulations for the various DLCs are to be binned and combined together via their probability of occurrence, which is based on the expected site joint-probability of wind and wave distributions and best estimates of system availability. In the case of an MP tower, additional damage may be due to pile-driving actions and should be added to the fatigue budget reckoning. Through load binning and a rainflow cycle-counting method [23,87], a representation of load-ranges (Sa,i) versus number of cycles (ni) can be achieved. However, the direct use of Eq. [10.24] (or better Eq. [10.15]) would require the transformation of each load range into stress ranges (sa,i). Material non-linearities may require multiple FEA runs to achieve this transformation. It may be more efficient to employ damage equivalent loads (DELs). For a generic load component, the DEL represents the zero-mean load range that, if applied for a select number of cycles, would yield the same damage as that produced by the application of the actual load cycles. In other words, the DEL is a measure of the damage accumulated in a specific structural part due to the oscillating loads that can be rainflow-counted from the FLS DLC outputs and combined as described above for the lifetime of the OWT. Therefore, by equating the actual and the

308

Offshore Wind Farms

equivalent damages as expressed in Eq. [10.24], and making use of Eq. [10.25] where the stress symbols are used to indicate loads, it can be shown that: 0

11 m n DEL A @ DEL ¼ P ni m i s a;i

[10.27]

where nDEL is the reference number of cycles for the DEL load range; and sa,i is the i-th bin load range. The DELs can finally be transformed into damage equivalent stresses (DESs) via detailed FEAs, which render a 3D description of the stress field, including stress concentrations and hot-spots. The obtained DES (or, better yet, the seN curve that it will refer to) should be modified by the same knockdown factors on the load, material resistance due to quality of construction, importance of the detail, and material characterization as shown in Eq. [10.25a]. Additional PSFs may be included based on expected fatigue corrosion rates (see Section 10.6.1). As discussed in Section 10.5.1.1, load, material resistance, and consequence of failure PSFs are recommended by the primary OWT standards (eg, Refs [23,1]), while further knockdown factors for specific elements (such as welds and flanges) are directly incorporated in the seN curves (eg, detail category curves as those in Fig. 10.23) given by the appropriate codes of reference (eg, Refs [30,39,44]; see also Section 10.3). The DES, together with Eq. [10.25a], yields the number of cycles at failure for the DES stress range (NDEL), and Dfat can be calculated as: Dfat ¼

X ni nDEL ¼ b NDEL i Ni

[10.28]

Eq. [10.28] may be used to readily calculate the FLS material utilization at various stations along the tower and in the welds, in the door reinforcements, and in the bolted connections. For the ULS verification, data obtained from AHSE simulations are further post-processed to arrive at ultimate loads. For operational DLCs, this process may require extrapolation of stochastically based data as directed by the standards [1,23]. The range of DLCs to investigate includes fault transients, extreme gusts, wind directional changes, as well as emergency stops, and possible shocks from ship impacts and sea ice as discussed in Section 10.4.1.5. ULS structural checks amount to ensuring that the material utilization be below 1, and are based on the verification of steel members under compression-bending. In summary, one should prove that, for the various tower segments, the stresses are kept below the allowable yield strength and that stability at a global and local level is guaranteed. These constraints can be expressed by the following equations [3,2]: svm

fy gm gn

[10.29]


309

Nd b Md þ m þ Dn 1 kNp Mp

[10.30]

sz;Ed sz;Rd

[10.31a]

sq;Ed sq;Rd

[10.31b]

szq;Ed szq;Rd

[10.31c]

sz;Ed sz;Rd

kz

þ

sq;Ed sq;Rd

kq

sz;Ed sq;Ed szq;Ed ks 1 ki þ sz;Rd sq;Rd szq;Rd

[10.31d]

Eq. [10.29] states the constraint on the yield resistance of the material (fy). Eq. [10.30] is the global (Eulerian) buckling constraint; where Nd and Md are the design axial and bending moment loads; Np and Mp are the associated characteristic, buckling critical, resistance values; k and bm are reduction factors for flexural buckling and bending moment coefficient, respectively [3]; Dn is a function of the slenderness (l) of the tower as in Eq. [10.32]: 2

Dn ¼ 0:25kl

[10.32]

Eq. [10.31] is the local (shell) buckling constraint; sz,Ed, sq,Ed, szq,Ed are the design values of the axial, hoop, and shear stress, respectively; sz,Rd, sq,Rd, szq,Rd are the corresponding characteristic buckling strengths; kz, kq, ks, and ki are constants given by the standards [2]. In the calculation of the resistance values in Eqs. [10.30]e[10.31], the standards of reference offer guidance on the selection of manufacturing quality factors. In the case of MP towers, the stability of the pile at the seabed must be verified. This is done for the largest (ULS) pile-head loads, where head displacement and rotation are checked against allowable values from the standards. Additionally, the embedment length must develop sufficient vertical friction to counteract the maximum normal force. These checks are commonly performed by the foundation engineering group. That group is also responsible for ensuring sufficient stiffness in the SSI superelement to provide the necessary boundary condition to the SSt. In addition to FLS and ULS checks, an important structural check for both blade and tower design is the verification of potential blade strike on the tower, which may be considered part of SLS. Blade maximum deflection may occur under a transient event as in the case of extreme gusts or fault situations. In order to verify that a safety margin remains in the deflection of the blade before tower strike, the maximum deflection across all ULS DLCs must be determined. Standards such as Refs [23,15] offer guidance on the PSFs to employ in this calculation: Ref. [23] uses the same load PSF as for any other ULS DLC, whereas Ref. [15] states that the blade clearance shall not be less than 30% of the value under unloaded (at rest) conditions.

310

Offshore Wind Farms

For other types of towers, such as lattice and concrete towers, additional structural checks are necessary. For lattices, each member and each joint must be assessed under all DLCs (eg, against structural constraints described in Ref. [19]), which is a critical and time-consuming activity. For concrete towers, for all ULS DLCs, one should verify: concrete’s compressive strength; reinforcement’s bending strength; punching, anchorage, and pull-out strength of the reinforcement; and the shear strength of structural members with and without shear reinforcement. For SLS and FLS, analyses should assess deformation limits, crack width limits, and stresses on the prestressed members. Possible guideline standards are listed in Table 10.1, but it is to be noted that the high load cycle number fatigue for prestressed concrete structures is still a focus of significant research.

10.5.3

Flanges and main detail components

Flanges and other detail components make up an important portion of the structural performance and functionality of the tower. Bolted connections and weldments are among the most important details as they assure the structural integrity and safety of the structure. Just as important are the details of door and man-hole reinforcements, and of the brackets for supporting the internals and secondary steel. Secondary steel is discussed in Section 10.6. In general, cracks in welded steel structures almost certainly depart from welds. The reason is that the welding process inevitably leaves behind microscopic defects from which cracks may grow. As a result, one should assume the existence of microcracks, and that most of the fatigue cycles will contribute to the growth of the crack rather than to its nucleation. Additionally, welds are characterized by changes in surface slope, as at the toes of butt welds and at the toes and roots of fillet welds. These areas are characterized by important stress concentration factors (SCFs). The door reinforcement is also a location of high SCF. In order to restore the structural strength of the tower at the door opening, either a welded reinforcement ring (stiffener), or an increased shell wall thickness, or a combination of the two is commonly utilized. The reinforcement has the double goal of controlling the local stresses, and of supporting the main shell against buckling. Structural codes (eg, Refs [30,39]) describe fatigue curves that can be used for general design and are based on test results of actual characteristic weld shapes (often referred to as detail categories). However, especially for the door opening area, actual stresses and hot-spots near welds need to be assessed via accurate FEA analyses (eg, Fig. 10.24). Ref. [2] provides guidance on extrapolating “hot-spot” stresses, and discusses the complex geometrically and materially non-linear analyses with imperfection modes (GMNIA) process to assess buckling resistance. Weakening of a structural weld can significantly reduce the overall buckling strength. Therefore fatigue design of the connections between shells, and between flanges and shells, is critical to the lifetime integrity of the OWT, and indirectly even on its ULS capacity. Additionally, the effect of a corrosive environment can accelerate crack propagation and effectively increase fatigue damage (see Section 10.6.1) around the connection


(a)

311

(b)

Plate stress:VM (MPa) 355.0000 301.7500 266.2500 230.7500 195.2500 159.7500 124.2500 88.7500 53.2500 17.7500 0.0000

Figure 10.24 Examples of FEA of the tower door opening: (a) a mesh of the lower tower segment; (b) results of the analysis verifying shell buckling around the door region. From M. Veljkovic, C. Heistermann, W. Husson, M. Limam, M. Feldmann, J. Naumes, D. Pak, T. Faber, M. Klose, K. Fruhner, L. Krutschinna, C. Baniotopoulos, I. Lavasas, A. Pontes, E. Ribeiro, M. Hadden, R. Sousa, L. da Silva, C. Rebelo, R. Simoes, J. Henriques, R. Matos, J. Nuutinen, H. Kinnunen, High-Strength Tower in Steel for Wind Turbines (Histwin). Final Report EUR 25127 EN. Directorate-General for Research and Innovation e European Commission, Luxembourg, Europe, 2012, Contract No RFSR-CT2006e00031.

details. Corrosion-fatigue of bolts may contribute to stiffness degradation with the previously mentioned problems of dynamic response in undesirable frequency bands (see Section 10.4.1.2). Tubular tower flanges connect tower segments together and the whole tower to the TP, or foundation in the case of onshore installations. They make up a significant contribution to the material and manufacturing costs. Flanges and bolts need to be checked against FLS and ULS limit states. A common way to analyze the bolted flanges follows the well-known Petersen method [89], or “segment-model.” In that model, the flange connection is replaced by an individual circular segment, where a single bolt is considered with the associated collaborating flange and shell regions. The new unit connection is loaded with the equivalent normal force (Fz) deriving from the shell normal stresses as shown in Fig. 10.25 and Eq. [10.33]. Fz ¼

4Md Nd nb Dsh;m nb

[10.33]

where Md and Nd are the factored bending moment and the factored normal load at the flange elevation; nb is the number of bolts in the flange connection; and Dsh,m is the shell mid-wall diameter. Because Nd acts in a favorable direction, it is usual practice to assign gfg ¼ 0.9 to the gravitational loads that make up Nd. Eq. [10.33] is easily derived considering trigonometry rules, the symmetry in the connection, and a simple moment balance about a horizontal axis.

312

Offshore Wind Farms

Stresses in shell

Fz

Segment Ring flange

Fz

Figure 10.25 Reduction of the flange bolted connection to an equivalent, unitebolt connection (flange segment model). Modified from P. Schaumann, M. Seidel, Failure analysis of bolted steel flanges, in: X. Zhao, R. Grzebieta (Eds.), Proceedings of the 7th International Symposium on-Structural Failure and Plasticity (IMPLAST 2000), Melbourne, Australia, 2000.

The connection capacity is calculated through the plastic hinge theory considering the three failure methods as shown in Fig. 10.26 (ie, bolt rupture, bolt rupture and plastic hinge in shell, plastic hinges in shell and flange), and as described by Eqs. [10.34]e[10.36]: Fu;A ¼ Ft;RD ts a

b

[10.34a]

Failure mode A: Bolt rupture

Failure mode B: Bolt rupture and plastic hinge in shell Fu MPl,3

Fu tf

R

MPl,3 a

b

Z < Ft,R

R

R∙a

M = Ft,R∙b b

MPl,3

MPl,2

Ft,R

Ft,R

a

Failure mode C: Plastic hinge in shell and flange Fu

MPl,2 a

MPl,3 b

Figure 10.26 Approximation of main failure modes for flange connections. Adapted from P. Schaumann, M. Seidel, Failure analysis of bolted steel flanges, in: X. Zhao, R. Grzebieta (Eds.), Proceedings of the 7th International Symposium on- Structural Failure and Plasticity (IMPLAST 2000), Melbourne, Australia, 2000.


313

Fu;B ¼

Ft;RD a þ MPl;3;MN aþb

[10.34b]

Fu;C ¼

MPl;2 þ MPl;3;MN b

[10.34c]

with Ft;RD

fy;b Ab 0:9fu;b Ab ¼ min ; gMb;y gMb;u

! [10.35]

where a and b are distances from the bolt centerline to the flange edge and the shell mid-wall; Ft,RD is the bolt strength load; MPl,3,MN and MPl,2 are the shell and flange plastic bending resistances, respectively; Fu,A, Fu,B, and Fu,C are the ultimate loads for the three different failure modes. Note that MPl,3,MN accounts for tension bending interaction. The MPl,2 and MPl,3,MN plastic hinge moments are given by: MPl;2 ¼

cf tf2 fy;f 4 gMf;y

[10.36a]

MPl;3 ¼

cs ts2 fy;s 4 gMs;y

[10.36b]

"

MPl;3;MN

Fult ¼ 1 NPl;3

2 # MPl;3

[10.36c]

where MPl,3 is the simple plastic bending resistance, without tension bending interaction; tf is the flange thickness; fy,f is the flange characteristic yield strength; gMf,y is the material PSF for the flange characteristic yield strength; ts is the shell thickness; fy,s is the shell characteristic yield strength; gMs,y is the material PSF for the shell characteristic yield strength; and NPl,3, the plastic resistance of the shell, is given by: NPl;3 ¼

cs ts fy;s gMs;y

[10.37]

In Eqs. [10.36a]e[10.37], cs and cf are defined by: cs ¼

pDsh;m nb

[10.38a]

314

Offshore Wind Farms

cf ¼

pDbc db nb

[10.38b]

with Dbc the bolt circle diameter in the flange connection; and db the bolt hole diameter. From Eq. [10.36c], it can be seen how MPl,3,MN introduces a non-linearity that requires some iterations to get to the final connection resistance Fult, which is the minimum of the expressions obtained through Eq. [10.34]. The connection is thus verified if: Fz Fult ¼ min Fu;A ; Fu;B ; Fu;C

[10.39]

Bolt stresses vary non-linearly as a function of shell loads due to the preload. Petersen’s method assumes a simplified bolt-load behavior as shown in Fig. 10.27(a), and as described by the following equations: Ft1 ¼ Fp þ Ka lf Fz Ft2 ¼ lf Fz

for Fz < Fzcr

[10.40a]

for Fz Fzcr

[10.40b]

where lf ¼

aþb a

Fzcr ¼

[10.40c]

Fp lf Kb

[10.40d]

(a)

(b)

Ft

Ft

Ft1

Ft2

Ft1

Ft2

b

Ft3

Fz

Ft

ts

tf Fp

tf

Fp

ts

a Ft Fzcr

Fz

Fz1

Fz2

Fz Fz

Figure 10.27 Bolt load (Ft) as a function of shell load (Fz): (a) experiments (dashed curve) and approximated trend according to Petersen [89]; (b) experiments (dashed curve) and approximated trend according to Schmidt/Neuper’s method [93] with the three linear regions.


315

where Fp is the bolt preload, and Ka and Kb are the fractions of the total stiffness for bolt and flanges, respectively: Ka ¼

Kb Kb þ Kf

[10.41a]

Kb ¼

Kf Kb þ Kf

[10.41b]

The bolt axial stiffness (Kb) can be written as: 0 Kb ¼ @

Z

Lb

0

11 1 Eb Ab dzA x Eb Ab 2tf

[10.42]

where Lb, ie, the bolt effective length, can be approximated by 2tf. The equivalent stiffness of the compressed flange pair (Kf) can be written as: 0 Kf ¼ @2

Z

tf 0

11 1 dzA Ef Af

[10.43]

where Af is the effective cross-sectional area for the compressed flanges. Kf is sometimes taken as three to six times the Kb, or approximated as [91]: Ef db2 þ 1:36db tf þ 0:26tf2 Ef Af Kf ¼ x 2tf 2tf

[10.44]

Other standards and literature, such as Refs [40,92], can also be consulted to determine the Af and thus Kf. Petersen’s approach is considered adequate for ULS assessments. Compared to experiments (see Fig. 10.27(a)), the method is conservative below Fzcr, but, because of the assumed ideal-elastic and pure edge-bearing behavior, could become unsafe in the case of ring-flanges with imperfections at and above Fzcr. For more accurate FLS verifications, Schmidt/Neuper’s method [93] may be used. That method uses a three-region, piecewise-linear function to approximate the bolt load as shown in Fig. 10.27(b) and as described by Eq. [10.45]. Ft1 ¼ Fp þ Ka Fz

for Fz Fz1

Fz Fz1 Ft2 ¼ Fp þ Ka Fz1 þ lf Fz2 ðFp þ Ka Fz1 Þ Fz2 Fz1

[10.45a] for Fz1 < Fz Fz2 [10.45b]

316

Offshore Wind Farms

Ft3 ¼ lj Fz

for Fz Fz2

[10.45c]

where Fz1 ¼ Fp

a 0:5b aþb

[10.45d]

Fz2 ¼ Fp

1 lf Kb

[10.45e]

0:7a þ b 0:7a

[10.45f]

lf ¼

As seen in Fig. 10.27, Schmidt/Neuper’s method is less conservative than Petersen’s for low shell forces Fz. The third part of the function, which considers the final edge-bearing behavior of the flange connection (Fz Fz2), is however devised to account for the effects of small imperfections. The range of applicability of Schmidt/ Neuper’s method is given by: aþb 3 tf

[10.46]

Neither of these methods considers bending stresses in the bolt; therefore caution is advised in the selection of the fatigue detail curve and a final FEA is recommended. For a state of the art in flange design consult Ref. [94] and later work by the same author.

10.6

Secondary steel, other structure details, and coatings

While the structural engineer is normally concerned with the main structure, secondary steel and auxiliary systems cannot be overlooked. On the one hand, access and safety systems guarantee the possibility of efficient and safe maintenance of the OWT. On the other hand, optimizing tower internals will allow for an optimized and slender primary steel distribution in the tower main shell. To this end, the adoption of damping systems may lead to more efficient overall design, but the dampers must be carefully analyzed and sized with a view to potential effects onto the remainder of the OWT both in terms of performance and costs. Additionally, attention must be paid to the protection of the materials from the environment corrosive actions. Failure to do so may bring forth unexpected consequences, such as reduction of the shell wall thickness and changes in the dynamic characteristics of the SSt, or even more serious consequences associated with fatigue corrosion.


317

10.6.1 Secondary steel In the design of a tower, man-hole and door reinforcements, access staircase, internal ladder, man and cargo hoists, lifts, platforms, electrical conduits, light panels and cable trays constitute the so-called secondary steel or internals (see Figs. 10.28 and 10.29). In some cases, tuned mass dampers can also be present, and offshore, other appurtenances near the deck or along a monopile, such as boat landing, helipads, crane pedestals, anodes, and J-tubes may also be considered as secondary steel. The access hatches, manlifts, and interfaces to the SbS and to the nacelle must comply with workmanship safety regulations and must offer enough versatility to allow for maintenance tools to be moved safely and effectively. Engineers should especially strive to find adequate solutions for guaranteeing safe and quick access to the nacelle. Although larger offshore wind turbines may make use of helipads on the nacelle, providing for a quick transportation to the top of the RNA is crucial to reduce off-line periods and revenue losses. Solutions that allow reducing the welds inside the tower are often sought to thin the shell and save on material quantities. Some original equipment manufacturers (OEMs) have adopted the use of magnets to attach platforms and other internals to the walls of the tower. Others have successfully employed either guy-wires secured to flanges, or adhesives to the tower inner walls. Both strategies reduce welded fittings and achieve savings in labor and material costs. TMDs (also known as vibration absorbers or dampers, see Fig. 10.30) can be used to reduce low- and high-frequency vibrations in towers and OWTs due to aeroelastic

Figure 10.28 Examples of tower internals and auxiliary platform.

318

Offshore Wind Farms

Figure 10.29 Examples of tower internals and secondary steel: (a), (b) ladder, cable tray, rest platform, and climb-assist system; (c) a man-lift.

Figure 10.30 Examples of wind turbine TMDs. From ESM Energie- und Schwingungstechnik Mitsch GmbH, http://www.esm-gmbh.de.


319

(a)

(b) y(t)

KS

KD

θ (t)

Dynamic amplification factor

(c)

LD AD A1D

F(t)

CD

CS

y(t)

KS

F(t)

ρ D, ν D

CS

MD

No damping Medium damping High damping 20

10

1 0.8

1 Frequency ratio

1.2

Figure 10.31 Conceptual diagrams of an oscillator and expected DAF. The basic structure stiffness and damping constants are denoted by KS and CS. (a) Physical model of a TMD, with the damper stiffness, damping, and mass denoted by KD, CD, and MD respectively. (b) Model of a TLCD, where the symbols denote: characteristic length (LD), cross-section (AD), choke section (A1D), and fluid densities and viscosity (rD and vD). (c) Potential effect of increased damping in terms of DAF.

and inertial forcing. The function of a damper is based on a spring/mass system, which counteracts and reduces the structural response (see Fig. 10.31). Once the performance targets are determined (eg, deflections, loads and accelerations, resonance frequency), dampers can be designed based on first principles and vibrational engineering. In certain cases, more complex solutions with viscous dampers, force-amplifying truss systems, and actively controlled actuators are needed to achieve the identified targets. These solutions are more difficult to properly design for, because they require renewed attention to detail and FEAs analyses coupled to system aeroelastic simulations. There are fundamentally two kinds of dampers: TMDs and active mass dampers (AMDs). Multiple tuned mass dampers (MTMDs) are of the first kind, but with multiple application points within the same structure. The main parameters, ie, the mass of the damper and its displacement, are a function of the mass of the system being damped. Therefore, for large OWTs, TMDs may

320

(a)

Offshore Wind Farms

(b) u

D

u u

L1

ψ H

L1

θ

Figure 10.32 Example of scissor-brace (a) and toggle-brace (b) MTMD. From A. Tsouroukdissian, C. Carcangiu, I. Pineda, T. Fischer, B. Kuhnle, M. Scheu, M. Martin, Wind turbine tower load reduction using passive and semi-active dampers, in: EWEA 2011, 2011.

require sizable devices and strokes to be effective, which can be an impediment to the installation inside towers. Typical damper mass entities are 1e10% of the system modal mass. Realistically 2e3% is all that can be achieved inside the tower. The TMD geometry, however, can be devised to amplify the primary structural drift and to achieve higher damping coefficients, for example through the use of trusses, toggles, and scissor-jacks as shown in Fig. 10.32 [95e98]. Besides truss systems, other devices have recently received attention for wind tower configurations following their success in the civil and offshore engineering industries. Among those are the tuned liquid column damper (TLCD) [96,99e102] (see also Fig. 10.31(b)) and other kinds of semipassive devices, such as the ball vibration absorber tested in Refs [103,104]. These systems can overcome some of the logistical limitations of the simpler TMDs. AMDs can be fairly effective [105,106] and bypass the same TMD limitations, but have an inherent power requirement and generally are more expensive. Partially active devices [96] retain advantages of the passive TMDs, but can also guarantee extra action on demand, as for example under sudden wind gusts. The theoretical effect of TMDs (as shown in Fig. 10.31(c)) is the complete reduction of the first eigenfrequency resonance peak. The original undamped mode is split into two modes with equal damping ratios. While the TMD is normally tuned to the f0 of the system, any misalignment between damper and SSt eigenfrequencies (for example due to soil conditions) can actually lead to increased loads. It might be more efficient to tune the TMD to the frequency spectrum of the exciting loads. Another advantage of the TMD over the other dynamic control strategies is the fact that it is effective during both operational and parked/idling cases. For floating turbines, the role of the oscillation damper is primarily on the surge and pitch DOFs, but for fixed-bottom OWTs, the primary role could be on the sway


321

or sideeside (SS) DOF, which has very little aerodynamic damping. Increasing damping has the potential of reducing deflections and loads, and of increasing the lifetime of the entire OWT. Direct consequences of increased damping on the tower translate to a reduced amount of required steel. Additionally, the reduction in tower deflection might have positive benefits on the blade-to-tower clearance, allowing for a lighter rotor. These effects may combine and allow for a lighter SbS as well. The ramifications of these kinds of design choice can further expand in the realm of the BOS costs, as manufacturing and transportation costs may all be affected. It is then clear how attention to the design of certain components, even within the secondary steel group, can actually have an important repercussion on the entire system performance and LCOE.

10.6.2 Designing for corrosion In addition to the loads discussed in Section 10.4, offshore steel SSts must also sustain the corrosive action of sea water, sea spray, wetedry cycles, temperature changes, and biological fouling. Corrosion affects component fatigue by reducing overall life at any given stress-level, by removing the fatigue limit for steels (see Fig. 10.33), and by inducing multicrack growth in contrast to a single fatigue crack of a clean environment. Corrosion pits may spur cracks earlier than what is predicted under non-corrosive conditions, and crack propagation is faster. For a good overview of the corrosion fatigue issue, the reader can refer to textbooks on material mechanics [107e110]. Examples of corrosion damage are seen in Fig. 10.34. These aspects are particularly important for the joints and welds of the SbS, but also for the tower flange connections, 80

Stress, MN/m

2

70

60

400

(b) 1000

50

300

Stress, ksi

Dry air Air, 93% R.H. Deaerated 3% NaCl Aerated 3% NaCl

500

Log failure stress

(a)

K 125 (no corrosivity)

100

K 125 (with corrosion protection) C (sea water with cathodic protection)

10

K 125 (sea water without corrosion protection) (sea water without corrosion protection)

C

40 1 30

200 10

5

6

10 Cycles of failure

10

7

4

10

10

6

8

10

Log loading cycles

Figure 10.33 Effect of corrosive environment on steel fatigue strength. (a) (From H.H. Lee, H.H. Uhlig, Corrosion fatigue of type 4140 high strength steel, Metall. Trans. 3 (11) (1972) 2949e2957.) deleterious effect of aerated aqueous chloride solution on the high fatigue cycle life of AISI4140 steel; note how the fatigue limit at high cycles is removed by the corrosive environment. (b) (From A. Momber, Corrosion and corrosion protection of support structures for offshore wind energy devices (owea), Mater. Corros. 62 (5) (2011) 391e404. http://dx.doi. org/10.1002/maco.201005691.) fatigue curves for an element of an offshore wind turbine tower in a clean environment, in a corrosive environment with corrosion protection, and in a corrosive environment without protection.

322

Offshore Wind Farms

Figure 10.34 Examples of extensive corrosion damage: on an O&G jacket (a) and inside a wind turbine monopile (b) (note the disintegrated ladder wrung). From R. Sheppard, F.J. Puskar, MMS TA&R Project 627 Inspection Methodologies for Offshore Wind Turbine Facilities Final Report. Energo Report Energo Project No.: E08147. Minerals Management Service (MMS), Houston, TX, January 2009.

Figure 10.35 Corrosion at the tower flange connection with the transition piece.

see Fig. 10.35. Therefore, the tower designer should be aware of the expected corrosive load on the structure right from the start; a corrosion protection system should be devised as integral to the structural design; and a balance should be found between costs of protection and repair costs. The O&G industry has developed extensive experience with corrosion protection, and special coatings have been developed to protect offshore structures. In contrast to O&G platforms, however, turbine towers are unmanned structures without permanent inspection of corrosion protection systems (CPSs). Therefore, while O&G platforms are promptly repaired upon damage detection, towers cannot be repaired without incurring large costs [112]. M€ uhlberg [114] states that repair of CPSs at sea can


323

cost up to 50 times more than the initial application during tower manufacturing. It is of note that repairs might be accompanied by expensive downtimes. Ref. [115] offers guidance on the inspection of offshore wind turbine structures, and particular attention is given to the CPS. The physics and chemistry of corrosion are quite complicated; whereas many references are available on the subject, it is good practice to request the help of material and corrosion experts during the design phase. Corrosion may be described as a process system consisting of three subsystems: medium, material, and interphase. For offshore applications, the medium can be either air, water, a combination of the two as in sea spray, condensation water in internal spaces, and soil on foundations. In all cases, a high salt content should be expected. The materials include unalloyed, low-alloyed and stainless steels, cast iron, aluminum and copper alloys, and composites. For the scope of this chapter, tower sections are considered to be made from steel. The interphase consists of chemical compounds that form from the oxidation of the materials within the medium. Depending on the environment and the loading level, various types of corrosion can be identified, for example, uniform corrosion, galvanic corrosion, pit corrosion, crevice corrosion, microbiologically influenced corrosion (MIC), stress (fatigue) corrosion, and erosion corrosion. All of these forms of corrosion can occur on offshore SSts. In general, the presence of moisture in the medium (conductive environment) spurs an oxidation reaction. The metal loses electrons in the medium, but metal oxide is not formed, as opposed to dry oxidation, and the formation of the reaction product might not occur at the reaction site. This is the basic mechanism of uniform corrosion. If another metal is in electrical contact with the primary one and located in the same environment, further (galvanic) corrosion will occur, as the metal at lower electrochemical potential will corrode (anode), protecting the other (cathode) as an electronic/ionic current is established. This explains why a zinc-coated (galvanized) steel tower performs efficient protection in the atmosphere where the moisture film is thin. Also, galvanic corrosion can occur at joints, where fasteners with different electrochemical potential are in contact with the tower main steel. Because of the presence of the medium, cathodic and anodic reactions might happen on the surface of the same metal, as water and oxygen interact with metal ions to form more complex ions and change the thermodynamics of the reaction. This is what makes the splash zone so active from a corrosion point of view. The presence of salt accelerates the reaction and allows for the formation of metal chlorides (the interphase) that undergo hydrolysis and thus lower the pH of the medium solution. The galvanic corrosion rate is proportional to the current intensity, and this is inversely proportional to the area of the electrodes. Pits and crevices represent small anodic areas, where a high salt concentration and more extensive presence of moisture are to be expected. For these reasons, pit and crevice corrosion are examples of fast-rate corrosion processes. MIC is a consequence of biological fouling, which can initiate and accelerate corrosion due to the interaction between microbial activity and steel. In general, underneath a biological layer that attaches to the SSt (eg, seaweeds and mussels), an anaerobic environment develops. Within the anaerobic environment, organisms like sulfate-reducing

324

Offshore Wind Farms

Figure 10.36 Cracks in a secondary structure due to corrosion fatigue. From P. Hogg, Durability of wind turbine materials in offshore environments, in: SUPERGEN Wind Phase 2e4th Training Seminar, SUPERGEN Wind, Manchester, UK, 2012.

bacteria (SRB) enhance the development of MIC. Metal reacts with hydrogen sulfides forming metal sulfides and hydrogen as corrosion products. Hydrogen tends to permeate the metal matrix making it more brittle and prone to cracking. This type of corrosion can also occur near the seabed. Stress, or fatigue corrosion, occurs when any of the other types of corrosion is accompanied by cyclic loading. Corrosion-fatigue fracture surfaces may or may not be coated with corrosion product depending on the relative effects of corrosion and stress (see Fig. 10.36). More evidence of corrosion can be expected at lower stress levels or lower frequencies of stress cycling, because of the increased time of exposure to an aggressive environment. Finally, erosion corrosion is associated with the abrasive action due to boat impacts, ice floe impact, surface icing, and biological fouling. Standards for corrosion protection (eg, Ref. [25]) subdivide the media and the environments in terms of corrosivity categories and exposure classes, as seen in Table 10.7. For offshore installations, category C5-M applies, and corrosion rates can be determined based on recommendations in Refs [1,3,11,25]. Further attention must be paid to the climatic region of the installation, as warmer climates, as for example in tropical and subtropical regions, give rise to higher corrosion rates, as do environments with higher salinity. In Arctic conditions, ice scoring can also increase corrosion rates through the removal of coatings and oxidation layers, but boat impacts can be even more damaging (see Fig. 10.37).


325

Table 10.7 Environment categories and expected thickness loss after one year exposure Thickness loss (mm)

Category

Low carbon steel

Zinc

Environment (examples)

C1dvery low

1.3

0.1

Dry indoor spaces, clean atmosphere.

C2dlow

25

0.7

Indoor spaces with occasional condensation. Mostly rural inland outdoor atmosphere.

C3dmedium

50

2.1

Indoor spaces with high humidity and low pollution. Urban and industrial outdoor air with low salinity.

C4dhigh

80

4.2

Indoor spaces: Chemical plants, swimming pools, boatyards. Industrial outdoor areas with moderate salinity.

C5-I/Mdvery high

200

8.4

Indoor spaces with permanent condensation and high pollution. Very humid and chemically aggressive outdoor atmosphere, coastal and offshore areas with high salinity.

From ISO 12944-2, 1998 Paints and Varnishes e Corrosion Protection of Steel Structures by Protective Paint Systems e Part 2: Classification of Environments, 1998.

Figure 10.37 Examples of coating damage and corrosion due to boat impacts on SbSs: (a) (from P. Hogg, Durability of wind turbine materials in offshore environments, in: SUPERGEN Wind Phase 2e4th Training Seminar, SUPERGEN Wind, Manchester, UK, 2012.); (b) (from M. de Jong, Adaptations to a Marine Climate, Salt and Water OWEZr11120101020 e Results Corrosionins Inpections Offshore Wind Farm Egmond aan € Zee, 2007e2009. Tech. Rep. 50863231 e TUS/NRI10-2242, KEMA Nederland B.V., Arnhem, The Netherlands, October 2010.).

326

Offshore Wind Farms

5-NZ

4–AZ

3–SZ

2–IZ z

1–UZ

Figure 10.38 Conventional corrosion zones for offshore wind turbines.

Conventionally, five corrosion zones are identified for wind turbines (see also Fig. 10.38): underwater corrosion zone (UZ); intermediate corrosion zone (IZ); splash corrosion zone (SZ); atmospheric corrosion zone (AZ) (C5-M); and internal or nacelle corrosion zone (NZ) (see also Refs [1,11]). For each of the zones, an appropriate design corrosion rate can be assigned, and an adequate CPS could be devised. CPSs include: design choices (eg, through the choice of structural materials and approaches to drainage), coating application, electrochemical protection, and monitoring and regular inspection [115,118,119]. The AZ, above the SZ, is exposed to uniform corrosion and must be protected by coatings as specified by standards such as Ref. [25]. For external surfaces, such as the tower outer surface, a zinc-based primer is normally covered by additional epoxy and polyurethane coatings. Coatings need to be inspected and repaired at prescribed intervals. For internal surfaces that can be exposed to external air, such as nacelle and tower inner surfaces, corrosion protection through coatings is commonly employed. Corrosion allowance can be granted in place of CPSs for components of minor significance in the AZ [3] and where inspections and repairs are possible. Corrosion-resistant materials for fastening devices (eg, stainless steel) and grating (GFRP) are acceptable as well.


327

Corrosion rates are highest in the SZ, where wave loading is also maximum, thus particular attention to the CPS must be paid in this zone. The SZ is bound by the 1-year crest height at highest still-water level (HSWL) and 1-year trough at lowest still-water level (HSWL). Here, maintenance of the CPS is not very effective, and neither is cathodic protection (CP). Coatings are mandatory and should be made of materials of proven reliability and following the usual standards [24e26], but they cannot be the only defense against corrosion. Therefore, corrosion allowance (CA) should be used for components such as towers, SbSs, and transition pieces if in the SZ. CA can be calculated as [11]: _ L Tc Þ CA ¼ CðT

[10.47]

where C_ is the corrosion rate, TL is the expected lifetime of the component, and Tc is the expected useful lifetime of the coating. Note that TL should account for storage time after production, installation time, and effective operational use of the OWT. C_ is assumed to vary between 0.15 and 0.20 mm/year for internal surfaces, and 0.30 and 0.40 mm/year for external surfaces [11]. Below the MSL, CP should also be used. In the IZ and UZ, CP is necessary, and coatings are optional and used to reduce the required CP capacity. Internal surfaces should be protected by either CP or CA. Ref. [11] recommends using a corrosion rate of C_ 0.10 mm/year. CP makes use of aluminum- or zinc-based, sacrificial anodes fastened to the main steel between the seabed and the MSL. The CP system must be designed for a minimum lifetime equal to the OWT’s, and guidance is offered by various standards (eg, Ref. [3]) on how to size anodes based on I, a protective current density (A/m2), and Q, a practical current capacity (Ah/kg) based on anode manufacturer specifications: ma ¼

As IðTL Tc Þ j Q

[10.48]

where ma is the anode mass; As is the total surface area to be protected; and j is a protective current density (A/m2) usually set to 1.1. Impressed current cathodic protection (ICCP), where the current is produced by a dedicated power source, can be used in lieu of the galvanic CP, and in principle it could be provided by tapping the grid connection. Dedicated cables, rectifiers, control, and monitoring devices are also needed. The extra costs associated with these devices should be weighed against the advantages of a more controlled cathodic polarization of the structure through ICCP, and the generally fewer anode elements that can also be located off the main structure. One must recognize that because of the presence in the structure of hatches and openings, perfect water- or air-tightness is almost impossible to achieve, even in zones that are initially thought of as such, as for instance the foundation cavity near the mudline. Moreover, air-sealed zones in towers and SbSs can still give rise to corrosion due to infiltration of seawater and the presence of SRB, and due to contamination of external air any time maintenance access needs to be granted. If a seal fails, tidal

328

Offshore Wind Farms

variations can occur within the structure cavities. This situation then resembles conditions near ports and harbors where accelerated corrosion occurs due to alternating wet and dry cycles with semistagnant conditions.

10.7

Optimization considerations

The tower is a relatively low-technology component that can also be modularized depending on the application. Offshore tower and SbS make up a significant portion of the OWT’s total installed cost; therefore they lend themselves to be one of the best candidates for component optimization. The taper of a typical offshore tower, for example, favors an optimal distribution of the wall thickness as the diameter increases, while still retaining a required stiffness level. On land, the maximum outer diameter (OD) at the base of the tower is constrained by transportation considerations. Offshore, stiffness and buckling strength requirements may demand even larger diameters than onshore, but the transportation constraints can be sufficiently relaxed. On the other hand, as discussed in this chapter, offshore SSt must satisfy even more structural requirements than their onshore counterparts, and their structural optimization (SO) is non-trivial. The discipline of SO has matured mostly within the aerospace industry, where advances in computational resources have opened up new ways to designing innovative solutions leveraging the progress made in CAE and FEA tools. In the wind industry, component optimization and, especially, system automatic optimization are still in their infancy. Several challenges exist, such as the non-linear dynamics of WTGs, the complex and stochastic loading scenarios, the importance of vibrational response and fatigue, and the consequent need for very specialized simulation tools. As far as SO within fatigue FLS-loading scenarios, not much literature support is available; in contrast, the majority of the work has been done within static (or ULS) scenarios, which is more readily achieved, requiring limited knowledge of the system response. In addition, the tight coupling among various components (see Section 10.7.1) demands multidisciplinary design optimization, which, to be accurate, should also be performed in the time domain. Hence, the analysis is quite complex and time-consuming. Frequency domain approaches, however, promise [120] to drastically reduce analysis times. An excellent review of methods and challenges of SO for OWTs is given by Ref. [121]. To try to overcome these challenges, new efforts are underway by turbine OEMs and by research laboratories [52,122e126]. In general, SO may encompass three different sets of problem [127]: 1. sizing optimization: eg, the cross-sectional area of the tower segment must be found 2. shape optimization: eg, when the profile of the tower (the taper in a tubular case) is sought as design variable 3. topology optimization: eg, when multiple members are allowed to participate and be removed from a truss, or lattice tower.

The basic constraint within the SO is that the response of the structure should be acceptable as per the various specifications, ie, it should at least be a feasible design


329

per the applicable standards (eg, those of Section 10.3). Since there can be many feasible designs, it is desirable to choose the optimal one in terms of either minimum cost, minimum weight, maximum performance. or a combination of these. For example, in the case of a tubular tower for OWTs, the SO reduces to a sizing optimization with discrete variables (eg, wall thickness and outer diameter of the tower segments) within both ULS and FLS loading scenarios. The ultimate objective should be that of minimizing the system LCOE. At the base of SO lies the mathematical formulation of a structural problem, which allows for finding optimal solutions via computer algorithms [128]. The mathematical nature of the problem is that of a non-linear programming problem, where approximations and successive linearizations are utilized to arrive at an iterative solution of the problem. The general SO can be written in mathematical terms as: ( SO

minxd fobj ðxd ; uðxd ÞÞ with gcnt ðxd ; uðxd ÞÞ 0

[10.49]

where xd is the array of design variables; u(xd) is a general (state) function of design variables (eg, as, for instance, the modal coordinates of the reduced model, or more simply the displacements of the elastic axis of a beam); fobj is the objective function; and gcnt represents the constraint function(s). The functions are of the array form, ie, they can have multiple dimensions in output as well as input. Different methods have been developed to solve Eq. [10.49] depending on whether the problem can be defined as either linear, convex, non-linear, nonconvex, or with discrete vs. continuous parameters [127,129e132]. Derivatives and Jacobians of the functions fobj and gcnt with respect to the array xd are usually needed to solve Eq. [10.49] (via the so-called gradient based optimization (GBO)). If analytical versions of the derivatives can be attained, the optimization will proceed much faster than the alternative, where numerical derivatives are employed. However, rarely one can confirm the convexity of the structural problem functions; therefore a GBO solution to Eq. [10.49] may not necessarily yield the global optimum, and multiple iterations may be necessary to determine the best solution. To overcome this drawback, genetic algorithms (GAs) can be employed, which are methods suitable for complex optimization problems with either continuous or discrete variables. GAs are robust at locating the global optimum, but at the cost of a large number of function evaluations; therefore they become advantageous when the number of variables is large. As an example of optimization, a 10-MW turbine tower shell was optimized for mass via a GBO algorithm. The tubular tower was envisioned as mounted on top of a jacket, at a base elevation of approximately 21 m MSL. The steel density was augmented to account for flanges, hardware, secondary steel, and coatings. For sake of simplicity, only two ULS DLCs were considered, and only the tower-base and tower-top cross-sectional properties, and the length (Htwr2) of a constant cross-section segment (tower waist) were selected as design variables. A more refined model could include all of the shell-segment elevations and associated Dsh and ts as optimization variables, as well as more load cases

330

Offshore Wind Farms

and constraints. Here, this example aims to demonstrate the power of an optimization algorithm during preliminary design, and thus first-order approximations are acceptable. Further verifications would demand higher fidelity, and AHSE could then be used to assess fatigue damage to be included as extra gcnt functions in the SO problem. The main turbine and environmental parameters for this SO example are given in Table 10.8, together with the imposed bounds for manufacturability (eg, minimum and maximum Db, minimum and maximum DTRs for weldability). The main structural checks performed, besides targeting modal performance in terms of desired f0 (see Table 10.8), are those described by Eqs. [10.29]e[10.31]. Results of the optimization process can be seen in Fig. 10.39 and Table 10.9. Note that, for this example, DLC 1.6 (a maximum thrust, operational condition) tends to drive the design of the tower shell (see Fig. 10.39(b)). This is in agreement with what was mentioned earlier regarding the overall importance of the RNA loads with the exception of very deep-water sites where FLS may be driven by hydrodynamics. In order to further optimize the shell, one could include more shell stations as design variables, and this effect will be shown later. As visible from the results in Table 10.9, the tower mass is substantially reduced from an initial guess at the layout. Moreover, the target eigenfrequency is better captured, and the ULS maximum local-buckling utilization reaches unity. This is a good starting point for the more detailed design, and it was achieved in a matter of a few minutes on a personal computer. A verification of the results was also conducted through the ANSYS® FEA package (FEA) (see Fig. 10.39(c)e(d)). The design space could be further investigated for solution improvement, or to assess mass penalties associated with changes in design variables, as for instance due to manufacturing or transportation constraints. In order to address these sensitivities, a sweep over some 4480 cases was carried out by varying the design variables in their allowed ranges, and some of the results are shown in Fig. 10.40. The graph shows contours of tower mass, eigenfrequencies, and structural utilization (ie, global buckling utilization ratio (GLUtil) and shell buckling utilization ratio (EUUtil) that tend to drive the design). Because of the discretization of the various design variables in the sweep analysis, the graph represents only an approximate cross-section of the problem hyperspace in proximity of the attained optimal solution (ie, Dt ¼ 3.5 m, rather than 3.65 m; DTRt ¼ 170, rather than 156; and Htwr2 ¼ 23.95 m, rather than 24 m). Yet, it can be seen how the obtained layout (denoted by the green cross) is indeed a global optimum, and that moving away from that would imply an increase in mass. For example, assuming a hard limit of 7.5 m on Db, a feasible solution (one that satisfies the constraints) would likely result in a penalty of 100 t (the red cross in the graph of Fig. 10.40), at least as long as the Htwr2 parameter is not significantly changed. Similar analyses to that in Fig. 10.40 could very well help in and speed up design decisions. Other graphs such as that of Fig. 10.41 can be used to further explore the sensitivity of mass and other parameters to the design variables and objectives. In Fig. 10.41(a), the general trend in terms of tower mass and GLUtil is shown together with obtained eigenfrequencies and average DTRs along the tower span. It can be observed how a stiffer tower (higher f0) demands larger steel mass as expected, but also how the


331

Table 10.8 Main load and environmental parameters for the simple optimization study of an offshore tower Parameter

Value

Unit

Comments

Mass

677

t

CMzoff

2.5

m

Hub height

119

m

f0

0.25

Hz

Target frequency

Uref

33

m/s

Hub height reference wind speed in simple DLC loads’ analysis

Max thrust

3.411Eþ06

N

RNA ULS unfactored thrust force

MxRNA, MyRNA

9.95Eþ06, 1.322Eþ07

Nm

RNA ULS moments associated with max thrust load

Uref

70

m/s


Max thrust

2.1Eþ06

N

RNA ULS unfactored thrust force

MxRNA, MyRNA

0.0, 1.572Eþ06

Nm

RNA ULS moments associated with max thrust load

Cd

0.7

e

Tower aerodyanmic drag coefficient assumed constant for simplicity

r/fy

8740/345

kg/m3, MPa

Steel density and yield strength; density augmented to account for secondary steel, coatings, hardware, etc.

Kx, Kqx

5.8e7, 4.4e10

N/m, N m/rad

Equivalent lateral and rotational spring constants at tower base

Kz, Kqz

1.3E9, 6.46e9

N/m, N m/rad

Equivalent vertical and torsional spring constants at tower base

Tower flange interface

20.7

m MSL

RNA data RNA CM vertical offset from tower-top flange

DLC 1.6

DLC 6.1

Boundary data

Continued

332

Offshore Wind Farms

Table 10.8

Continued

Parameter

Value

Unit

Comments

Design variable bounds Db [min,max]

[4,7.5]

m

Tower-base OD allowed range

Dt [min,max]

[3,4]

m

Tower-top OD allowed range

DTRb [min,max]

[120,200]

e

Tower-base DTR allowed range

DTRt [min,max]

[120,200]

e

Tower-top DTR allowed range

Htwr2 [min,max]

[0,18]

m

Tower-constant cross-section top elevation from tower base

(a)

40

20

0 0.0

0.5

(c) 1Nodal solution Step = 1 Sub = 1 Frequy = 0.249894 (avg) RSYS = 0 DMX = 1 SMX = 1

1.0 Utilization ratios

1.5

60

40

20

0 0.0

2.0

(d) Plot no. 1

0 0.222222 0.888889 0.444444 0.666667 0.111111 1 0.777778 0.333333 0.555556 10MW_offshoretower

Vonmises util1 Vonmises util1 EUsh util1 EUsh util1 GL util1 GL util2

80

Z from base of tower (m)

60

Utilization ratio

100 Vonmises util1 Vonmises util1 EUsh util1 EUsh util1 GL util1 GL util2

80 Z from base of tower (m)

(b)

Utilization ratio

100

1 Element solution Step = 1 Sub = 22 Time = 1 SEQV (noavg) DMX = 2.52662 SMN = 0.153E+08 SMX = 0.253E+09

0.5


1.5

2.0

Plot no. 1

0.682E+08 0.174E+09 0.227E+09 0.153E+08 0.121E+09 0.417E+08 0.946E+08 0.148E+09 0.200E+09 0.253E+09 10MW_offshoretower

Figure 10.39 Tower profile (in the middle of the graphs) and utilization distributions for the DLC 1.6 (denoted by suffix 1) and 6.1 (denoted by suffix 2): (a) shows utilizations for the initial guess configuration, as (b) does for the optimized design. ANSYS® results showing the first eigenmode and the ULS stress distribution are given in panels (c) and (d).

wall thickness can be reduced (DTR increased) to save mass for a given target frequency. The two trends show how the DTR-versus-mass surface is less steep than the f0-versus-mass one, illustrting how the wall thickness is not as effective at reducing mass as tower-base diameter, for example, as is shown in the analogous graph in


333

Initial guesses and optimized configuration for minimum mass [t] of the tower simple optimization study

Table 10.9

Design variable

Initial guess

Optimized value

Multisegment-optimized

Unit

Db

6.9

7.99

8.42 (8.15)a

m

3.57

b

3.91 (4.89)

m

a

e

b

e

4

Dt

100

DTRb

120

DTRt

a

183

212 (309)

156

236 (185) c

c

Htwr2

15

24

48 , 79

m

Mass

855

624

552

t

f0

0.24

0.25

0.256

Hz

Diameter and DTR at first station above tower base. Diameter and DTR at second station above tower base. Elevation of intermediate stations above tower base.

b c

Mass & f1; Dt = 3.5 m, DTRt =170, Htwr2 = 23.95 m

1400

200 1200

160

1000

140

800

Mass (t)

Tower base DTR

180

120 600 100 7.0

7.5 f0 (Hz)

8.0

Tower base OD (m) GLutil

8.5

400

EUutil

Figure 10.40 A cross-section of the design hyperspace: mass-filled contour as a function of tower-base OD and DTR; eigenfrequencies are shown in white line contours, whereas global and local utilizations (denoted by GLutil and EUutil in the graph) are given in black solid and dotted contours. The tower-top diameter and DTR are 3.5 m and 170 respectively; the tower-waist elevation over base is at 23.95 m. The green cross denotes the approximate location of the optimized configuration obtained in this study. The red cross denotes the approximate location of a feasible design with a tower-base OD hard limit at 7.5 m.

334

Offshore Wind Farms 1.00

1.00

(b)

0.95

0.95

0.90

0.90

0.80

0.2 0.2 0 0.2 1 0.2 2 0.243 0.2 0.2 5 0.2 6 0.287

1000

0.75

800

0.70

600

0.65

Mass (t)

1200

0.85

1400 GL utilization ratio

0.85

Mass (t)

1400

0.80

1200

0.75

1000

0.70

800 0.65

600 0.60

100

120

140

160 DTR

180

8.6 8.4 8.2 8.0 7.8 0.24 7.6 7.4 7.2 7.0 0.23

200 0.55

0.60 0.55 0.50

0.50

(c)

0.28 0.27 0.26 0.25

GL utilization ratio

(a)

950 0.96 900 0.90

Mass (t)

0.84 800 0.78 750 0.72 700 0.66

EUsh utilization ratio

850

650 0.60 600 550 0.23

0.54 0.24

0.25

0.26

0.27

0.28

Figure 10.41 General trends for the solution space in the preliminary design of an offshore tower supporting a 10-MW turbine: (a) mass (along the z axis) and GLUtil (color-coded markers, with black colors indicating GLUtil > 1) as a function of average DTR (y axis) and first natural frequency (x axis); (b) as in (a) with Db along the x axis and first natural frequency along the y axis; (c) EUUtil ratios (color-coded markers) and associated tower mass (y axis) and calculated eigenfrequency (x axis).

Fig. 10.41(b). Fig. 10.41(c) shows a two-dimensional (2D) view of the same calculated solution space, with emphasis on the EUUtil ratio. This graph demonstrates that if the frequency target is pushed above the current 0.25 Hz, the buckling constraint would no longer be the major driver, being replaced by the modal performance requirement. This, of course, assumes no changes in the boundary conditions, as for example in the stiffness of the SbS. It can also be observed that a change of some 12% in calculated f0 would bring forth a 20% mass penalty. These results were achieved with just a few design variables, ie, with just two tower segments, and keeping a constant cross-section within the bottom segment. By relaxing the latter condition, and by adding one more intermediate station, and thus allowing for three tapered segments, the tower mass can be further reduced as shown in the last column of Table 10.9. A plot of the obtained tower profile and utilization distribution along the span is given in Fig. 10.42.

10.7.1

Component versus system optimization

Traditionally, towers have been specified by the turbine OEM and designed by either the same OEM or in collaboration with a third-party manufacturing firm.


335

Utilization ratio

100

Z from base of tower (m)

Vonmises util1 80

Vonmises util1 EUsh util1 EUsh util1

60

GL util1 GL util2

40

20

0 0.0

0.5


1.5

2.0

Figure 10.42 Tower profile (in the middle of the graph) and utilization distributions for the DLC 1.6 (denoted by suffix 1) and 6.1 (denoted by suffix 2) for the tower with multiple segment optimization.

SbS and foundation, on the other hand, are normally designed by civil engineering firms, which would also define the TP. As such, there has been a need to iteratively exchange information among the RNA and tower OEM engineers, and those in charge of the rest of the SSt. The turbine OEM would pass the admissible system frequency bands to the SbS OEM together with the predicted loads at the base of the tower; the SbS engineers would in turn couple those loads to hydrodynamic calculations and ensure the verification of the limit states for the SbS-foundation system. New stiffness values at the tower base would also be returned to the turbine OEM that would then repeat the loads’ analysis based on this new information. The iterations would continue until a balance is reached between calculated loads and modal constraints. Communication and data-sharing protocols must thus be put in place for this process, which can raise intellectual property (IP) questions and difficulties in the informational exchange. In principle, ULS analysis can employ the sequential approach because ultimate loads can be more easily superimposed under the various DLCs, though potentially yielding a very conservative solutions. FLS analysis, however, is more complicated as it must capture the vibrational coupling excited by the combined hydro-aeroelastic-servo-dynamics, and time domain simulations conducted in sequentially coupled fashion are very onerous. Nonetheless, the sequential approach has worked relatively well for the smaller installations of the North Sea (eg, 3-MW turbine ratings and 20 m water depths) with only a few structural problems. In some cases, fatigue issues concerning monopile TP grouted sections have led to some expensive retrofits. As a result, newer monopile TP designs incorporate shear keys and tapered sections to minimize the tensile stresses in the concrete. Lattice SbSs, or jackets, have also witnessed a few fatigue issues, especially near the welded joints.

336

Offshore Wind Farms

Another flaw of the sequentially coupled design is that system optimization is inherently hampered, because a systems engineering approach cannot be directly applied. As a result, the burden may be passed from the tower designers onto the SbS and foundation engineers, who may need to increase the amount of steel to guarantee modal compliance beyond what is needed for pure load resistance. This normally leads to suboptimal utilizations. The simultaneous optimization of the entire SSt, in contrast, could lead to an overall minimum mass if one accepts some small penalty on the tower mass. So far, the optimization has been limited to a single component and to a single objective function, eg, the tower mass. However, the optimal design of a complex system, such as an OWT, should satisfy several merit functions, including cost functions for manufacturing, installation, maintenance, and decommissioning. Even more importantly, optimum design through subsystem objectives, as the component mass minimization, does not necessarily lead to optimal designs of the entire system LCOE. Ideally, multiobjective function optimization should therefore be employed, which requires accurate models of not only the OWT structural dynamics, but also of the system interaction across a wind plant, and of the processes associated with the entire BOS. As should be clearly deduced from reading this chapter, the design of the offshore tower is intimately connected to that of the SbS. Consequently, in order to arrive at the optimal design of the entire SSt, the tower and SbS should be simultaneously designed and analyzed. As an example, a simple mass optimization of a jacket-tower SSt for a 5-MW turbine was performed. The main environmental, geometric and loading parameters are given in Table 10.10, which also lists the RNA data relevant for this study and the assigned modal constraint. First, an example of sequential optimization was pursued, where the four-legged SbS was optimized assuming a frozen tower configuration. The tower was modified from an onshore design that was originally devised to sustain the same loads as those of the offshore system, and for the same target frequency. The original tower’s peak ULS utilizations can be seen in Fig. 10.43(a). The structure had to be truncated at a height of 22 m (flange level as shown in Table 10.10) to be fitted on top of the TP. The final tower profile and utilizations are shown in Fig. 10.43(b). The jacket geometry (batter, dimensions for the members in the legs, braces, TP girder, as well as pile including embedment length) was optimized under the additional constraints on joint, member, and pile utilizations per [19]. Only two ULS DLCs (see Table 10.10) were considered for the sake of simplicity, with no marine growth or corrosion effects, no windewave misalignment, and load directed along two non-adjacent piles. A 3D view of the optimized SSt skeleton above the seabed is presented in Fig 10.43(c). The associated mass schedule is given in Table 10.11. The optimization exercise was repeated, but this time by simultaneously seeking an optimum design for the entire SSt. Again, the objective was to minimize the overall system mass, and constraints were set to satisfy modal performance criteria and design code checks, for both the tower and the SbS. The mass schedule resulting from the optimized configuration is given in Table 10.11. As can be seen from that table,


337

Main load and environmental parameters for the simple optimization study carried to show the differences between component optimization and system optimization

Table 10.10

Parameter

Value

Unit

Comments

Mass

350

t

Ixx, Iyy, Izz, Ixz

1.15Eþ08, 2.20Eþ07, 1.88Eþ07, 5.04e5

kg m2

Inertial quantities

CMxoff, CMyoff, CMzoff

1.13, 0.0, 50.9e-1

m

RNA CM offsets from tower-top centerline

Hub height

90

m

f0

0.28

Hz

Target frequency

Uref

33

m/s


FxRNA, FyRNA, FzRNA

1.28Eþ06, 0., 1.12Eþ05

N

RNA ULS forces

MxRNA, MyRNA, MzRNA

3.96Eþ06, 8.96Eþ05, 3.47Eþ05

Nm

RNA ULS moments

Uref

70

m/s


FxRNA, FyRNA, FzRNA

1.88Eþ05, 0., 1.65Eþ04

N

RNA ULS forces

MxRNA, MyRNA, MzRNA

0.0, 1.31Eþ05, 0.0

Nm

RNA ULS moments

RNA data

DLC 1.6

DLC 6.1

Environmental data Soil

Sand, stiff

Generic stiff soil assumed

Water depth

41

m

Wave height

17.6

m

50-year wave

Tp

12.5

s

Wave period associated with 50-year wave

e

Tower aerodynamic drag coefficient assumed constant for simplicity

Additional auxiliary data Cd

0.7

Continued

338

Offshore Wind Farms

Table 10.10

Continued

Parameter

Value

Unit

Comments

r/fy

8740/345

kg/m3, MPa

Steel density and yield strength; density augmented to account for secondary steel, coatings, hardware, etc.

Deck height

16

m MSL

Tower flange interface

22

m MSL

Max footprint

16

m

Mudline distance between legs

Max pile Lp

70

m

Max pile embedment length

(a)

60

Z from base of tower (incl.AF) (m)

70

GL util2

50 40 30 20 10 0

Utilization ratio

70

Vonmises util1 Vonmises util1 EUsh util1 EUsh util1 GL util1

80 Z from base of tower (m)

(b)

Utilization ratio

90

0.0

0.5

1.0 1.5 GL and EU utilization ratios

Vonmises util1 GL util1 EUsh util1 Vonmises util2 GL util2 EUsh util2

60 50 40 30 20 10 0 0.0

2.0

0.5 1.0 1.5 Vonmises, GL, and EU utilization ratios

2.0

(c) 120 100 80 60 40 20 0 60 40 –60

–40

0 –20

0

20

40

60

20

–20 –40 –60

Figure 10.43 Examples of jacket optimization based on a fixed tower previously optimized for onshore, and modified to fit the jacket: (a) initial tower configuration and utilizations; (b) modified tower and utilization profiles for the offshore application; (c) jacket-tower configuration obtained.

all subcomponents benefited from the simultaneous approach, with an overall mass saving of some 6%. This example did not consider FLS cases, but one could make the argument that similar conclusions may be derived for those DLCs as well, for a given constant modal response.


339

Table 10.11 Mass [t] schedule for the various components of the SSt in simple optimization study

Component

Sequential optimization

Simultaneous optimization

Relative difference (simultaneous vs. sequential optimization)

Piles

300

290

0.03

Jacket lattice

320

290

0.09

Jacket TP

250

230

0.08

Total jacket

570

520

0.09

Tower

230

205

0.13

Total

1090

1020

0.06

In other situations, it could be shown that increasing the stiffness of the tower, hence its mass, could actually amount to an overall reduced system mass due to a significantly lighter SbS, or yield an SSt reduced footprint [133]. A smaller footprint can even bring positive repercussions on transportation costs. It is important to try to capture these “collateral” effects, as they participate in the final project cost-budget. Note, for example, that if the jacket legs’ OD could be lessened, the associated hydrodynamic loads would further decrease with a positive feedback on the design of the entire OWT and render a compounded effect toward mass savings. While these and other collateral consequences on the LCOE are not captured in this simple example, it is easy to appreciate the effects of simpler manufacturing, transportation, and installation of the SSt. Emphasis should therefore be placed on a system engineering approach to the design of OWTs, as their components (including tower and SbS) are technically and economically more interconnected than their onshore counterparts.

10.8

Final remarks

While the initial and natural progression from land to sea called for a simple extension of the tower in the form of an MP configuration, deeper waters, heavier RNAs, and higher hub heights demand the analysis of more complex support systems. The stringent requirement on modal performance can only partially be tackled by more substantial SbSs, and in order to prevent them from becoming prohibitive, towers will need to become more significant. Compared to onshore, offshore tower design faces new technical challenges in the expanded loading scenarios that include actions from the hydrodynamics, seaeice dynamics, and stress-corrosion. Moreover, while a number of design and certification standards are available, no unified guideline exists. Finally, the biggest challenge still lies within the quest for reduced system LCOE. The tower engineer must be aware of these formidable challenges that the experience gained on land and that of the O&G industry can only partially overcome. Nonetheless,

340

Offshore Wind Farms

relaxed transportation constraints, new materials, renewed control strategies, active damping systems, coupled with more powerful computational resources, are new and exciting technical opportunities to achieve a robust and economical design. This chapter has offered an overview of the loading scenarios and load sources that must be considered when designing an offshore tower, and of the techniques and tools available to calculate the loads. Limit states verification was also presented within the LRFD philosophy following the principal standards of reference and making use of the concepts of reliability and exposure levels. The so-called secondary steel and CPS strategies were also given special attention, as they directly and indirectly affect the distribution of structural steel, and are all but secondary aspects. Simplified design equations were provided notwithstanding emphasis on more rigorous FEA assessments. Focus was on the design of traditional steel tubular towers, with their flanges and shells, as these are still the most common configurations for offshore applications. However, other layouts can be devised following the same principles illustrated herewith, and adding additional detail analyses. For instance, reinforced concrete towers may offer very economical solutions, especially in integrated gravity-based SSts with tall hub heights. Concrete has excellent dynamic (damping) and durability properties, is very versatile in terms of shapes and physical characteristics (strength, stiffness, density), and lends itself to on-site production as well as precast, modular subcomponents. Furthermore, tendon prestress is an efficient way to guarantee overall strength, to connect separate precast rings, and can also be used to fine-tune the tower natural frequencies. All these aspects open up a number of attractive options for the tower designer. Other concepts, such as lattice structures, or a combination of guys and steel/concrete structures can also be envisioned as potential candidates for future installations. In any case, the successful layout will be the one that renders the minimum wind-plant LCOE. To this end, the tower and the remainder of the components of an OWT ought to be simultaneously designed, and in parallel with considerations for BOS and O&M. Offshore systems appear more tightly coupled than their onshore counterpart. For this reason, single-part optimization will likely yield a suboptimal overall design and a greater LCOE. Multiobjective system optimization should thus be pursued, but it remains a formidable task for several reasons. First, multiple engineering disciplines (eg, geotechnical, civil, mechanical, naval, electrical, and industrial engineering), with their different levels of model fidelity, must be combined to perform accurate analyses; second, the number of parameters and design variables is of daunting proportions. Third, the number of CAE tools that can account for all of these aspects is limited, and even fewer exist that can be easily automated within the optimization process. In order to achieve this comprehensive view, it is even more important to be able to characterize the system load response as a fully coupled one, simultaneously accounting for aerodynamics, hydrodynamics, structural and control-system dynamics. For instance, the tower (and entire SSt) design can and should take advantage of advanced control devices and techniques. The economical advantages of the so-called co-design, where parametric control configurations can be modeled together with the OWT structural dynamics to arrive at tunable optimum system layouts, should not be overlooked. As of yet, full optimization is still a non-trivial exercise, with high demand on computational resources and analysis time. Within this context, simplified models are still very


341

valuable and the efficient communication among all the design teams (eg, the WTG OEM and the SSt design firm) is crucial for the success of the wind project. Finally, the role of the experienced designer within each field cannot be overestimated. Policy and energy challenges will soon demand that offshore wind LCOE be further reduced. Although historically OWT component design has been done sequentially, and within a compartmentalized process, multidisciplinary optimization and system engineering approaches seem indispensable for the offshore industry to flourish in the future.

Glossary 2D

Two-dimensional

3D

Three-dimensional

AGL

Above ground level

AHSE

Aero-hydro-servo-elastic

AMD

Active mass damper

ANSYS®

ANSYS® finite element analysis (FEA) package

AOA

Angle of attack

AWEA

American Wind Energy Association

AZ

Atmospheric corrosion zone

BEMT

Beam element momentum theory

BOS

Balance of station

CA

Corrosion allowance

CAD/CAM

Computer-aided design/manufacturing

CAE

Computer-aided engineering

CFD

Computational fluid dynamics

CM

Center of mass

COD

Co-directional wave and wind

CP

Cathodic protection

CPS

Corrosion protection system

CVA

Certification and verification agency

DAF

Dynamic amplification factor

DEL

Damage equivalent load

DES

Damage equivalent stress

DLC

Design load case Continued

342

Offshore Wind Farms

DOF

Degree of freedom

ECD

Extreme coherent gust with direction change

ECM

Extreme current model

EDC

Extreme direction change

EOG

Extreme operating gust

ESS

Extreme sea state

ETM

Extreme turbulence model

EWH

Extreme wave height

EWLR

Extreme water level range

EWM

Extreme wind speed model

EWS

Extreme wind shear

FA

Fore-aft

FAST v8

CAE tool created by National Renewable Energy Laboratory (NREL)

FEA

Finite element analysis

FEED

Front-end engineering design

FLS

Fatigue limit state

FWVM

Free wake vortex method

GA

Genetic algorithm

GBF

Gravity-based foundation

GBO

Gradient-based optimization

GDW

Generalized dynamic wake

GFRP

Glass fiber-reinforced polymer

GMNIA

Geometrically and materially non-linear analyses with imperfection modes

HSWL

Highest still-water level

HSWL

Lowest still-water level

ICCP

Impressed current cathodic protection

IP

Intellectual property

IPC

Independent pitch control

ISO

International Standardization Organization

IZ

Intermediate corrosion zone

LCOE

Levelized cost of energy

LF

Load factor

LRFD

Load resistance factor design


MIC

Microbiologically influenced corrosion

MIS

Misaligned

MP

Monopile

MSL

Mean sea level

MTMD

Multiple tuned mass damper

MUL

Multi-directional wave directions

NCM

Normal current model

NREL

National Renewable Energy Laboratory

NSS

Normal sea state

NTM

Normal turbulence model

NWLR

Normal water level range

NWP

Normal wind profile

NZ

Nacelle corrosion zone

O&G

Oil and gas

O&M

Operation and maintenance

OC3

Offshore code comparison collaboration

OC4

Offshore code comparison collaboration continuation

OD

Outer diameter

OEM

Original equipment manufacturer

OTM

Overturning moment

OWT

Offshore wind turbine

PSD

Power spectral density

PSF

Partial safety factor

RC

Reinforced concrete

RNA

Rotor nacelle assembly

RP

Return period

RSR

Reserve strength ratio

RWH

Reduced wave height

RWM

Reduced wind speed model

SbS

Substructure

SCF

Stress concentration factor

SLS

Serviceability limit state

SO

Structural optimization

343

Continued

344

Offshore Wind Farms

SRB

Sulfate-reducing bacteria

SS

Sideeside

SSI

Soilestructure interaction

SSS

Severe sea state

SSt

Support structure

SWH

Severe wave height

SWL

Still-water level

SZ

splash corrosion zone

TLCD

Tuned liquid column damper

TLP

Tension-leg platform

TMD

Tuned mass damper

TP

Transition piece

ULS

Ultimate limit state

UNI

Unidirectional wave

UZ

Underwater corrosion zone

WTG

Wind turbine generator

List of symbols 1P

The fundamental rotational frequency

Ab

Bolt stress area

Af

Effective cross-sectional area for the compressed flanges

As

Surface area to be protected

Amid

Area inscribed by the mid-thickness line

A

Cross-sectional area

CMxoff

Offset of rotor nacelle assembly (RNA) along x from tower-top centerline

CMyoff

Offset of RNA along y from tower-top centerline

CMzoff

Offset of RNA along z from tower-top centerline

Cd

Drag coefficient

Cm

Added mass coefficient

C

Constant in the seN curve

DTRb

Tower-base DTR


345

DTRt

Tower-top DTR

DTR

Diameter-to-thickness ratio

Db

Tower-base outer diameter (OD)

Dt

Tower-top OD

Dbc

Bolt circle diameter in the flange connection

Dfat

Fatigue damage

Dsh,m

Shell mid-wall diameter

Dsh

Shell OD

EUUtil

Shell buckling utilization ratio

Eb

Bolt material Young’s modulus

Ef

Flange material Young’s modulus

E

Young’s modulus

Fd

Generic, design (factored) load within the load resistance factor design (LRFD) approach

Fj

Forcing associated with the j-th mode of vibration

Fk

Generic, characteristic load within the LRFD approach

Fp

Bolt preload

Ft

Bolt tension load

Fz

Shell tension load

Fsi

Horizontal load from vessel impact

Ft,RD

Bolt strength load

Fu,A

Shell equivalent load in the failure mode A according to the flange segment-model

Fu,B

Shell equivalent load in the failure mode B according to the flange segment-model

Fu,C

Shell equivalent load in the failure mode C according to the flange segment-model

Fult

Ultimate resistance load of the flange connection

FxRNA

Force from the RNA along the x axis

FyRNA

Force from the RNA along the y axis

Fz1

Threshold value of Fz tension load in Schmidt/Neuper’s method

Fz2

Threshold value of Fz tension load in Schmidt/Neuper’s method

FzRNA

Aerodynamic force from the RNA along the z axis

Fzcr

Value of Fz tension load that would cause flange separation Continued

346

Offshore Wind Farms

GLUtil

Global buckling utilization ratio

Gf

Gust factor

Hw

Wave height

Htwr2

Height of tower waist

Ixx

Mass second moment of inertia about the x axis

Ixz

Mass cross-moment of inertia about the x and z axes

Iyy

Mass second moment of inertia about the y axis

Izz

Mass second moment of inertia about the z axis

I

Protective current density (A/m2)

Jxx

Cross-sectional area moment of inertia

Ka

Ratio of bolt stiffness to total connection stiffness

Kb

Ratio of flange stiffness to total connection stiffness

Kb

Bolt axial stiffness

Kf

Equivalent stiffness of the compressed flange pair

Kx

Equivalent soilestructure interaction (SSI) spring constant along x (¼y) axis

Kz

Equivalent SSI spring constant along z axis

Kqx

Equivalent SSI rotational (about x (¼y) axis) spring constant

Kqz

Equivalent SSI torsional (about z axis) spring constant

Lb

Bolt effective length

Lp

Pile embedment length

L

Tower length, or SSt length

Md

Design (factored) bending moment load at the station of interest

Mp

Bending moment load resistance at the tower station of interest

Mx

Component of the bending moment along the x axis at the station of interest

My

Component of the bending moment load along the y axis at the station of interest

Mz

Torsion moment load along the z axis at the station of interest

MPl,2

Flange equivalent plastic bending (resistance) load in failure mode C according to the flange segment-model

MPl,3,MN

Shell equivalent plastic bending (resistance) load in failure modes B and C according to the flange segment-model, accounting for bending and tension interaction

MPl,3

Shell equivalent plastic bending (resistance) load in failure modes B and C according to the flange segment-model

MxRNA

RNA aerodynamic moment along the x axis


347

MyRNA

RNA aerodynamic moment along the y axis

MzRNA

RNA aerodynamic moment along the z axis

Nd

Design (factored) normal load at the tower station of interest

Ni

Number of cycles at failure for the i-th load range level

Np

Normal (axial) load resistance at the tower station of interest

NDEL

Number of cycles at failure for the damage equivalent stress (DES) stress range

NPl,3

Shell equivalent plastic (resistance) load in failure modes B and C according to the flange segment-model

PeD

PeD effect

Q

Practical current capacity (Ah/kg)

R(fd)

Probability distribution of the generic, design (factored) material resistance within the LRFD approach

S(Fd)

Probability distribution of the generic, design (factored) load within the LRFD approach

St

Strouhal number

Sa,i

i-th bin load-range

TL

Expected lifetime of the component

Tc

Expected useful lifetime of the coating

Td

Design (factored) shear load at the station of interest

Tx

Component of the shear load along the x axis at the station of interest

Ty

Component of the shear load along the y axis at the station of interest

Ucr

Critical wind velocity for vortex shedding

Uref

Hub height reference wind speed in simple design load case (DLC) loads’ analysis

C_

Corrosion rate

bi N

Number of cycles at failure for the i-th load range level, accounting for partial safety factors (PSFs)

bi

Unit vector along x axis

bj

Unit vector along y axis

Uw

Wave and current velocity

Uhub

Wind velocity at hub height

U

Wind velocity

fa

Force per unit length due to wind aerodynamic drag

fw

Force per unit length due to wave and current kinematics Continued

348

Offshore Wind Farms

u0,j

First natural frequency in rad associated with the j-th mode of vibration

asi

Added mass coefficient during collision (1.4e1.6 sideway impact, 1.1 bow or stern collision)

a

Distance from bolt centerline to flange edge

b

Distance from bolt centerline to shell mid-wall centerline in the flange connection

cf

Flange circular segment net length (segment-model)

cj

Damping coefficient for the j-th mode of vibration

cs

Shell circular segment length (segment-model)

cc, j

Critical damping coefficient for the j-th mode of vibration

csi

Stiffness of the impacting part of the vessel

db

Bolt hole diameter

dw

Water depth

f0

First natural frequency in Hz

fd

Generic, design (factored) resistance within the LRFD approach

fk

Generic, characteristic material resistance within the LRFD approach

fy

Characteristic yield stress

fobj

Objective function in the optimization problem formulation

fu,b

Bolt characteristic ultimate strength

fy,b

Bolt characteristic yield strength

fy,f

Flange characteristic yield strength

fy,s

Shell characteristic yield strength

gcnt

Constraint function in the optimization problem formulation

g

Gravity acceleration

j

Protective current density (A/m2)

ki

Interaction (axial-hoop stresses) factor in the local buckling utilization calculation

kj

Stiffness associated with the j-th mode of vibration

kw

Dynamic pressure factor to calculate hoop stressesda function of cylinder dimensions and external pressure buckling factor per [2]

kz

Exponent factor for the axial stress ratio, in the local buckling utilization calculation

klat

Lateral restraint spring constant

krot

Rotational restraint spring constant


349

ma

Anode mass

mj

Modal mass associated with the j-th mode of vibration

mRNA

RNA mass

msi

Displacement mass of impact vessel

mtwr

Tower mass, or SSt mass

m

Inverse exponent in the seN curve

nP

n-blades times the fundamental rotational frequency

nb

Number of bolts in the flange connection

ni

Number of cycles at the i-th load range

nDEL

Reference number of cycles for the damage equivalent load (DEL) load range

n

Number of rotor blades

qmax

Maximum wind dynamic pressure

tf

Flange thickness

ts

Shell thickness

u(xd)

States in the optimization problem formulation

vsi

vessel impact speed

xd

States in the optimization problem formulation

xj

Degree of freedom (DOF) variable associated with the j-th mode of vibration

y

Local tower station horizontal displacement

zRNA

z coordinate of the RNA center of mass (CM)

zhub

Hub height above mean sea level (MSL)

z

Altitude above MSL

List of greek symbols Dfn

Factor accounting for member slenderness in the global buckling utilization calculation

as

Wind power-law exponent

l

Reduced slenderness, see Ref. [3]

bm

Bending moment coefficient in the global buckling utilization calculation

dj

Logarithmic decrement damping for j-th mode of vibration Continued

350

Offshore Wind Farms

gK

Stiffness correction factor

gM

Mass correction factor

gf

Generic load PSF

gm

Material PSF

gn

Consequence of failure PSF

gMb,u

Material PSF for the bolt characteristic ultimate strength

gMb,y

Material PSF for the bolt characteristic yield strength

gMf,y

Material PSF for the flange characteristic yield strength

gMs,y

Material PSF for the shell characteristic yield strength

gfa

Aerodynamic load PSF

gfg

Gravitational load PSF

by

Linear combination of assumed eigenmodes

k

Reduction factor in the global buckling utilization calculation

lf

Geometric parameter for the flange connection in Schmidt/Neuper’s method

lf

Geometric parameter for the flange connection

uj

Forcing frequency associated with the j-th mode of vibration

fj

j-th modal coefficient, or periodic function of time

ra

Air density

rw

Sea water density

r

Material density

seN

SeN curve

su

Ultimate strength

sq,Ed

Hoop design (factored) stress

sq,Rd

Hoop buckling strength

sa,i

i-th bin stress-range

seq,i

i-th bin equivalent stress range after Goodman’s correction

sm,i

i-th bin stress-range mean

svm

Von-Mises stress

sz,Ed

Axial (meridional) design (factored) stress

sz,Rd

Axial (meridional) buckling strength

szq,Ed

Shear design (factored) stress

szq,Rd

Shear buckling strength

xj

Damping ratio associated with the j-th mode of vibration


351

x

Damping ratio

z

Dummy coordinate along z axis

ks

Exponent factor the shear stress ratio, in the local buckling utilization calculation

kq

Exponent factor the hoop stress ratio, in the local buckling utilization calculation

Acknowledgments The author would like to thank Senu Sirnivas (NREL) for reviewing this text; Dr. Braulio Barahona and Andy Platt (both from NREL) for providing sample data for postprocessing analysis; Patrick Fullenkamp (GLWN) and Joshua Bauer (NREL communications department) for providing some of the photos and illustrations.

References [1] IEC 61400e3 Wind Turbines e Part 3: Design Requirements for Offshore Wind Turbines, 2009. [2] European Committee for Standardisation, Eurocode 3: Design of Steel StructuresdPart 1e6: General RulesdSupplementary Rules for the Shell Structures, 1993. [3] Germanischer Lloyd, Guideline for the Certification of Offshore Wind Turbines, 2005. [4] S. Tegen, E. Lantz, M. Hand, B. Maples, A. Smith, P. Schwabe, 2011 Cost of Wind Energy Review, Technical Report NREL/TP-5000e56266, National Renewable Energy Laboratory, Golden, Colorado, March 2013. Contract No. DE-AC36-08GO28308. [5] P. Vionis, D. Lekou, F. Gonzalez, J. Mieres, T. Kossivas, E. Soria, E. Gutierrez, C. Galiotis, T. Philippidis, S. Voutsinas, D. Hofmann, Development of a mw scale wind turbine for high wind complex terrain sites; the MEGAWIND project, in: EWEC 2006, EWEA, Athens, Greece, 2006. [6] S. Lim, C. Kong, H. Park, A study on optimal design of filament winding composite tower for 2 mw class horizontal axis wind turbine systems, Int. J. Compos. Mater. 3 (1) (2013) 15e23, http://dx.doi.org/10.5923/j.cmaterials.20130301.03. [7] UDRI exploring composite towers for wind turbines, Compos. Technol. 16 (2) (2010) 14. Trade publication. [8] A. Kayaran, C.S. Ibrahimoglu, Preliminary study on the applicability of semi-geodesic winding in the design and manufacturing of composite towers, in: The Science of Making Torque from Wind 2012, 2012, http://dx.doi.org/10.1088/1742-6596/555/1/012059. [9] J. Cotrell, T. Stehly, J. Johnson, J. Roberts, Z. Parker, G. Scott, D. Heimille, Analysis of Transportation and Logistics Challenges Affecting the Deployment of Larger Wind Turbines : Summary of Results, Technical Report NREL/TP 5000e61063 TP 500061063, National Renewable Energy Laboratory, 2014. [10] A. Bromage, A.H. Triclebank, P.H. Halberstadt, B.J. Magee, Concrete Towers for Onshore and Offshore Wind Farms, Tech. rep., The Concrete Center and Gifford, UK, 2007. [11] DNV, Design of Offshore Wind Turbine Structures, 2013. [12] B. Skaare, F.G. Nielsen, T.D. Hanson, R. Yttervik, O. Havmller, A. Rekdal, Analysis of measurements and simulations from the hywind demo floating wind turbine, Wind Energy 18 (6) (June 2015) 1105e1122.

352

Offshore Wind Farms

[13] D. Roddier, C. Cermelli, A. Aubault, A. Weinstein, WindFWind: a floating foundation for offhore wind turbines, J. Renew. Sustain. Energy 2 (3) (2010). [14] O. Dalhaug, P. Berthelsen, T. Kvamsdal, L. Froyd, S. Gjerde, Z.Z. Zhang, K. Cox, E. Van Buren, D. Zwick, Specification of the NOWITECH 10 Mw Reference Wind Turbine, Technical report, Norwegian Research Centre for Offshore Wind Technology, Trondheim, Norway, 2012. [15] Germanischer Lloyd, Guideline for the Certification of Offshore Wind Turbines, 2012. [16] EN 10025, 2004-European Structural Steel Standard, April 2004. [17] EN 14399, High-Strength Structural Bolting Assemblies for Preloading, 2005. [18] AWEA, Offshore Compliance Recommended Practices Recommended Practices for Design, Deployment and Operation of Offshore Wind Turbines in the United States, 2012. [19] API, Planning, Designing and Constructing Fixed Offshore Platforms e Working Stress Design, aPI RECOMMENDED PRACTICE 2A-WSD, November 2014. [20] ISO 19901-3:2014-Petroleum and Natural Gas Industries e Specific Requirements for Offshore Structures e Part 3: Topsides Structure, 2014. [21] ISO 19902:2007-Petroleum and Natural Gas Industries e Fixed Steel Offshore Structures, 2007. [22] ISO 19903:2006-Petroleum and Natural Gas Industries e Fixed Concrete Offshore Structures, Revised in 2010, 2006. [23] IEC 61400e1. Wind Turbines e Part 1: Design Requirements, 2005. [24] NORSOK, M-501: Surface Preparation and Protective Coating, June 2004. [25] ISO 12944-2:1998 Paints and Varnishes e Corrosion Protection of Steel Structures by Protective Paint Systems e Part 2: Classification of Environments, 1998. [26] ISO 20340:2009 Paints and Varnishes e Performance Requirements for Protective Paint Systems for Offshore and Related Structures, 2009. https://www.iso.org/obp/ui/#iso:std: iso:20340:ed-2:v1:en. [27] API, ANSI/API Recommended Practice 2met e Derivation of Metocean Design and Operating Conditions, November 2014. [28] ISO 19901-1:2005 (modified) e Petroleum and Natural Gas Industries e Specific Requirements for Offshore Structures e Part 1: Metocean Design and Operating Considerations, aNSI/API Recommended Practice 2MET, November 2014. [29] DNV-Risø, Guidelines for Design of Wind Turbines, 2002. [30] European Committee for Standardisation, Eurocode 3: Design of Steel StructuresdPart 1e9: Fatigue, 2005. [31] AISC, ANSI/AISC 360-10-Specification for Structural Steel Buildings, Supersedes the 2005 Edition, 2010. [32] ACI 318-14-Building Code Requirements for Structural Concrete and Commentary, 2014. [33] ACI 357R-84-Guide for the Design and Construction of Fixed Offshore Concrete Structures, Reapproved 1997, 1984. [34] European Committee for Standardisation, Eurocode 4: Design of Composite Steel and Concrete Structures, Re-Approved 2004, 1994. [35] European Committee for Standardisation, Eurocode 2: Design of Concrete Structures, December 2004. [36] Model Code for Concrete Structures 2010, November 2013, 434 p. [37] NS 3473-Concrete Structures- Design Rules. [38] ABS, Guide for Building and Classing e Floating Offshore Wind Turbine Installations, Revised July 2014, January 2013. [39] AWS, AWS 01.1:2000-Structural Welding Code- Steel, 2000.


353

[40] VDI-Fachbereich Produktentwicklung und Mechatronik, Vdi 2230 Systematic Calculation of Highly Stressed Bolted Joints -Part I, December 2014. [41] DNV, Fatigue Design of Offhore Steel Structures, October 2012. [42] ABS, Guide for Building and Classing e Bottom-founded Offshore Wind Turbine Installations, Revised July 2014, January 2013. [43] ABS, Guide for Buckling and Ultimate Strength Assessment for Offshore Structures, February 2014. [44] ABS, Guide for Fatigue Assessment of Offshore Structures, February 2014. [45] API, ANSI/API Recommended Practice 2sim e Structural Integrity Management of Fixed Offshore Structures, November 2014. [46] API, ANSI/API Recommended Practice 2geo e Geotechnical and Foundation Design Considerations, Also ISO 19901-4:2003 (Modified), Petroleum and Natural Gas IndustriesSpecific Requirements for Offshore Structures, Part4Geotechnical and Foundation Design Considerations, November 2011. [47] NORSOK, DNV NO304: Foundations, February 1992. [48] IEC 60364-Electrical Installations of Buildings -part 5-54: Selection and Erection of Electrical Equipment Earthing Arrangements, Protective Conductors and Protective Bonding Conductors, June 2002. [49] ISO 12944-1:1998 Paints and Varnishes e Corrosion Protection of Steel Structures by Protective Paint Systems e Part 1: General Introduction, 1998. [50] ISO 12944-8:1998 paints and Varnishes e Corrosion Protection of Steel Structures by Protective Paint Systems e Part 8: Development of Specifications for New Work and Maintenance, 1998. [51] ISO 19900:2013-Petroleum and Natural Gas Industries e General Requirements for Offshore Structures, Revision to 2002 edition, 2013. [52] R. Damiani, W. Musial, Factors affecting design and reliability of offshore wind turbines and supports for the vowtap project, in: AWEA OffshoreWind- Power 2013, American Wind Energy Association, Providence, RI, 2013. Poster. [53] R. Damiani, W. Musial, Hurricane design guidance for us waters: engineering offshore wind systems to survive hurricanes, in: AWEA Offshore Wind- Power, American Wind Energy Association, Atlantic City, NJ, 2014. Invited Talk. [54] F. Moses, L. Russell, Applicability of Reliability Analysis in Offshore Design Practice: Api-prac Project 79e22, Final report, Tech. rep., American Petroleum Institute, API Publishing Services, 1220 L Street, NW, Washington, DC 20005, 1980. [55] DNV, Guideline for Offshore Structural Reliability Analysisapplications to Jackets, Tech. Rep. 95e3203, Det Norske Veritas AS, Veritasveien 1-N-1322 HVIK, Norway, November 1996. Joint Industry Project. [56] G. Stewart, M. Efthymiou, J. Vughts, Ultimate strength and integrity assessment of fixed offhore platforms, in: T. Moan, N. Janbu, O. Faltinsen (Eds.), International Conference on Behaviour of Offshore Structures (BOSS1988), Tapir Publishers, Trondheim, Norway, 1988, pp. 1205e1221. [57] O. K€ubler, M. Faber, Optimality and acceptance criteria in offshore design, J. Offshore Mech. Arct. Eng. 3 (126) (2004) 258e264, http://dx.doi.org/10.1115/1.1782641. [58] G. Ersdal, O. K€ubler, M.H. Faber, J.D. Sørensen, S. Haver, I. Langen, Economic optimal reserve strength for a jacket structure, in: ASRANet International Colloquium 2004, Barcelona, Spain, 2004. [59] I. Energo Engineering, Reliability vs. Consequence of Failure for API RP2A Platforms Using RP2MET, Final Report 609, US DOI e MMS, 381 Elden St., Herndon, VA 20170, March 2009.

354

Offshore Wind Farms

[60] M. Efthymiou, J.W. van de Graaf, Reliability and (re)assessment of fixed steel structures, in: ASME 2011 30th International Conference on Ocean, Offshore and Arctic Engineering (OMAE2011), vol. 3, ASME, Rotterdam, The Netherlands, 2011, pp. 745e754, http://dx.doi.org/10.1115/OMAE2011-50253. [61] Din 1055-4 (2005-03), Action on Structures e Part 4: Wind Loads, March 2005. [62] European Committee for Standardisation, Eurocode 1: Actions on StructuresdPart 1-4: General Actionsdwind Actions, April 2010. [63] SEI, Minimum Design Loads for Buildings and Other Structures, American Society of Civil Engineers, 2005 aSCE Standard, ASCE/SEI 7e05. [64] SEI, Minimum Design Loads for Buildings and Other Structures (ASCE/SEI 7e10), third ed., American Society of Civil Engineers, 1801 Alexander Bell Drive, Reston, Virginia 20191, 2010. ASCE Standard, ASCE/SEI 7e10. [65] T. Burton, D. Sharpe, N. Jenkins, E. Bossanyi, Wind Energy Handbook, first ed., JohnWiley and Sons, Inc., 111 River St., Hoboken, NJ 07030, USA, 2005. [66] D.A. Peters, C.J. He, Correlation of measured induced velocities with a finite-state wake model, J. Am. Helicopter Soc. 36 (3) (1991) 59e70. [67] R. Damiani, Algorithmic Outline of Unsteady Aerodynamics (Aerodyn) Modules, RRD Engineering, LLC Final Report Subcontract No. AFT-1-11326-01, NREL, Golden, CO, September 2011. [68] J.F. Manwell, J.G. Mcgowan, A.L. Rogers, Wind Energy Explained: Theory, Design and Application, second ed., John Wiley & Sons, Ltd., Chichester, West Sussex, PO19 8SQ, United Kingdom, 2010, 705 p. [69] S. Adhikari, S. Bhattacharya, Vibrations of wind-turbines considering soilstructure interaction, Wind Struct. 14 (2) (2011) 85e112. [70] T. Fischer, W. de Vries, Final Report Task 4.1-Deliverable d 4.1.5-(wp4: Offshore Foundations and Support Structures), Upwind Project 4.1, Universit at Stuttgart, Allmandring 5B, 70569 Suttgart, Germany, 2011. Contract No.:019945 (SES6). [71] I. Prowell, A. Elgamal, C.-M. Uang, J.E. Luci, H. Romanowitz, E. Duggan, Shake table testing and numerical simulation of a utility-scale wind turbine including operational effects, Wind Energy 17 (7) (2013) 997e1016, http://dx.doi.org/10.1002/we.1615. [72] P.J. Murtagh, B. Basu, B.M. Broderick, Gust response factor methodology for wind turbine tower assemblies, J. Struct. Eng. 133 (1) (2007) 139e144, http://dx.doi.org/ 10.1061/(ASCE)0733-9445(2007)133:1(139). [73] J.D. Wheeler, Method for calculating forces produced by irregular waves, J. Pterol. Eng. 249 (1970) 359e367. [74] S.K. Chakrabarti, Hydrodynamics of Offhore Structures, WIT Press, Southampton, UK, 1987, 440 p. [75] Germanischer Lloyd, Guideline for the Construction of Fixed Offhore Installations in Ice Infested Waters, 2005. [76] European Committee for Standardisation, Eurocode 8: Design of Structures for Earthquake Resistance Part 1: General Rules, Seismic Actions and Rules for Buildings, December 2004. [77] R.E. Kj€orlaug, A. Kaynia, A. Elgamal, Seismic response of wind turbines due to earthquake and wind loading, in: A. Cuhna, E. Caetano, P. Ribeiro, G. M€ uller (Eds.), 9th International Conference on Structural Dynamics, EURODYN2014, Porto, Portugal, 2014. [78] H. Huang, J. Ma, R. Xu, X. Wu, Seismic consideration for high rise concrete wind turbine towers, in: 14th World Conference on Earthquake Engineering, Bejiing, China, 2008.


355

[79] P. Passon, M. Khn, S. Butterfield, J. Jonkman, T. Camp, T. Larsen, OC3 benchmark exercise of aero-elastic offshore wind turbine codes, J. Phys. Conf. Ser. 75 (012071) (2007), http://dx.doi.org/10.1088/1742-6596/75/1/012071. The Science of Making Torque from Wind. [80] W. Popko, F. Vorpahl, A. Zuga, M. Kohlmeier, J. Jonkman, A. Robertson, Offshore code comparison collaboration continuation (OC4), phase i results of coupled simulation of offshore wind turbine with jacket support structure, in: 22nd International Offshore and Polar Engineering Conference (ISOPE), vol. 1, ISOPE, Rhodes, Greece, 2012, pp. 337e346. [81] W. Popko, F. Vorpahl, J. Jonkman, A. Robertson, OC3 and OC4 projects verification benchmark of the state-of-the-art coupled simulation tools for offshore wind turbines, in: 7th European Seminar Offshore Wind and Other Marine Renewable Energies in Mediterranean and European Seas (OWEMES), Rome, Italy, 2012, pp. 403e407. [82] ANSI, Specification for Structural Steel Buildings, 3rd printing: 2013, June 2010. [83] European Committee for Standardisation, Eurocode 3: Design of Steel StructuresdPart 1e1: Genstructures and Rules for Buildings, 2005. [84] DNV, Offshore Standard DNV-OS-C501-composite Components, October 2010. [85] A.Z. Palmgren, Die lebensdauer von kugellagcrn, Z. Ver. Deutsch. Ing. 68 (1924) 339e341. [86] M. Miner, Cumulative damage in fatigue, J. Appl. Mech. 12 (1945) A159eA164. [87] S.D. Downing, D.F. Socie, Simple rainflow counting algorithms, Int. J. Fatigue 4 (1) (1982) 31e40. [88] M. Veljkovic, C. Heistermann, W. Husson, M. Limam, M. Feldmann, J. Naumes, D. Pak, T. Faber, M. Klose, K. Fruhner, L. Krutschinna, C. Baniotopoulos, I. Lavasas, A. Pontes, E. Ribeiro, M. Hadden, R. Sousa, L. da Silva, C. Rebelo, R. Simoes, J. Henriques, R. Matos, J. Nuutinen, H. Kinnunen, High-Strength Tower in Steel for Wind Turbines (Histwin), Final Report EUR 25127 EN, Directorate-General for Research and Innovation e European Commission, Luxembourg, Europe, 2012. Contract No RFSR-CT-2006e00031. [89] C. Petersen, Grundlagen der berechnung und baulichen ausbildung von stahlbauten, Stahlbau 66, Vieweg-Verlag, Braunschweig, Germany. [90] P. Schaumann, M. Seidel, Failure analysis of bolted steel flanges, in: X. Zhao, R. Grzebieta (Eds.), Proceedings of the 7th International Symposium on- Structural Failure and Plasticity (IMPLAST 2000), Melbourne, Australia, 2000. [91] R. Juvinall, K. Marshek, Fundamentals of Machine Component Design, fifth ed., John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ, 2012. [92] K. Brown, C. Morrow, S. Durbin, A. Baca, Guideline for Bolted Joint Design and Analysis: Version 1.0, SANDIA REPORT SAND2008e0371, Sandia National Laboratories, Albuquerque, New Mexico, January 2008. [93] H. Schmidt, M. Neuper, Zum elastostatischen tragverhalten exzentrisch gezogener l-ste mit vorgespannten schrauben, Stahlbau 66 (3) (1997) 163e168. [94] M. Seidel, Zur bemessung geschraubter ringflanschverbindungen von windenergieanlagen (Ph.D. thesis), Hannover University, Hannover, Germany, 2001. [95] M. Brodersen, J. Hgsberg, Damping of offshore wind turbine tower vibrations by a stroke amplifying brace, in: Energy Procedia, ERA Deep-Wind 2014, 11th Deep Sea Offshore Wind R&D Conference, Trondheim, Norway, vol. 53, 2014, pp. 258e267. [96] A. Tsouroukdissian, C. Carcangiu, I. Pineda, T. Fischer, B. Kuhnle, M. Scheu, M. Martin, Wind turbine tower load reduction using passive and semi-active dampers, in: EWEA 2011, 2011. [97] A. Sigaher, M. Constantinou, Scissor-jack-damper energy dissipation system, Earthq. Spectra 19 (1) (2003) 133e158. eERI.

356

Offshore Wind Farms

[98] M. Constantinou, P. Tsopelas, W. Hammel, A. Sigaher, Toggle-brace-damper seismic energy dissipation systems, J. Struct. Eng. 127 (2) (2001) 105e112. aSCE. [99] H. Gao, K. Kwok, B. Samali, Optimization of tuned liquid column dampers, Eng. Struct. 19 (6) (1997) 476e486. [100] S. Colwell, B. Basu, Tuned liquid column dampers in offshore wind turbines for structural control, Eng. Struct. 31 (2) (2008) 358e368, http://dx.doi.org/10.1016/ j.engstruct.2008.09.001. [101] C. Roderick, Vibration Reduction of Offshore Wind Turbines Using Tuned Liquid Column Dampers, Mechanical Engineering, University of Massachusetts Amherst, Amherst, MA, September 2012. [102] A. Wilmink, J. Hengeveld, Application of tuned liquid column dampers in wind turbines, in: EWEC 2006, EWEA, Athens, Greece, 2006. [103] J. Li, Z. Zhang, J. Chen, Experimental study on vibration control of offshore wind turbines using a ball vibration absorber, Energy Power Eng. 4 (2012) 153e157, http://dx.doi.org/10.4236/epe.2012.43021. [104] M. Pirner, Acutal behavior of a ball vibration absorber, Wind Eng. Ind. Aerod. 90 (8) (2002) 987e1005. [105] B. Fitzgerald, B. Basu, Active tuned mass damper control of wind turbine nacelle/tower vibrations with damaged foundations, Key Eng. Mater. 569e570 (2013) 660e667, http:// dx.doi.org/10.4028/www.scientific.net/KEM.569-570.660. [106] G.M. Stewart, Load Reduction of Floating Wind Turbines Using Tuned Mass Dampers, Mechanical Engineering, University of Massachusetts, Amherst, Amherst, MA, February 2012. [107] N.E. Dowling, Mechanical Behavior of Materials e Engineering Methods for Deformation, Fracture, and Fatigue, fourth ed., Prentice Hall, 2012, p. 960. [108] J. Barsom, S.T. Rolfe, Fracture and Fatigue Control in Structures, third ed., in: Applications of Fracture Mechanics, ASTM, 100 Barr Harbor Dr., West Conshohocken, PA 19428-2959, 1999, p. 548. [109] S. Suresh, Fatigue of Materials, second ed., Cambridge University Press, 1998 http:// dx.doi.org/10.1017/CBO9780511806575. [110] R.P. Gangloff, Corrosion Tests and Standards, No. MNL 20 in ASTM Manual Series, second ed., ASTM, West Conshohocken, PA, 2005, pp. 302e322. Astm Manudal Series 26- Environmental Cracking e Corrosion Fatigue. [111] H.H. Lee, H.H. Uhlig, Corrosion fatigue of type 4140 high strength steel, Metall. Trans. 3 (11) (1972) 2949e2957. [112] A. Momber, Corrosion and corrosion protection of support structures for offshore wind energy devices (owea), Mater. Corros. 62 (5) (2011) 391e404, http://dx.doi.org/10.1002/ maco.201005691. [113] R. Sheppard, F.J. Puskar, MMS TA&R Project 627 Inspection Methodologies for Offshore Wind Turbine Facilities Final Report, Energo Report Energo Project No.: E08147, Minerals Management Service (MMS), Houston, TX, January 2009. [114] K. M€uhlberg, Corrosion protection on offshore wind turbines a challenge for the steel builder and paint manufacturer, J. Prot. Coat. Linings 27 (3) (2010) 20e32. [115] R.E. Sheppard, F.J. Puskar, C. Waldhart, Inspection guidance for offshore wind turbine facilities, in: OTC (Ed.), Offshore Wind Energy Special Session- Offshore Technology Conference, OTC, Houston, TX, 2010, http://dx.doi.org/10.4043/20656-MS. [116] P. Hogg, Durability of wind turbine materials in offshore environments, in: SUPERGEN Wind Phase 2e4th Training Seminar, SUPERGEN Wind, Manchester, UK, 2012.


357

[117] M. de Jong, Adaptations to a Marine Climate, Salt and Water OWEZr11120101020. Results Corrosion Inspections Offshore Wind Farm Egmond aan Zee, 2007e2009. Tech. € Rep. 50863231 e TUS/NRI10-2242, KEMA Nederland B.V, Arnhem; The Netherlands, October 2010. [118] L.R. Hilbert, A.R. Black, F. Andersen, T. Mathiesen, Inspection and monitoring of corrosion inside monopile foundations for offshore wind turbines, in: E. Conference (Ed.), Eurocorr 2011, no. 4730, EFC, Stockholm, Sweden, 2011. [119] A.R. Black, Hi, Corrosion monitoring of offshore wind foundation structures, in: EWEA Offshore 2013, EWEA, 2013. [120] J. Van der Tempel, Design of Support Structures for Offhore Wind Turbines (Ph.D. thesis), TU Delft, Stevinweg 1, 2628 CN Delft, The Netherlands, April 2006, 209 p. [121] M. Muskulus, S. Schafhirt, Design optimization of wind turbine support structures-a review, J. Ocean Wind Energy 1 (1) (2014) 12e22. [122] K. Dykes, R. Meadows, F. Felker, P. Graf, M. Hand, M. Lunacek, J. Michalakes, P. Moriarty, W. Musial, P. Veers, Applications of Systems Engineering to the Research, Design, and Development of Wind Energy Systems, Tech. Rep. TP-5000e52616, NREL, 1617 Cole Boulevard, Golden, Colorado, December 2011. Contract No. DE-AC3608GO28308. [123] AWEAWindPower 2012, AWEA, Atlanta, GA, 2012, p. 1. Poster NREL/PO-500054717. [124] K. Dixon, E. Mayda, Blade design at siemens wind power, in: NREL 2nd Wind Energy Systems Engineering Workshop, NREL, 2013. Presentation. [125] S.A. Ning, R. Damiani, P.J. Moriarty, Objectives and constraints for wind turbine optimization, J. Solar Energy Eng. 136(4), 12. [126] http://www.fusedwind.org/. [127] P.W. Christensen, A. Klarbring, An Introduction to Structural Optimization, in: Solid Mechanics and Its Applications, vol. 153, Springer, 2009. [128] J. Arora, Introduction to Optimum Design, third ed., Academic Press e Elsevier, 225 Wyman St., Waltham, MA 02451, USA, 2011, 896 p. [129] S.A. Burns, Recent Advances in Optimal Structural Dlesign, SEI-ASCE, 2002, 312 p. [130] R. Levy, O. Lev, Recent developments in structural optimization, J. Struct. Eng. ASCE 113 (9) (1987) 1939e1962. [131] M.P. Kamat, Structural Optimization: Status and Promise, in: Progress in Astronautics and Aeronautics, vol. 150, AIAA, 1993. [132] A. Cherkaev, Variational Methods for Structural Optimization, in: Applied Mathematical Sciences, vol. 140, Springer-Verlag, New York, Berlin, Heidelberg, 2000, 627 p. http:// www.math.utah.edu/books/vmso/. [133] R. Damiani, H. Song, A jacket sizing tool for offshore wind turbines within the systems engineering initiative, in: Offshore Technology Conference, Houston, Texas, USA, 2013, http://dx.doi.org/10.4043/24140- MS.


Design of floating offshore wind turbines

11

M. Collu Cranfield University, Bedford, United Kingdom M. Borg Technical University of Denmark (DTU), Lyngby, Denmark

11.1

Introduction

As countries aim to generate more electricity from wind energy, the exploitation of offshore wind energy resources has become increasingly important. The significantly larger wind resources available offshore in water depths greater than 50 m have attracted interest in developing floating support structures for wind turbines as fixed foundations do not remain economically viable (Collu et al., 2010). As many countries, such as Japan, the United States and European countries along the Atlantic Ocean, have very limited areas with shallow water depths (below 50 m), the economic barrier of fixed support structures, such as monopoles and jacket structures, can be overcome with floating support structures. In the pursuit of deploying floating wind turbine technology in the offshore environment, a small number of large-scale prototypes have been tested offshore, bringing the technology closer to market and raising the Technology Readiness Level of such technology to 7e8. However, these designs have been conservative in nature to reduce the risks involved, and hence more costly. The use of traditional offshore oil and gas industry design standards and technology also increases costs as safety margins inflated associated capital costs.

11.1.1 Classification of floating wind turbines 11.1.1.1 Classification based on static stability Floating wind turbines are generally classified based on the configuration of the support structure adopted. The support structure is classified based on the main approach adopted to fulfil the static stability requirements in the rotational degrees of freedom (pitch and roll), ie, how the structure counteracts the inclining moment due mainly to the aerodynamic forces acting on the wind turbine.


360

Offshore Wind Farms

As will be derived in Section 11.2.2, the total pitch/roll restoring moment (counteracting the inclining moment) can be calculated as the sum of three contributions: 1

0

C B C B MR;roll ¼ B rgIxx þ FB $zCB mg$zCG þ C44;moor CsinðfÞ @ |ffl{zffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflffl{zfflfflfflffl} A a

0

a

b

b

g

g

1

[11.1]

Bzffl}|ffl{ zfflfflfflfflfflfflfflfflfflfflfflfflfflffl}|fflfflfflfflfflfflfflfflfflfflfflfflfflffl{ zfflfflfflffl}|fflfflfflffl{C MR;pitch ¼ @ rgIyy þ FB $zCB mg$zCG þ C55;moor AsinðqÞ

where a is the ‘waterplane area’ contribution, proportional to the seawater density (r), the gravitational acceleration constant (g), and the second moment of the waterplane area (in roll, Ixx, and in pitch, Iyy). If this is the main contribution to the roll/pitch restoring moment, the support structure is said to be ‘waterplane stabilised’; b is the contribution due to the relative position of the centre of buoyancy (B) and the centre of gravity (G), and it is usually called somewhat inaccurately the ‘ballast’ contribution, even if it is linked not only to the vertical position of the G (inertial characteristic), but also to the vertical position of the B (geometric characteristic). The name derives from the fact that usually a large amount of ballast material is used in a position close to the keel of the structure to lower the global G position. FB is the buoyancy force, while m is the total mass of the support structure. Due to the fact that the mooring system imposes a downward force on the floating support structure, in general FB is higher than the total weight of the floating offshore wind turbine system (FOWT) (mg). If this is the main contribution to the roll/pitch restoring moment, the support structure is said to be ‘ballast stabilised’; g is the contribution due to the mooring system. While this contribution can be considered negligible for catenary mooring systems, it can be the main pitch/roll restoring mechanism for TLP (Tension-Leg Platform) systems. In this case, the FOWT is said to be ‘mooring stabilised’.

11.1.1.2 Classification societies Bureau Veritas (BV, 2010) adopts the classification criterion illustrated in the previous section, having three categories: ballast floating platforms (term b), tension leg platforms (term g), and buoyancy floating platforms (term a). The American Bureau of Shipping (ABS, 2013) adopts a criterion based on the structural elements of the different floating support structures, without expressly mentioning the stabilising mechanism. It distinguishes among ‘TLP-type’, ‘SPARType’, and ‘Column-Stabilised’ floating support structures. Nonetheless, these three categories can be related, respectively, to the terms g, b, and a in Eq. [11.1]. In the Det Norske Veritas (DNV, 2013) offshore standard ‘Design of Floating Wind Turbine Structures’, the criterion adopted is different from the previous ones, and it considers whether a structure is restrained (displacements in the order of


361

centimetres) or compliant (displacements in the order of metres and above) in global modes of motions.

11.1.2 Floating wind turbines: examples 11.1.2.1 SPAR: Hywind demo by Statoil Hywind demo by Statoil is a 2.3-MW (Siemens SWT-2.3e82) FOWT system using a SPAR as a floating support structure. Installed in 2010, 12 km off the island of Karmøy, north of Stavanger, in Norway, it is the world’s first full-scale floating wind turbine (Statoil, 2010). As for every SPAR system, it is characterised by a large draught (100 m), making it a suitable solution only for deep waters. As mentioned in Section 11.1.2.1, a SPAR system relies mainly on term b of Eq. [11.1] to ensure its static stability in roll and pitch, ie, on a low vertical position of the G relatively to the B. For Hywind, this is achieved through a combination of heavy materials (rocks) and seawater ballast tanks at the bottom of the hull.

11.1.2.2 Semi-submersible/tri-floater: WindFloat prototype (WF1) by Principle Power and Fukushima FORWARD phase I (Mirai) WindFloat prototype (WF1) The WindFloat prototype was deployed in October 2011, 5 km off the coast of Aguçadoura, Portugal, by Principle Power, equipped with a Vestas 2.0-MW wind turbine (Principle Power, 2012). It is a three-legged, semisubmersible type, floating platform. This configuration is also called a ‘Trifloater’. In this case, the static stability is achieved mainly through the large second moment of the waterplane area (term a in Eq. [11.1]), thanks to three relatively large water-piercing cylindrical columns (10.7 m diameter), and a column centre-to-centre distance equal to 56.4 m. This allows the structure to have a relatively shallow draft, and therefore the possibility to be deployed in relatively shallow waters. Principle Power claims that a 5-MW WindFloat system can have a draught lower than 20 m, and be deployed in a water depth higher than 40 m. This system is fitted with the patented water entrapment (heave) plates at the base of each column. This system has two main effects: it augments the added mass in heave, reducing its natural frequency (augmenting its natural period, moving it further from the wave peak energy frequency) and it augments the damping in heave, reducing the response of the structure to waves, especially near the system heave natural frequency. Further information is available at http://www.demowfloat.eu/. Fukushima FORWARD Phase I After the Great East Japan Earthquake of 2011, Japan has started a number of initiatives looking at developing its renewable energy technology capabilities and portfolio. Among this, the Fukushima FORWARD (Floating OffshoRe Wind fARm Demonstration project) consortium was formed, funded by the Japanese Ministry of Economy, Trade and Industry. There are two

362

Offshore Wind Farms

phases to the project FORWARD: Phase I (2011e13), during which a floating substation (Fukushima Kizuna) and a compact semi-submersible FOWT of 2 MW (Fukushima Mirai) were designed, manufactured, and commissioned, which started to produce electricity in November 2013, and Phase II (2014e2015), currently looking at the design, manufacturing and commissioning of two 7 MW FOWT alternative designs: an advanced SPAR and a V-shaped semi-submersible. The Fukushima Mirai, whose floater design, manufacturing, and commissioning have been coordinated by Mitsui Engineering & Shipbuilding Co. Ltd., consists of one central column, on top of which is installed the 2-MW wind turbine, and three water-piercing side columns, with diameter around 7.5 m, height of 32 m and a draft around 16 m, and a centre-to-centre distance of about 50 m, which, similarly to the WindFloat prototype, provide the static stability thanks to their second moment of area. The three columns are connected to the central one through six horizontal braces and three diagonal braces. The main difference is that for the Phase I prototype FORWARD the wind turbine is installed on the central column, while the WF1 prototype does not have a central column and the wind turbine is installed on top of one of the side columns. Another interesting characteristic of the FORWARD Phase I prototype is that the wind turbine rotor control system not only optimises the power generation depending on the wind speed, but also minimises the floater motion through the wind turbine. This is an advanced feature pioneered by Statoil with the Hywind project (Driscoll et al., 2015). Further information is available at http://www.fukushimaforward.jp/english/research/index.html.

11.1.2.3 TLP: BlueH phase 1 prototype by BlueH As proof of concept, Blue H Group Technologies Ltd deployed off the coast of southern Italy a 75% scale prototype of their TLP system, equipped with a small wind turbine (0.08 MW). After 6 months at sea, the unit was decommissioned early in 2009 (Blue, 2004b). In general, a TLP system could potentially be one of the most suitable platforms for offshore wind turbines, as the displacements can be the smallest, if compared to the other floating support structures. The major drawback is the high cost of the mooring system. The TLP concept has been investigated and proposed by a number of companies and research institutes. Unfortunately, little information is available in the literature about the BlueH, but it has been one of the first of the few deployed TLP prototypes. BlueH claims that this technology can be deployed in water in excess of 60 m, thanks also to its patented deployable TLP system (Blue, 2004a). In principle, a TLP system can have very little waterplane area and a relatively high vertical position of G, since the necessary stability in pitch/roll is achieved through the stiffness of the tension legs of the platform (g in Eq. [11.1]). Nonetheless, a small waterplane area and relatively high G can pose substantial challenges in terms of stability during the transport phases, while the mooring system is not connected to the platform yet (ie, g in Eq. [11.1]). For this reason, the floating support structure is usually a hybrid configuration, exploiting the other two stability terms during the transport phases, and the mooring contribution in the operational site.


363

This is also the case for the BlueH prototype, where the stability during the transport phase is achieved through a combination of a water-piercing multi-column system (second moment of waterplane area) and a heavy ballast system lowering G (part of the deployable TLP mooring system).

11.2

Design of floating offshore wind turbines: main preliminary steps

If the present floating wind turbine concepts and prototypes are analysed, the focus of the design has been on the floating support structure, rather than on the design of the whole system, since the tendency is to adopt an already commercially available horizontal axis wind turbine (HAWT) developed for the fixed offshore wind market. This approach has its advantages, but also its limitations (Sections 11.2.4 and 11.3.1). In the following sections the main requirements and constraints driving the design of a FOWT will be illustrated. Then they will focus on how to analyse the hydrostatics and ensure the static stability of the structure. Considerations about the best approach to analyse the dynamic response of the structure to the metocean conditions will be given, and to conclude an overview of further aspects to be analysed will be provided. The present approach is useful to investigate and narrow down the design space of the potential suitable configurations. For a more detailed design, the reader is referred to the documentation issued by the classification societies.

11.2.1 Main requirements and constraints 11.2.1.1 Floatability The first requirement is to have the sum of the vertical forces equal to zero, ie, the buoyancy force is able to counteract the total weight of the system plus all the other downward forces, such as the mooring forces. This requirement can be translated into a minimum required draught, sufficient to have the necessary displaced volume V. Nonetheless this requirement is usually less stringent than the other requirements on draught (eg, minimum draught to avoid slamming loads). As a consequence, ballast tanks are required (generally filled with seawater) to satisfy the other more stringent minimum draught requirements.

11.2.1.2 Maximum inclination angle While FOWT systems can experience relatively large inclination angles (in roll and/or pitch), onshore and fixed-to-seabed offshore wind turbines do not experience such angles. As a consequence, there is very little experience in estimating the performance of wind turbines at large inclination angles, and relatively few data have been presented in literature.

364

Offshore Wind Farms

Taking also into account the fact that many of the sub-systems of an offshore wind turbine (bearings, gearbox, generator, etc.) have been designed to operate close to the upright condition, it is necessary to impose a maximum roll/pitch inclination angle. The exact value of this maximum inclination angle is still open to discussion, but according to the literature a good starting value is 10 degree. It is important to remember that this is the total angle of inclination, the sum of the static and the dynamic angles of oscillations, due, respectively, to the average value (mainly due to the wind) and the oscillation amplitude (mainly due to waves) of the inclining moments. In terms of the design, this requirement can be translated into a floating support structure minimum rotational stiffness (Section 11.2.2).

11.2.1.3 Freeboard height and minimum draught In heavy storms, the relative vertical motion between the FOWT and the waves can potentially lead to so-called ‘green water loads’, due to a large body of water flowing on the top of the support structure. In order to avoid these loads it is necessary to have a minimum vertical distance between the mean undisturbed seawater level and the level at which no green water loads are considered, such as the top of the support structure: this distance is the minimum freeboard height. Its value depends on the local metocean conditions, but for a conceptual/preliminary design a good starting value is 10 m. Slamming can be defined as ‘a phenomenon described broadly as severe impacting between a water surface and the side or bottom of a hull where the impact causes a shock-like blow’ (ITTC, 2014). It occurs when the relative motion between the structure and the water surface is such that the bottom of the floating structure is above the water elevation. In order to avoid this a minimum draught requirement is imposed that depends on the local metocean conditions. As previously mentioned, in general this minimum draught requirement is more stringent than the one required for floatability, and therefore ballast tanks are required. A good starting value is 15 m.

11.2.1.4 Optimum dynamic response to wind and wave forces Wave forces impose oscillatory motions to the FOWT system, and these motions should be minimised since they may impact negatively on the system performance. In order to evaluate the response to wind and wave forces two main approaches are adopted: frequency domain and time domain (see Section 11.2.3).

Frequency approach Through a frequency analysis approach, it is possible to estimate the system regime responses. Once the wind farm site has been chosen, it is possible to model the relative wave spectrum, illustrating how the energy of the waves is spread with respect to the frequencies. Once a conceptual/preliminary design of the FOWT system is defined, it is possible to derive the main inertial (mass, G and moments of inertia) characteristics of the


365

whole system and the geometric characteristics of the submerged part of the system. From these, it is possible to derive the two main transfer functions linking the amplitude of the wave to the amplitude of the motion of the system. The final transfer function between the amplitude of the oscillation of the structure (at a frequency u) in the ith degree of freedom (d.o.f.) and a wave of frequency u of unit amplitude is called RAOi (Response Amplitude Operator). Using the wave spectrum for the given location and the RAOi of the considered FOWT system, it is possible to estimate the wave response spectrum of that FOWT system in the given location. This wave response spectrum should be minimised in order to minimise the displacements and accelerations of the FOWT system. The important point is that the natural frequencies (periods) of the FOWT system should be outside the most energetic frequency (period) range of the wave spectrum. This depends on the location, but in general wave spectra are most energetic between the 5and 25 s period (1.25e0.25 rad/s), and therefore the structure should aim at having natural periods above 25 s or below 5 s in all the d.o.f. An important aspect should be highlighted, since it is a source of misunderstandings. The RAO concept is strictly valid only to estimate the regime response to waves, and by definition is a linear approach. Since the FOWT system experiences substantial aerodynamic forces as well, mathematically speaking if these forces are considered the RAO concept is no longer valid. Nonetheless, the RAO concept is still used, specifying that it is valid just for a given wind velocity (and for a given wind turbine rotational velocity), and sometimes these are called ‘pseudo-RAOs’. It is therefore important to remember that: (1) these ‘pseudo-RAOs’ are valid only for a given wind speed and (2) since the aerodynamic loads are not linear with the wind velocities, an RAO for each wind speed condition is required. This also highlights the limitations of a frequency approach when analysing the dynamics of FOWT systems.

Time domain With a time-domain approach, it is possible to adopt a time-domain coupled model of dynamics and therefore be able to take into account nonlinear forces and also estimate the transient regimes. In this case, it is possible to estimate the displacements, velocities, accelerations, and time responses of the system in all d.o.f., but also the loads acting on the structure. Through a statistical analysis, it is then possible to estimate the maximum, minimum, mean, variance, standard deviation, and significant values of each of these parameters. The advantage is that it is possible to have a more realistic estimate of these values, but the disadvantage is that it is more difficult to understand in depth how to modify the design in order to obtain a more suitable response to wind and wave forces.

11.2.2 Hydrostatics and stability In the following sections a quick review of the hydrostatics and stability analysis applied to FOWT systems will be given, largely based on Borg and Collu (2015).

366

Offshore Wind Farms

11.2.2.1 Simplified approach and relevant hypotheses The basic theory of hydrostatics of floating bodies is well known e see, for example, Patel (1989). In this section it is applied to FOWT systems. The hydrostatic characteristics of a floating body can be derived applying the prime principle, ie, the integration of the pressures acting on the submerged area of the body, but here a simplified approach and the relevant hypotheses are presented, useful at conceptual/preliminary design level to quickly explore and narrow down the design space. The simplifying hypotheses are: • • •

the fluid in which the body is immersed is considered at rest, the body is always in equilibrium and therefore the amount of submerged volume is constant during the (quasi-static) rotation, the angle of inclination of the body is small (small angle approximation). In most cases, for FOWT systems, it means an angle lower than 10e15 degrees.

This is usually called ‘initial stability’ analysis. The present approach analyses the pitch rotational degree of freedom (rotation about the y axis), but can be easily extended to roll.

11.2.2.2 Axis system and reference points An orthogonal axis system is defined, with x aligned with the direction of the wind, z perpendicular to x and vertical upward, and the origin coincident with the centre of flotation (F) (Fig. 11.1). Z' Fenv

CP(env) Additional submerged volume

Volume no longer submerged

θ

FB

O=F

B Fmoor,H = Fenv

Z

X' X

MLA MR,moor

G

Fmoor,V mg

Figure 11.1 Forces and moments acting on a floating wind turbine system, longitudinal plane. Adapted from Borg, M., Collu, M., 2015. A comparison between the dynamics of horizontal and vertical axis offshore floating wind turbines. Philosophical Transactions of the Royal Society A 373 (2035).


367

The centre of buoyancy (B) is the centroid of the submerged volume of a body through which the total buoyancy may be assumed to act. The centre of flotation (F) is the centroid of the waterplane area, i.e. the area enclosed by a waterline. A waterline is the intersection line of the free water surface with the moulded surface of the body. The centre of gravity (G) is the centre through which all the weights constituting the system may be assumed to act. The centre of mooring line action (MLA) is here defined as the intersection of the line of action of the horizontal component of the mooring force with the z axis, and is the reference point of the mooring line action. The environmental forces acting on the FOWT system will be: aerodynamic forces, hydrodynamic forces, and current forces. If an equilibrium state is considered with no waves, constant wind speed and current forces, the centre of pressure of environmental forces (CP(env)) is defined as the point where the sum of the environmental forces (Fenv) acts on.

11.2.2.3 Balance of vertical forces The main vertical forces acting on the structure in an equilibrium state are (Fig. 11.1): the total weight of the system (mg), the vertical component of the total force due to the mooring system, Fmoor,V, and the buoyancy force FB. Therefore: FB mg Fmoor;V ¼ 0 FB ¼ mg þ Fmoor;V

[11.2]

11.2.2.4 Inclining and restoring moments Referring to Fig. 11.1, the horizontal component of the sum of the environmental forces Fenv is counteracted by the horizontal component of the mooring system force (F ¼ Fenv). Then the inclining moment (in the xz plane) MI can be estimated as Fenv times the vertical distance between CP(env) and the point where Fenv is counteracted, CMLA, or: MI ¼ Fenv zCPðenvÞ zMLA cos q

[11.3]

The moments counteracting the inclining moment, whose sum is the restoring moment, can depend on three system characteristics: geometrical, inertial (mass and G), and in the case of tensioned mooring systems (eg, TLP), on the mooring system. The initial position of B and the second moment of the waterplane area (linked to the movement of B when the platform is inclined) are the characteristics determining the geometry contribution to the restoring moment. For freely floating bodies (like ships), FB is equal to mg (total system weight), and all the contributions are summarised in the parameter called metacentric height, GM. For an offshore floating wind turbine, FB can be higher than mg, due to the downward force of the mooring system. It is then preferred to classify the stabilisation mechanisms in (Eq. [11.4]): a term taking into account the waterplane area contribution

368

Offshore Wind Farms

(geometric), a term taking into account the relative position of G and the initial B position (geometric-inertial) (also called ‘ballast’ term), and a term related to the (possible) contribution of the mooring system (mooring), as follows: MR ¼

rgIy q þ ðFB zCB mgzCG Þq þ C55;moor q ¼ C55;tot q |fflffl{zfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflffl{zfflfflfflfflfflffl} waterplane B-G relative position mooring contribution

ð‘ballast’Þ

contribution

contribution

[11.4]

¼ MR;WP þ MR;CG þ MR;moor where Iy is the second moment of area of the initial waterplane area (within the approximation of small inclination, the waterplane area remains constant) with respect to the x axis, q is the pitch inclination angle, FB is the buoyancy force, zB is the vertical position of B, m is the total mass of the system, zG is the vertical position of G, and C55,moor is the contribution of the mooring system to pitch stiffness. There are several mooring systems that have been adopted by the offshore floating wind industry. In general, its contribution to the total stiffness of the FOWT system may be represented by a 6 6 matrix, since it can generate counteracting forces in all degrees of freedom, and there can be coupled terms. For the present 1 d.o.f. analysis it is assumed that the restoring moment in pitch is only proportional to the rotational displacement in pitch (decoupled from the other d.o.f.).

11.2.2.5 Floatability and maximum inclination angle requirements As regard floatability, the requirement is then imposed using Eq. [11.2]. Imposing a maximum angle (qmax) of inclination in the design phase is equivalent to, given an inclining moment (Eq. [11.3]), imposing a minimum total stiffness, or: MI C55;tot

¼ qequilibrium qmax

C55;tot

MI qmax

C55;totðminÞ ¼

11.2.3

[11.5]

MI qmax

Dynamic response

11.2.3.1 Time domain versus frequency domain One of the major choices in modelling FOWTs is to perform the analysis in the frequency or time domain to evaluate the dynamic response of the structure. Frequency-domain analysis has been used extensively in the offshore oil and gas


369

industry, since it is computationally very efficient and, knowing the wave spectrum of the given site and the RAOs1 of the system considered, it is relatively quick to assess the system response spectrums (Journée and Massie, 2001; Patel, 1989). Frequency-domain methods have also been used for the preliminary design of a number of offshore floating wind turbine support structures (Bulder, 2002; Lee, 2005; Wayman et al., 2006; Collu et al., 2010; Lefebvre and Collu, 2012). Anyway, the linearisation required for frequency domain analysis does not allow for any nonlinear dynamics to be easily accommodated. One example is the RAO: this is defined as the response versus frequency of the system (in the considered d.o.f.) to a wave of unitary amplitude and specified frequency. This concept is often ‘stretched’ with FOWT systems, including nonlinear aerodynamic forces in the analysis of the response, leading to the need for multiple RAO representations for the same d.o.f. (sometimes called pseudo-RAO) in order to present the response with different wind speeds. A frequency approach is also able to represent only the response at regime, and not the transient phase of the dynamic response, which may be critical in the design of a floating wind turbine. This is evident in all current floating HAWT wind turbine design codes (Cordle and Jonkman, 2011), and also for vertical axis wind turbines (VAWTs): the aero-elastic-hydro-servo coupled model of dynamics tend to adopt a time domain approach (Borg et al., 2014). A major contribution to time domain integrated dynamics design codes was made by Jonkman (2007). Jonkman developed a comprehensive simulation tool for the coupled dynamic response of floating HAWTs, and then performed integrated dynamic analysis on an HAWT mounted on a barge-type platform according to the IEC 61400-3 design standard (IEC, 2007). This tool has become integrated into FAST, one of the most widely used offshore HAWT design codes.

11.2.3.2 Maximum inclination angle When considering the requirement on the maximum inclination angle, a dynamic response analysis can provide an estimate of the dynamic oscillation in the pitch/roll d.o.f. For HAWTs this is typically the sum of the ‘static’ angle due to the more or less constant inclining moment illustrated in Eq. [11.3], plus a dynamic angle due to the oscillatory nature of the wave forces on the system. For VAWTs, due to the highly oscillatory nature of the aerodynamic forces (and therefore also the thrust force and the inclining moment), there will be an oscillating response in pitch and roll even without wave forces. In any case, the maximum inclination angle requirement should be revisited, taking into account the maximum dynamic oscillation in the relevant rotational degrees of freedom.

1

Response Amplitude Operators.

370

Offshore Wind Farms

11.2.4

Comment: further aspects to consider

The present approach is illustrating the main requirements and considerations to take into account in a conceptual/preliminary phase, nonetheless it is good practice to try to consider other important aspects as early as possible in the design. Guidelines, recommended practices and classification and certification documents issued by the main certification authorities are a good point of reference also in the early stages of the design (see section Sources of further information). It is important in the early stages to take into consideration also the other phases (not only the operational phases), ie, the construction, transport to operational site, installation and commissioning phases, as they can impose further and more stringent requirements even at the hydrostatic and stability levels. For example, one of the main limitations of a TLP system for FOWT is that the static stability requirement is satisfied by the tensioned mooring system, but this system cannot be used during the transport phase from the port to the operational site. Even in these early stages a rough estimate of the total cost can be derived. It can be simply based on a ‘bill of material’ approach. Only the main materials are considered (ie, steel and/or concrete), and their estimated weight is multiplied by an estimate of the cost per tonne. Even if approximated, it can be quite useful to compare the several configurations of the design space analysed, and narrow it down to the three to five most suitable configurations that will be analysed in more detail in the following design phases.

11.3

Key issues in design of floating offshore wind turbines

In the following sections some of the key challenges that the offshore floating wind turbine industry is facing/will face are explained.

11.3.1

Lack of design integration

Citing from the DNV GL report on the ‘Project FORCE’ (DNV GL, 2013): . the turbine manufacturer designs a turbine optimised to deliver the lowest lifecycle cost possible before releasing technical information to enable the separate design of the support structure. The trouble with this approach is that the design of each element has subtle but significant implications for the design of the other. It may be possible to design a turbine with more advanced features that is, perhaps, slightly more expensive but reduces the loading of the support structure enough to save cost in the steel fabrication and result in a net overall saving. By performing this kind of optimisation exercise on the turbine/support structure system as a whole, any unintended conservatism resulting from isolated design of components can be eliminated e and cost saved.


371

In the same report, it is stated that by adopting an integrated wind turbine support structure approach, a 10% lower cost of energy can be achieved in the short term (5 years) (HAWT). An important observation can be made differentiating HAWT from VAWT. The same report also recognises that: ‘The cost savings are real and achievable with today’s technology: the only barrier is commercial’ and ‘Currently, wind turbines are being procured by developers under separate contracts to the support structures; this is a barrier to the integrated design approach and generally results in non-optimal designs’. The maturity of the offshore HAWT market is in this case an obstacle, and one that will be difficult to overcome. On the other hand, this may be an opportunity for the novel wind turbine concepts emerging for the offshore floating wind turbine market (eg, VAWT), where their relative novelty can allow an integrated design of the wind turbine and the floating support structure from the early design stages.

11.3.2 Oil and gas industry legacy The offshore wind industry is, at the moment, in a position not so different than the oil and gas industry in the 1940se50s, when the first far offshore oil and gas reserves started to be exploited, and novel floating concepts were defined for far from shore and deeper sites. The substantial experience accumulated over the decades, as well as the new research fields developed and investigated to allow the developments of the offshore floating platforms for the oil and gas industry, represent an ideal starting point for the design of floating support structures. Nonetheless, it should always be considered that some of the driving parameters for a typical offshore oil and gas platform and an offshore wind turbine are substantially different. The first important difference is that while the oil and gas platforms are designed considering that these are permanently manned structures, in general an offshore wind turbine system should be designed as an unmanned system. This has a substantial impact on the design criteria, since, for example, the safety factors considered for an oil and gas structure can be over-conservative for a renewable energy device, leading to an over-engineered wind turbine system, that is reflected ultimately in a higher final cost for the energy produced. A second important difference is linked to the number of platforms. A floating support structure designed for the oil and gas industry is typically a bespoke, one-off design. The oil field is investigated and characterised in detail and the relative oil rig is designed for that particular site, also taking into account the environmental loads determined on a case-by-case basis, and the relative design standards are based on this bespoke process. On the contrary, the new standards developed for offshore floating wind turbines (eg, (DNV, 2013), see also Section Sources of further information) are based on the different concept of ‘environmental class’. Basically, the wind turbine designed and certified with this approach is not suitable only for a specific site, but for a ‘class’ of sites, or region (DNV, 2013), where the metocean conditions can be considered

372

Offshore Wind Farms

substantially similar. This encourages a mass production approach, needed also even if considering only a single wind farm. The ‘London array’ wind farm, even if it utilises fixed to seabed wind turbines, can be considered as a precursor of the future offshore floating wind farms. It became fully operational in April 2013, and with a total rated electrical power of 630 MW, it consists of 175 wind turbines on an area of around 100 km2. The ‘mass production’ advantage is a much needed opportunity to lower the costs of offshore wind energy, and the floating wind farms to be developed will certainly need to exploit it.

11.3.3

Numerical modelling limits

During the design of a floating wind turbine, ideally the highest fidelity numerical models are used to assess and optimise design aspects. However, current computational resources available to design engineers limit the use of such numerical models, such as integrated computational fluid dynamics and finite element analysis, to the final stages of the design, to investigate very specific operating conditions. Hence reduced-order engineering models are more convenient for carrying out the preliminary design and optimisation studies. The use of such engineering models implies a number of assumptions in the formulation of the numerical model, and it is important to assess the validity of these assumptions when simulating a floating wind turbine in the offshore environment. Environmental loads arise from incident wind inflow, a large range of surface waves, sea currents and tides, and ice loading in colder regions, providing a challenge to adequately represent these natural phenomena in engineering numerical models. Currently, the majority of state-of-the-art floating wind turbine design tools utilise blade element momentum-based aerodynamic models to derive the loads from the interaction of the turbine with the incident wind inflow. This quasi-steady approach has been modified to include secondary effects, such as dynamic stall, tip losses and skewed flow to provide significantly better agreement with experimental data. Whilst this has been sufficient for HAWTs, supported by design guidelines (DNV, 2013), in the case of floating VAWTs this is much more uncertain. The inherent unsteady nature of VAWT aerodynamics leads to quasi-steady models insufficiently predicting instantaneous blade forces, even for onshore turbines, that are crucial for turbine structural design, as illustrated by Ferreira et al. (2014). Despite the slight decrease in computational efficiency, transitioning to higher-order engineering models, such as vortex models, would generate more cost-effective designs as there would less be uncertainty during structural design (Borg et al., 2014b). Hydrodynamic loading regimes are heavily dependent on the floating support structure geometry, relative size with respect to the incident ocean waves and operating sea states. State-of-the-art design tools make use of a combination of the potential flow-based Cummins equation and Morison equation (Borg and Collu, 2014). The combination of these two models, theoretical and semi-empirical, respectively, allows design engineers to model a wide array of low- and high-volume bodies that would constitute the floating support structure. Higher-order engineering models have also been developed to investigate particular situations with extreme conditions of natural


373

phenomena, such as the nonlinear impulse theory by Sclavounos (2012) and domain decomposition strategies for the hydrodynamic field by Paulsen et al. (2014). The structural flexibility of the various components can have a critical impact on numerical simulations, and state-of-the-art design tools mainly utilise reduced-order FEM, although the multibody formulation is more predominant due to improved computational efficiency and sufficient numerical accuracy (Borg et al., 2014a). Likewise, state-of-the-art design tools implement the multibody formulation to the model mooring system, including hydrodynamic loading through the Morison equation mentioned above (Borg et al., 2014). In the quest to develop significantly more structurally efficient and cost-effective floating wind turbine designs, the verification and validation2 of design tools are of critical importance to reduce simulation uncertainties and improve design guidelines. Ideally, experimental data from a full-scale floating wind turbine would be used to validate such design tools, however the significantly large financial backing and commercial issues render this scenario highly unlikely. In light of this, design tool verification through code-to-code comparisons based on reference systems has been done extensively though the International Energy Agency Task 23 (Jonkman et al., 2010) for floating HAWTs and initiated for floating VAWTs (Borg et al., 2014). Supporting this work, extensive model-scale tests in controlled conditions in ocean basins provide validation and insights into unforeseen dynamics not experienced in numerical simulations, most notably those carried out in the DeepCwind initiative (Robertson et al., 2013).

11.3.4 Floating platform impact on turbine loadings and control 11.3.4.1 Loads Floating support structures obviously provide very soft foundations when compared to fixed offshore or onshore foundations, thereby significantly modifying the system Eigen modes and frequencies, which subsequently influences the dynamic response and experienced loads due to environmental excitations. Jonkman and Matha (2010) carried out a systematic study of the three floating HAWT concepts (a barge, spar and TLP), finding that for all concepts the turbine experienced increased fatigue and extreme loads. The immediate implication of this is that the turbine components need to be strengthened when translating an onshore turbine design for offshore use. Borg and Collu (2015b) investigated in detail the influence of platform motion on VAWT aerodynamic loads in the frequency domain, concluding that whilst the peak aerodynamic forces did not substantially change, there were significant order of 2

Whilst verification and validation are sometimes used interchangeably, verification is the process of verifying that the numerical model has been correctly implemented (‘solving the equations right’), and validation is confirming that the implemented numerical model actually represents what occurs in reality (‘solving the right equations’).

374

Offshore Wind Farms

magnitude increases across the frequency range of platform motion. This is mostly relevant to the fatigue assessment of various turbine components.

11.3.4.2 Control The main aim of the turbine control system is to reduce loads and maximise power production across the operating envelope of the floating wind turbine. Additional considerations need to be taken for FOWT, in particular the situation where negative aerodynamic damping occurs, that is, the controller actually increases turbine loads and platform motion (Larsen and Hanson, 2007). This arises as the platform pitches fore and aft, the turbine experiences variations in incident wind inflow and the controller attempts to correct for this cyclically. This is remedied by augmenting the controller parameters to have a slower response to changes in wind speed. Control of VAWTs differs somewhat to that of HAWTs; VAWTs usually have fixed-pitch blades and control is usually achieved solely through generator torque management. Whilst for HAWTs the floating platform inherent roll restoring stiffness accommodates the torque generated by the turbine, it is a different scenario for VAWTs, whereby the mooring system has to sufficiently accommodate the generator torque in the yaw d.o.f. In combination with the relatively low yaw mooring stiffness characterising most mooring system configurations, it can be challenging to adapt a suitable controller to maintain the specified control strategy (Merz and Svendsen, 2013).

11.3.5

Costs of floating wind turbines

As described in chapter “Remote sensing technologies for measuring offshore wind”, the costs of offshore wind turbines originate from different sources. Taking a lifecycle cost perspective, the costs can be divided into capital expenditures (CAPEX), operating expenditures (OPEX), and decommissioning expenditures (DECEX). CAPEX consists of upfront investment costs related to site assessment and development; engineering, procurement and manufacturing of turbines, support structures and grid infrastructure; and costs related to transport and installation of the units on site. OPEX includes all costs related to operation of the offshore wind farm; including maintenance and repairs; monitoring and control of the farm; and maintaining backup power capacity. DECEX relates to costs incurred at the end of the operational life, removing units from the offshore site and scrapping. According to Myhr et al. (2014), the OPEX for fixed and floating wind turbine farms (under the same conditions) is, respectively, (2014 EUR) 115 and 131 kV/MW, while the CAPEX is similar for fixed (monopile and jacket options considered) and floating wind farms, at around 3.5e3.8 mV/ MW, with the exception of the floating option ‘WindFloat’, with a cost around 4.6 mV/MW.3 3

It has to be said that the floating wind turbine ‘WindFloat’ was the first prototype of its kind, a one-off production, and therefore the CAPEX/MW reported here cannot be considered to represent the likely CAPEX/MW of the same configuration at a commercial mature stage.


375

These costs dictate the final price of electricity generated by the wind farm; typically called levelised cost of energy (LCOE). As there have not yet been any floating wind farms installed, predicting OPEX costs is not trivial and it involves some uncertainties. Whilst some costs can be reasonably estimated through experiences from bottom-fixed wind farms, differences between the two types of wind turbine structures lead to different operational practices. There have been relatively few studies investigating LCOE of floating wind turbines and cost drivers for this type of offshore wind turbine (Myhr et al., 2014; Paulsen et al., 2015; Laura and Vicente, 2014). Myhr et al. (2014) carried out a comparative study of the LCOE for a number of fixed and floating offshore wind turbines, estimating that the LCOE can be within the range of V130 to V180/MWh depending on wind farm size, distance to shore, amongst a number of other factors. One main outcome from this study is that the LCOE is heavily dependent on a large number of factors, and likely cannot be generalised for the floating wind industry but is very project-specific.

11.4

Summary: case study

11.4.1 Context and site description To illustrate the design stages detailed above, a case study on the preliminary design of a floating support structure for a 5-MW HAWT carried out by Lefebvre and Collu (2012) is presented here. The aim of this study was to design a floating support structure for the NREL 5-MW reference turbine (Jonkman 2010) situated in the Dogger Bank site in the North Sea, one of the offshore Round 3 sites proposed by the UK government, with site characteristics given in Table 11.1.

Table 11.1

Dogger bank site characteristics

Water depth (m)

18e63 (40 assumed)

Distance from shore (km)

192.7 JONSWAP wave spectrum parameters Operational

Survival

Hs (m)

4.928

10.0

Philips constant (a)

0.008074

0.008110

JONSWAP constant (b)

3.3

3.3

Peak frequency (rad/s)

0.6283

0.4488

376

11.4.2

Offshore Wind Farms

Development of floating support structure

An assessment of different floating platforms was undertaken, namely two barges, a spar, TLP, semi-submersible and Tri-floater, by carrying out preliminary sizing and cost estimates with the Tri-floater emerging as the most favourable configuration. A maximum allowable inclination angle of 10 degree was assumed following the work of Van Hees et al. (2002) to derive the minimum required pitch restoring stiffness considering the maximum overturning induced by the operating turbine (Eq. [11.5]). The next step taken was to carry out a detailed analysis of the Tri-floater structure, Fig. 11.2, investigating static and dynamic stability and potential mooring configurations. During this process the shell thickness was reduced from 20 to 15 mm by the inclusion of stiffeners, depicted in Fig. 11.3, as the dynamic analysis revealed that the original thickness resulted in an unnecessarily more costly structure. In

Wind turbine (1)

Tower (2)

Tower support (7) Main horizontal beam (5)

Column (3)

Footplate (4)

Bracings (6)

Figure 11.2 Structural components of the Tri-floater configuration (Lefebvre and Collu, 2012).


Figure 11.3 Illustration of column internal structure (Lefebvre and Collu, 2012).

377

378

Offshore Wind Farms

this stage the actual diameters of the bracings were also derived, depicted in Fig. 11.4, such that the steel yield strength was not exceeded considering the relevant safety factors. In conjunction with the structural assessment, the hydrostatic stability carried out in the preliminary assessment was reiterated, now considering the refined structural composition of the support structure. Both intact and damaged scenarios were considered, although in the latter case it was assumed that the turbine would not be operational in such a scenario and hence a maximum allowable inclination angle of 20 degree was prescribed. In any case, the stability simulations produced inclination angles of 9 and 14.5 degrees for the intact and damage scenarios, respectively, well within the prescribed allowable inclination angles. Likewise a hydrodynamic stability assessment was carried out, whereby numerical simulations were performed to derived the system RAOs and dynamic response in the operational and survival sea states. During initial hydrodynamic stability analyses it was seen that the system heave natural frequency was well within the range of most energetic wave spectrum frequencies, and to reduce heave motion, footplates were added to the bottom of each support column to augment the heave natural frequency, depicted in Fig. 11.2. Following the inclusion of the footplates, the system natural frequencies were found to be well outside the most energetic wave spectrum frequencies and that the maximum allowable inclination angles were not exceeded for either operational or survival sea states. Finally, due to the relatively shallow water depth, a taut mooring system was considered. Both a three-line and nine-line mooring system were considered, and using the hydrodynamic analysis, the line length, pre-tension and stiffness were tuned such that the above requirements were still met, in addition to the requirement that line elongation must not exceed 10%. As both systems satisfied the requirements, the three-line system was adopted as the nine-line system was significantly more costly. Ø6 × 0.03 Ø1 × 0

.02

Z

.02 ×0 Ø1 Ø1.5 × 0.02

Y X Z

Z Y

X

Ø1.5 × 0.02

Ø 1

×

0.

02

Y X

×0 Ø1

Figure 11.4 Bracings dimensions (Lefebvre and Collu, 2012).

.02

Z Y X


11.5

379

Future trends

11.5.1 Size 10 MW and beyond The size of the offshore wind turbine has been constantly increasing over the years, as, in general, the higher the rated power of the wind turbine, the lower is the final LCOE (Ashuri, 2012). While there are already commercially available wind turbines with a rated power of 8 MW,4 wind turbines with a rated wind power equal to 10e12 MW and beyond are already being studied and developed (INNWIND FP7 project,5 HiPRWind FP7 project6). This trend is certainly going to continue, with some of the most current ambitious projects looking at 15e20 MW rated power offshore wind turbines, even if for HAWT there are some indications that it will be more and more difficult to upscale these systems (Tjiu et al., 2015).

11.5.2 Vertical and horizontal axis wind turbines Whilst HAWTs have been the preferred configuration for onshore wind turbines, due to the significantly different conditions experienced in the floating offshore environment, other wind turbine configurations may be more advantageous from both technical and economic aspects. There is a re-emerging interest in VAWTs for floating foundation applications, due to several potential advantages. This led to a number of studies being performed for this class of turbine by different researchers (Shires, 2013). Nonetheless, there have been only a few research attempts to quantitatively compare HAWTs and VAWTs for floating offshore applications. Borg and Collu (2015) focused on comparing the static and dynamic responses of floating HAWT and VAWT systems. It was shown how a VAWT configuration can be characterised by a smaller inclining moment, a lower wind turbine mass and a lower wind turbine G, all factors that can be exploited to lower the cost of the support structure, and therefore the final LCOE. Nonetheless, the highly oscillatory nature of the aerodynamic forces acting on a VAWT system are also highlighted, bringing different challenges to the design of these wind turbines. Referring to Section 11.3.1, VAWT wind turbines, even though they are commercially less mature than HAWTs, still have the opportunity to adopt an integrated design approach (wind turbine þ support structure) from the early phases, ensuring that the optimisation process can investigate solutions modifying both the support structure and the wind turbine design. This should allow a substantial reduction of the LCOE.

11.5.3 Multipurpose platform integration In some cases, future floating wind turbines may be integrated with other energyharvesting devices, such as wave and tidal, as well as other ocean uses such as 4 5 6

http://www.mhivestasoffshore.com/Products-and-services/The-Turbines/V164. http://www.innwind.eu/. http://www.hiprwind.eu/.

380

Offshore Wind Farms

aquaculture. The main perceived benefit of integrating such different technologies is that infrastructure such as electrical grid connections, power conversion equipment and floating support structures can be shared, thereby reducing the cost of energy. There have been a number of initiatives to investigate such integrated designs; H2OCEAN (www.h2ocean-project.eu), TROPOS (www.troposplatform.eu) and MERMAID (www.mermaidproject.eu) projects assessed the feasibility of multipurpose platforms; the MARINA project (www.marina-platform.info), Aubault et al. (2009), Roddier and co-workers (Peiffer et al., 2011), Floating Power Plant (www. floatingpowerplant.com) and Borg et al. (2013) are several examples of investigations into combined windewave energy-harvesting devices; and the SKWID (www.modec.com/fps/skwid) concept combined wind and tidal current energy extraction. The main challenge is carrying out an integrated engineering and logistical design of such complex systems to provide competitive products.

11.5.4

Toward an integrated multi-disciplinary design and optimisation

Wind turbines, and especially offshore floating wind turbines, can be considered complex systems, similar in complexity to aeroplanes, cars and ships. The design, analysis and optimisation of these complex systems require a multi-disciplinary approach. Multidisciplinary design, analysis and optimisation (MDAO) is an engineering field focussing on the use of numerical tools for the design of systems involving a number of disciplines or sub-systems. The main reason for MDAO is that such systems cannot be optimally designed by designing, analysing and optimising separately the several sub-systems, but it is necessary to take into account their interactions (Martins and Lambe, 2013). MDAO was firstly applied for aircraft wing design, due to the strong coupling between the aerodynamics, structural, and control aspects of the problem. It has then been extended to complete aircraft and other engineering systems (eg, bridges, buildings, railway cars, automobiles, spacecraft, etc.) (Martins and Lambe, 2013). In recent years, the first studies to apply an MDAO approach to wind turbines have been presented (Fuglsang and Madsen, 1999; Fuglsang et al., 2002; Kenway and Martins, 2008), for onshore and offshore fixed foundation wind turbine systems. Furthermore, the first open-source programs are becoming available. FUSED-Wind (http://www.fusedwind.org/) is: ‘an open-source framework for multi-disciplinary optimisation and analysis (MDAO) of wind energy systems, developed jointly by DTU and NREL’. NREL-WISDEM (https://nwtc.nrel.gov/WISDEM) is ‘built on top of the FUSED-Wind software framework and includes wrappers for a full suite of wind plant models including turbine aerodynamics, component structural analysis, component costs, plant balance of station costs, plant operations and maintenance costs, financial models, wind plant layouts, and wind turbine aeroelastic simulations’.


381

Multidisciplinary design, analysis, and optimisation approaches will play a fundamental role to lower the costs of offshore (fixed and floating) wind structures in the near future.

Nomenclature B C55,moor C55,tot CP(env) d.o.f. F FB Fenv Fmoor,V G M FOWT Ixx Iyy MI MLA MR,roll MR,pitch RAOi TLP V X Y Z zCP(env) zMLA q r

Centre of buoyancy Rotational stiffness coefficient (pitch) due to the mooring stiffness (Nm/rad) Total rotational stiffness (pitch) (Nm/rad) Centre of pressure of environmental forces Degree of freedom Centre of flotation Buoyancy force (N) Sum of environmental forces (wind and current) (N) Vertical component of the total force due to the mooring system (N) Centre of gravity Total mass (kg) Floating Offshore Wind Turbine Second moment of the waterplane area in roll (m4) Second moment of the waterplane area in pitch (m4) Inclining moment (Nm) Centre of mooring force line action Restoring moment in roll (Nm) Restoring moment in pitch (Nm) Response Amplitude Operator (m/m), i-th d.o.f. (i ¼ 1,2,3) (m/m), Response Amplitude Operator, i-th d.o.f. (i ¼ 4,5,6) (deg/m) Tension Leg Platform Displaced volume (m3) Horizontal axis of reference (m) Lateral axis of reference (m) Vertical axis of reference (m) Vertical coordinate of CP(env) (m) Vertical coordinate of MLA (m) Angle of inclination (deg) Seawater density (kg/m3)

Sources of further information As previously mentioned, specific guidelines, recommended practices and certification documents are starting to be issued by the main classification and certification societies, and will probably be updated regularly. DNV GL, ABS and BV documents are freely available. There are several international peer-reviewed conferences with sessions specifically focused on FOWT, the most important being the conferences organised by the

382

Offshore Wind Farms

European Wind Energy Association (EWEC and EWEA offshore), the ASME International Conference on Ocean, Offshore and Arctic Engineering (OMAE), and the conference organised by the International Society of Offshore and Polar Engineers (ISOPE). At the moment there are no books specifically dedicated to offshore floating wind turbines, but there are a number of books that can be taken as references for offshore floating wind turbine support structures, and here only a few examples are given (with the caveat explained in Section 11.3.2): ‘Handbook of offshore engineering’ (Chakrabarti, 2005), ‘Offshore Hydromechanics’ (Journée and Massie, 2001), ‘Dynamics of Offshore Structures’ (Patel, 1989).

References ABS, 2013. Guide for Building and Classing Floating Offshore Wind Turbine Installations. Ashuri, T., 2012. Beyond Classical Upscaling: Integrated Aeroservoelastic Design and Optimization of Large Offshore Wind Turbines. Delft University of Technology. Available at: http:// mdolab.engin.umich.edu/content/beyond-classical-upscaling-integrated-aeroservoelasticdesign-and-optimization-large (accessed 01.12.14.). Aubault, A., Cermelli, C., Roddier, D., 2009. WindFloat: a floating foundation for offshore wind turbines part III: structural analysis. Engineering 1e8. Blue, H., 2004a. Blue H e Concept. Available at: http://www.bluehgroup.com/concept/ index.php (accessed 25.02.15.). Blue, H., 2004b. Blue H e Products e Phase 1. Available at: http://www.bluehgroup.com/ product/phase-1.php. Borg, M., Wang, K., et al., 2014. A comparison of two coupled model of dynamics for offshore floating vertical axis wind turbines (VAWT). In: Proceedings of the ASME 2014 33rd International Conference on Ocean, Offshore and Arctic Engineering OMAE 2014. ASME, San Francisco, CA. Borg, M., Collu, M., 2015. A comparison between the dynamics of horizontal and vertical axis offshore floating wind turbines. Philosophical Transactions of the Royal Society A 373 (2035). Borg, M., Collu, M., 2015b. Frequency-domain characteristics of aerodynamic loads of offshore floating vertical axis wind turbines. Appl. Energy 155, 629e636. Borg, M., Collu, M., 2014. Offshore floating vertical axis wind turbines, dynamics modelling state of the art. Part III: hydrodynamics and coupled modelling approaches. Renewable and Sustainable Energy Reviews 46, 296e310. Available at: http://www.sciencedirect.com/ science/article/pii/S1364032114009253 (accessed 01.12.14.). Borg, M., Collu, M., Brennan, F.P., 2013. Use of a wave energy converter as a motion suppression device for floating wind turbines. Energy Procedia 35, 223e233. Available at: https://nbl.sintef.no/project/DeepWind 2013/Deepwind presentations2013/E/Borg.M._ Cranfield Univ.pdf (accessed 18.11.13.). Borg, M., Collu, M., Kolios, A., 2014a. Offshore floating vertical axis wind turbines, dynamics modelling state of the art. Part II: mooring line and structural dynamics. Renewable and Sustainable Energy Reviews 39, 1226e1234. Available at: http://www.sciencedirect.com/ science/article/pii/S1364032114005747 (accessed 03.11.14.).


383

Borg, M., Shires, A., Collu, M., 2014b. Offshore floating vertical axis wind turbines, dynamics modelling state of the art. Part I: aerodynamics. Renewable and Sustainable Energy Reviews 39, 1214e1225. Available at: http://www.sciencedirect.com/science/article/pii/ S1364032114005486 (accessed 01.12.14.). Bulder, B.H., 2002. Study to Feasibility of and Boundary Conditions for Floating Offshore Wind Turbines (Studie naar haalbaarheid van en randvoorwaarden voor drijvende offshore windturbines). BV, 2010. Classification and Certification of Floating Offshore Wind Turbines, Bureau Veritas. Chakrabarti, S.K., 2005. Handbook of Offshore Engineering. Structure, I, pp. 2005e2006. Available at: http://books.google.com/books?hl¼en&lr¼&id¼Snbuzun9LUQC&oi¼ fnd&pg¼PP2&dq¼HANDBOOKþOFþOFFSHOREþENGINEERING&ots¼vceQzo3 Fnx&sig¼LgsOQlU5AvXIOAHYJ8F2k4SV2To. Collu, M., et al., 2010. A comparison between the preliminary design studies of a fixed and a floating support structure for a 5 MW offshore wind turbine in the North Sea. In: RINA Marine Renewable and Offshore Wind Energy. Cordle, A., Jonkman, J.M., 2011. State of the art in floating wind turbine design tools. In: ISOPE 2011 Conference, pp. 367e374. DNV, 2013. DNV-OS-J103 Design of Floating Wind Turbine Structures. DNV GL, 2013. PROJECT FORCE e Offshore Wind Cost Reduction Through Integrated Design. Driscoll, F., et al., 2015. Validation of a FAST model of the statoil hywind demo floating wind turbine. In: Proceedings of the ASME 2015 34th International Conference on Ocean, Offshore and Arctic Engineering. St. John’s, Newfoundland, Canada, pp. 1e10. Ferreira, C.S., et al., 2014. Comparison of aerodynamic models for vertical Axis wind turbines. Journal of Physics: Conference Series 524 (1), 012125. Available at: http://stacks.iop.org/ 1742-6596/524/i¼1/a¼012125 (accessed 01.12.14.). Fuglsang, P., Bak, C., Schepers, J., 2002. Site-specific design optimization of wind turbines. Wind Energy 5, 261e279. Available at: http://onlinelibrary.wiley.com/doi/10.1002/we.61/ abstract (accessed 01.12.14.). Fuglsang, P., Madsen, H., 1999. Optimization method for wind turbine rotors. Journal of Wind Engineering and Industrial Aerodynamics 80 (1e2), 191e206. Available at: http://www. sciencedirect.com/science/article/pii/S0167610598001913 (accessed 01.12.14.). Van Hees, M., et al., 2002. Study of Feasibility of and Boundary Conditions for a Floating Offshore Wind Turbines. IEC, 2007. IEC 61400-3, Wind Turbines e Part 3: Design Requirements for Offshore Wind Turbines. ITTC, 2014. ITTC Wiki. Available at: http://www.ittcwiki.org/doku.php. Jonkman, J., 2010. Definition of the Floating System for Phase IV of OC3. Available at: http:// www.nrel.gov/docs/fy10osti/47535.pdf (accessed 02.12.14.). Jonkman, J.M., 2007. Dynamics Modeling and Loads Analysis of an Offshore Floating Wind Turbine. ProQuest. Available at: http://books.google.com/books?hl¼en&lr¼&id¼66F_ VT7Px-kC&oi¼fnd&pg¼PA1&dq¼DynamicsþModelingþandþLoadsþAnalysisþofþ anþOffshoreþFloatingþWindþTurbine&ots¼N9lpZfiBw1&sig¼0lbYYU9zsz1HrcfoCVBt3leCYFs (accessed 10.02.12.). Jonkman, J.M., et al., 2010. Offshore code comparison collaboration within IEA wind task 23 : phase IV results regarding floating wind turbine modeling. In: EWEC. Jonkman, J.M., Matha, D., 2010. Dynamics of offshore floating wind turbines d analysis of three concepts y. Wind Energy.

384

Offshore Wind Farms

Journée, J., Massie, W.W., 2001. Offshore Hydromechanics. TU Delft, 2000. Kenway, G., Martins, J., 2008. Aerostructural shape optimization of wind turbine blades considering site-specific winds. In: Proceedings of 12th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference. Available at: http://arc.aiaa.org/doi/pdf/10.2514/6. 2008-6025 (accessed 01.12.14.). Larsen, T., Hanson, T., 2007. A method to avoid negative damped low frequent tower vibrations for a floating, pitch controlled wind turbine. Journal of Physics: Conference Series 75 (1), 012073. Available at: http://iopscience.iop.org/1742-6596/75/1/012073 (accessed 01.12.14.). Laura, C.-S., Vicente, D.-C., 2014. Life-cycle cost analysis of floating offshore wind farms. Renewable Energy 66, 41e48. Available at: http://www.sciencedirect.com/science/article/ pii/S0960148113006642. Lee, K.H., 2005. Responses of Floating Wind Turbines to Wind and Wave Excitation. Massachusetts Institute of Technology. Lefebvre, S., Collu, M., 2012. Preliminary design of a floating support structure for a 5 MW offshore wind turbine. Ocean Engineering 40, 15e26. Martins, J., Lambe, A., 2013. Multidisciplinary design optimization: a survey of architectures. AIAA Journal 51 (9). Available at: http://arc.aiaa.org/doi/abs/10.2514/1.J051895 (accessed 01.12.14.). Merz, K.O., Svendsen, H.G., 2013. A control algorithm for the deepwind floating vertical-axis wind turbine. Journal of Renewable and Sustainable Energy 5 (6), 063136. Available at: http://scitation.aip.org/content/aip/journal/jrse/5/6/10.1063/1.4854675 (accessed 01.12.14.). Myhr, A., et al., 2014. Levelised cost of energy for offshore floating wind turbines in a life cycle perspective. Renewable Energy 66, 714e728. Available at: http://linkinghub.elsevier.com/ retrieve/pii/S0960148114000469. Patel, M.H., 1989. Dynamics of Offshore Structures. Butterworth-Heinemann Ltd. Paulsen, B.T., et al., 2014. Forcing of a bottom-mounted circular cylinder by steep regular water waves at finite depth. Journal of Fluid Mechanics 755, 1e34. Available at: http://journals. cambridge.org/abstract_S0022112014003863 (accessed 01.12.14.). Paulsen, U., et al., 2015. Outcomes of the DeepWind conceptual design. In: 12th Deep Sea Offshore Wind R&D Conference. Trondheim, Norway. Peiffer, A., Roddier, D., Aubault, A., 2011. Design of a point absorber inside the WindFloat structure. In: Ocean Space Utilization; Ocean Renewable Energy, Volume 5. ASME, pp. 247e255. Available at: http://proceedings.asmedigitalcollection.asme.org/proceeding. aspx?articleid¼1624862 (accessed 01.12.14.). Principle Power, 2012. WindFloat Prototype. Available at: http://www.principlepowerinc.com/ products/windfloat.html. Robertson, A., et al., 2013. Summary of conclusions and recommendations drawn from the DeepCWind scaled floating offshore wind system test campaign. In: ASME 2013 32nd International Conference on Ocean, Offshore and Arctic Engineering. ASME, Nantes, FR. Sclavounos, P.D., 2012. Nonlinear impulse of ocean waves on floating bodies. Journal of Fluid Mechanics 697, 316e335. Available at: http://journals.cambridge.org/abstract_ S0022112012000687 (accessed 01.12.14.). Shires, A., 2013. Design optimisation of an offshore vertical axis wind turbine. Energy 166 (EN1), 7e18. Available at: http://www.icevirtuallibrary.com/content/article/10.1680/ener. 12.00007 (accessed 29.12.13.). Statoil, 2010. Hywind Demo. Available at: http://www.statoil.com/en/TechnologyInnovation/ NewEnergy/RenewablePowerProduction/Offshore/Hywind/Pages/HywindPuttingWind PowerToTheTest.aspx?redirectShortUrl¼http://www.statoil.com/hywind.


385

Tjiu, W., et al., 2015. Darrieus vertical axis wind turbine for power generation II: challenges in HAWT and the opportunity of multi-megawatt darrieus VAWT development. Renewable Energy 75, 560e571. Available at: http://www.sciencedirect.com/science/article/pii/ S0960148114006661 (accessed 01.12.14.). Wayman, E.N., et al., 2006. Coupled dynamic modeling of floating wind turbine systems. In: Offshore Technology Conference. Houston, Texas.


Part Three Integration of wind farms into power grids


Offshore wind farm arrays O. Anaya-Lara University of Strathclyde, Glasgow, United Kingdom

12

This chapter presents the reader with configurations and technologies used in an offshore wind farm electrical collector. It also introduces the industry’s current practices for their design and implementation. Array layouts and electrical collectors are designed on a site-specific basis to achieve a good balance between electrical losses and wake effects. The material presented in the chapter is oriented towards the electrical aspects once a layout has been identified to ensure maximising the energy yield from the wind. Issues addressed include electrical performance, wind farm capacity, wind turbine generator technology and voltage levels among others.

12.1

Fundamentals of offshore wind farm arrays

An offshore collector system consists of underwater cabling connecting all wind turbines within the wind farm to other turbines in an array structure, allowing the power generated by each unit to be exported to the substation(s). It can also be referred to as an interturbine/interarray system. This cabling e operating at a voltage of 33 kV in the UK and 20 kV across Europe e channels the energy generated by each turbine to the nearest substation. One important feature of wind farms is that wind turbines are required to be spaced a certain number of rotor diameters away from one another due to the wake effect phenomenon. This often leads to large sites spanning tens of kilometres in size; such large sites will require substantial cabling distances to connect all turbines to the farm substation(s). As collector system cabling lengths span across the entire site, the losses associated with the collector system result in the largest portion of total wind farm losses. Therefore, it is extremely important to minimize losses introduced by the collector system as reduction of these losses will significantly reduce the overall electrical power losses across the farm, ultimately leading to a more efficient farm and more power generated. This increase in generation will, in the long term, lead to an increase in returns produced for the site owners. These losses may be reduced by simply rearranging the interturbine cabling arrangement. A wind farm collector system is designed after all turbines and substation(s) have been suitably located. When designing a wind farm structure, engineers must go through many stages, one of which includes the design of the collector system. This process is a complex optimization task in which engineers try to strike a balance between minimising the collector cable losses whilst also minimising the development costs involving the submarine cabling. Offshore Wind Farms. http://dx.doi.org/10.1016/B978-0-08-100779-2.00012-X Copyright © 2016 Elsevier Ltd. All rights reserved.

390

Offshore Wind Farms

12.2

Design considerations

The electrical array system has to be designed to meet the following general requirements [1]: • • • • • • • • • •

Health and Safety: the system must be designed in a way that minimizes the risk of adversely affecting the health and safety of personnel and the public, and must comply with all relevant H&S legislation. Compliance with all relevant codes e national grid codes, distribution codes, STC (System Operator-Transmission Owner Code) e and standards, eg, IEC, IEEE. Export capacity to be at least 100% of wind farm output. Minimise CAPEX: material and installation costs. Minimise OPEX: losses and maintenance requirements. Maximise availability and reliability. Low environmental impact (low life cycle carbon emissions, recyclable, decommissionable). Strong supply chain of components. Adaptable to different wind turbine generators (WTGs) and wind farm sizes. Redundancy against a single fault to provide power to auxiliary demand.

In addition, bilateral agreements may dictate project-specific requirements depending on the location of a wind farm or grid connection point.

12.3 12.3.1

Main electrical components Wind turbines

Currently installed offshore wind turbines are adapted from standard onshore wind turbine designs with significant upgrades to account for sea conditions. These modifications include strengthening the tower to handle the added loading from waves, along with pressurized nacelles and environmental control to keep corrosive sea spray away from critical drive train and electrical components [2]. Offshore turbine power capacity is in general greater than standard onshore wind turbines, currently ranging from 2 to 5 MW (Fig. 12.1). The current generation of offshore wind turbines is typically three-bladed horizontal-axis, yaw-controlled, active blade-pitch-to-feather controlled, upwind rotors, which are nominally 80 m to approximately 130 m in diameter [3]. Offshore wind turbines are generally larger because there are fewer constraints on component and assembly equipment transportation, which limit land-based machine size. In addition, larger turbines can extract more total energy for a given project site area than smaller turbines [4]. A critical issue in developing very large wind turbines is that the physical scaling laws do not allow some of the components to be increased in size without a change in the fundamental technology. In onshore wind turbines, the drive train is typically designed around a modular, fixed-ratio, three-stage, gearbox with planetary stages on the low-speed side and helical stages on the high-speed side. Offshore towers are shorter than onshore towers for

Offshore wind farm arrays

Project Turbine type

391

Scroby sands Type C

RØdsand II Type D

Robin rigg Type C

Hub height

60 m

68 m

80 m

90 m

Rotor diameter

80 m

93 m

90 m

126 m/116 m

Capacity

2 MW

2.3 MW

3 MW

5 MW

90

60

12

Number of turbines

30

Alpha ventus Type C, D

Figure 12.1 Offshore wind turbine development [3].

the same output because wind shear is lower offshore, which reduces the energy capture potential of increasing tower height [4].

12.3.2 Offshore substations The main purpose of a substation is to step up the voltage from the collector level (eg, 33 kV) to a higher level suitable to export the electricity produced offshore to the grid, normally 132 kV. The voltage is required to be stepped up to reduced losses that occur with large transmission distances as offshore farms are often located great distances from shore. Substations are required for large offshore projects (>100 MW) or when farms are situated at distances greater than 15 km from shore. An offshore substation costs around 7% of the total cost of a wind farm. For a 500-MW offshore wind farm the offshore substation costs around £80 million and weighs up to 2000 tonnes [2,5]. Fig. 12.2 shows the schematic of an offshore substation [5,6]. One substation can typically support up to 500 MW of output from a wind turbine array. With an increase in wind farm size, the number of substations may increase too. An offshore substation is typically delivered as one component after contract by a supplier and weighs generally from 1800 to 2200 tonnes (for a 500-MW wind farm). The platform level is generally 25e30 m above sea level and has an area of 800e1200 m2. These offshore substations are generally not service-based but still have a small workshop available inside. The major components of an offshore substation are divided into three categories [5]: • • •

Components related to electrical systems; Components related to facilities; Components related to structure.

392

Offshore Wind Farms

Power transformer

Module E

Heli lift

Module D

Module A Module B

Module C Auxiliary transformer

Platform deck

132-kV cable transmission

Cable deck

Module A

33-kV cable transmission

Auxiliary transformer

Fire Module B Power transformer

Control

Standby diesel

33 kV

LV room

132 kV

Module C

Figure 12.2 Layout of an offshore substation [5,6].

The main components related to the electrical systems are as follows: • • •

The major component of the offshore substation is the transformer. The transformer steps up the voltage according to requirements for optimum power transfer and minimization of losses. A back-up diesel generator is also made available on the substation in case of loss of power via the export cable. Switchgear is part of an offshore substation to discriminate between the export cables and array cables.


• • •

393

If the onwards transmission is HVDC then converters are also installed in the offshore substation. Reactive power compensation equipment is used to provide optimum reactive power compensation required for the maximum power transfer to the onshore grid (eg, bank of reactors, capacitors, SVCs or STATCOMs) The substation is properly earthed to provide power safety in case of safety hazard or short circuit.

A typical substation can support about 500 MW of input from numerous wind turbines. Many farms require more than one substation to export the power generated due to the scale of the farm. However, if it is financially viable, farms will often have more than one substation to increase the farm’s export security. The world’s current largest operational wind site, London Array Phase 1, has two identical substations due to the scale of the farm.

12.3.2.1 Transformers As already mentioned, the main task of offshore substations is to convert the interarray voltage to a higher magnitude to reduce transmission losses. Gas (SF6)-insulated transformers are specifically designed to be non-flammable; this is an important feature as offshore substations are located great distances from shore and in normal circumstances unmanned, hence any signs of fire would not be able to be quickly dealt with, potentially leading to a catastrophic outcome. When maintenance is required on the substation’s transformers this can be easily and safely carried out as all live parts are contained within metal structures that are grounded.

12.3.2.2 Switchgear Switchgears are used to control, protect and isolate the array of wind turbines. Wind turbines are required to be isolated if they malfunction, in order to allow maintenance to be carried out whilst working turbines continue outputting power. Gas (SF6)-insulated switchgears are popular for offshore sites as not only they are compact but they also provide an improved level of safety than their vacuum- or oil-insulated counterparts.

12.3.2.3 Protection equipment When a fault occurs in the internal collection grid, the protection system needs to identify and to isolate the faulty component. AC protection systems consist of transducers, relays, circuit breakers, switches, auxiliary power and all wirings in between. DC breakers are still an immature technology. For instance, in high-voltage DC systems, protection systems exist currently without DC circuit breakers, all relying instead on circuit breakers on the AC side of the converters [7]. Medium voltage (MV) switchgears are used in offshore wind farms for protection and they are located inside each wind turbine and on the offshore substation. The number and type of protection used is different depending on the design, grid topology and required level of availability.

394

Offshore Wind Farms

In each turbine, there are breakers for LV and MV circuit protection. The LV protection clears internal wind turbine faults, while the MV protection is used to clear faults in the LV/MV transformer and faults in the cable connected to the next turbine. The MV circuit breakers may have the functionality to be operated remotely from shore. Gas-insulated switchgears (GIS) are also used to manually disconnect and ground transformers and cables. Protection elements on the offshore substation are used to protect against faults in the MV/HV transformer and disconnect feeders in the event of faults on the feeders. The offshore substation transformer will have additional protection for oil temperature, pressure and so on. Depending upon the internal grid topology, the type of protection differs. For example, in radial layout wind farms, it could be sufficient to place switchgear only at the beginning of the feeders, so that in case of a fault somewhere on the feeder, the whole feeder will be disconnected. Fig. 12.3 shows examples of different

(a)

(b) G

G

G

G

G

G

(c)

G

G

G

G

G

G

G

G

G

G

(d)

G

G

G

G

G

G

G

Circuit breaker Normally closed switch Normally open switch

Figure 12.3 Examples of configurations with circuit breakers and manual switches with different grid topologies. (a) String topology. (b) Star topology. (c) Ring feeder topology; a fault on any cable can be isolated and power from all six turbines can be transmitted to the substation (provided cable capacity is sufficient everywhere). (d) Ring topology where cable capacity is sufficient to carry the maximum generation from four turbines only; a fault in one of the first two cable segments, power output from two turbines will be lost [7].


395

protection schemes for different types of wind farm layouts and different levels of reliability. In case of a fault, it is the nearest circuit breakers that isolate and reduce the currents to zero in the fault area. Once this has been done, manually, remotely operated switches can possibly be used to isolate the fault even further, allowing other parts of the grid to be re-energized.

12.3.2.4 Optimum substation locations The location of any offshore substation is crucial when designing a collector system. This location has a great deal of influence on the layout of the cabling, and hence has a big impact on the expenditure of the project. Ideally, substations would be located in the farm where the amount of interarray cabling connecting all turbines is minimized. The number of substations is determined by [7]: • • •

Size of wind farm, hence the total length of cables. Voltage level, which affects maximum length of a feeder. Capacity of wind farm, and capacity of transformer and HV cable.

At a short distance from shore, small wind farms can be connected directly to shore. In such cases, the MV level is used to connect to shore. The feeders can be divided into different groups with each group having separate connection to an onshore substation. Switchgears are used at the shore end of transmission cables. Thus, there is no need for an offshore substation. As the capacity of the wind farm increases the need for having an offshore substation also increases. This is because the lifetime cost of power losses due to transmission of high power at the MV level becomes comparable to the capital cost of having an offshore substation. At a large distance from shore, offshore substations are required to transform voltage to HV level suitable for long-distance transmission. Besides greater installed power, large wind farms will also cover a larger area. This means that long cables are required within the wind farm area to connect each wind turbine (or group of wind turbine feeders) to the offshore substation. But with MV level, there is a limited length cable can have before power losses become too large and compensators for voltage regulation are needed. The solution is therefore to divide the total turbine area in two or more parts, depending on the total area coverage of the wind farm, and have an offshore substation within each area. Each offshore substation will have a separate transmission cable link to shore. Other constraints that affect the decision regarding the number of offshore substations include existing offshore industries, shipping, subsea cables, pipelines, etc., in the wind farm area. A higher voltage level increases the transmission capacity and allowable length of cables. Wind farms that have higher MV collection can be directly connected to shore which otherwise would be unfeasible to connect with standard 33 kV. A higher MV level increases the possibility of connecting small wind farms at longer distances and large wind farms at shorter distances to shore without an offshore substation. For wind farms covering a large area, with higher MV, longer array cables can be realized which would lower the need for having more than one offshore substation.

396

Offshore Wind Farms

Wind farms with large installed power, even if they are close to shore, would require an offshore substation because of the large power losses due to transmission at MV level. On the other hand, small wind farms can be directly connected to shore without the need for substation. The size of the offshore transformer can be a limiting factor for the total amount of power that can be transmitted. For wind farms with a capacity of several hundreds of MW, it could be possible to have more than one substation transformer connected in parallel in order to support higher wind farm production capacity. This will also increase the availability of the wind power because, in the case of transformer disconnection due to a fault, at least some of the power can be transported. The cost of transformers and their corresponding physical sizes and weight increase with their ratings. This also implies that the cost of offshore platforms needed to support the transformers and switchgears also increases.

12.3.3

Subsea cables

XLPE-insulated submarine cables are the most common types of interarray collection cables used in offshore wind farms. They have low electrical losses. In addition, they have high reliability and are environmentally friendly. The conducting material can either be aluminium or copper. Aluminium conductors, which are less commonly used, have a lower current-carrying capacity, and thus require large diameter and hence a larger bending radius [7]. Submarine cables can be three separate cables, one cable for each phase, also known as 3 1-core or they can be three cables bundled up together in a common shield and armouring (1 3-core). Single-core (3 1-core) cables have a higher current-carrying capacity than 3-core (1 3-core) cables. But at the same time, single-core cables have higher losses than 3-core cables. This is because in single-core cables there is current flowing through the armour that creates additional losses. There exists a magnetic coupling between the phases in single-core cables for separation distances lower than 50 m [7]. Since the cables are not transposed, the magnetic coupling results in different impedance in the cables and unbalance in the three-phase system. In 3-core systems, there is little or no coupling between the phases and therefore, the system remains symmetrical. All three cables are laid separately for 1-core cables, while for 3-core cables the cable is laid in one instance which makes the installation cost cheaper. In general, 1-core cables are more expensive than 3-core cables. Both single and 3-core cables’ shield has to be grounded to avoid overvoltages. To increase reliability an additional fourth cable can be laid in parallel to the other cables in single-core systems while for a 3-core system a second cable in parallel can be installed. There is a possibility of integrating a fibreoptic cable to be used for communication purposes with 3-core cables. Interarray cables closest to the substation carry the total sum of power produced by all the wind turbines linked to them. This required higher current-carrying capacity of these cables. The total number of wind turbine generators that can be connected in series is therefore limited by the maximum current-carrying capacity of the submarine


397

95 mm2

185 mm2

G

G 1–6 WT

7–9 WT

G

240 mm2 10 WT

33/138 kV

Figure 12.4 Different cable cross-sections in Lillgrund offshore wind farm [7].

cables. As already mentioned, higher current-carrying capacity for wind farm cables can be achieved if three 1-core cables are used instead of one 3-core cable. But the additional costs have to be justified. Submarine cables are available in a wide range of cross-sections ranging from 95 mm2 to 1000 þ mm2. One factor that determines collection system cable sizes is the maximum amount of current the cable segment is expected to carry. Depending on the wind farm layout, many different types of cable sizes can be used within a wind farm. For example, Lillgrund offshore wind farm in south Sweden has a radial layout. The cable sizes in one of the feeders are shown in Fig. 12.4. The cable sizes at the end of the feeder are smaller than cable sizes found at the beginning of the feeder closest to the substation transformer. Heating and losses are also factors that are considered when selecting the appropriate cable sizes. Some of the major submarine cable manufacturers in the world include ABB, JDR Cable Systems, Nexans, Prysmian, NSW and Parker Scanrope.

12.4

Topologies

As the power capacity of offshore wind farms increases, the adequacy of the wind farm electrical system becomes critical. The efficiency, cost, reliability and performance of the overall wind farm will depend, to a great extent, on the electrical system design [2,8]. The overall function of the electrical collector system is to collect power from individual wind turbines and maximise the overall energy generation. An electrical collector can be designed using different layouts depending on the wind farm size and the desired level of collector reliability. There are various arrangements for wind farm collector systems employed in existing offshore wind farms, whilst others are in a conceptual stage. Four basic designs shown in Fig. 12.5 are discussed below. a. Radial design The most straightforward arrangement of a wind farm collector system is a radial design (Fig. 12.5(a)), in which a number of wind turbines are connected to a single cablefeeder within a string. The maximum number of wind turbines on each string feeder is determined by the capacity of the generators and the maximum rating of the subsea cable in the string. This design offers the benefits of being simple to control and also inexpensive because the total cable length is smaller with tapering of cable capacity away from the hub being possible. The major drawback of this design is its poor reliability as, in the case of a cable or switchgear fault at the hub end of the radial string, it has the potential to prevent all downstream turbines from exporting power.

398

(a)

HV collector hub

(b)

MV collector hub

HV collector hub

MV collector hub

P Q

G1 G2 G3 G4 G5 G6 G7

P Q

P Q

P Q

33 – kV

Main hub

(d)

MV collector hub

P Q

33 – kV

MV collector hub

G1 G2 G3 G4 G5 G6 G7 P Q

HV collector hub

P Q

132 – kV

Main hub

(c)

G1

G2

G3

G4

G5

G6

G7

G8

G9

HV collector hub

B1

P Q

G14 G13 G12 G11 G10 G9 G8

P Q 132 – kV

132 – kV

33 – kV

Figure 12.5 Wind farm electrical collector basic designs [8].

Main hub

33 – kV

Offshore Wind Farms

P Q Main hub

G1 G2 G3 G4 G5 G6 G7

P Q

P Q 132 – kV

B 1


399

b. Single-sided ring design With some additional cabling, ringed layouts can address some of the security of supply issues of the radial design by incorporating a redundant path for the power flow with a string. The additional security comes at the expense of longer cable runs for a given number of wind turbines and higher cable rating requirements throughout the string circuit. A single-sided ring design, illustrated in Fig. 12.5(b), requires an additional cable run from the last wind turbine (ie, G7) to the hub. This cable must be able to handle the full power flow of the string (eg, 35 MW in a 5-MW seven-turbine string) in the event of a fault in the primary link to the hub end (denoted by the open breaker B1 in Fig. 12.5(b)). c. Double-sided ring design Fig. 12.5(c) illustrates a double-sided ring design [8,9]. In this configuration the last wind turbine in one string is interconnected to the last wind turbine in the next string (eg, G7 to G8 as shown in Fig. 12.5(c)). If the full output power of the wind turbines in one of the strings were to be delivered through the other string, then the cable at the hub end of the latter needs to be sized for the power output of double the number of wind turbines. d. Star design The star design shown in Fig. 12.5(d) aims to reduce cable ratings and to provide a high level of security for the wind farm as a whole, since one cable outage only affects one wind turbine in general. A cost implication of this design is the more complex requirements at the wind turbine in the centre of the star (that is, turbine G5 in Fig. 12.5(d)).

12.5

Converter interface arrangements and collector design

At present, offshore wind turbines are based on designs for onshore use, producing an AC output for direct connection to the electricity grid, and complying with the relevant grid codes for power quality and fault response. To gain an understanding on how wind turbine generator technology may influence control approaches, it is necessary to consider the electrical system as a whole, that is, the turbine topology (converter interface arrangement), wind farm collector, and the offshore transmission type (eg, AC or DC). Two cases are discussed next in the context of the converter interface arrangement and location [2,10].

12.5.1 Converters on turbine 12.5.1.1 AC string The conventional AC string arrangement is shown in Fig. 12.6. Turbines feature a squirrel-cage induction generator (SCIG), or a permanent magnet generator (PMG) connected to a fully rated converter, or alternatively a doubly fed induction generator (DFIG) and partially rated converter can be used. The output of the converter is stepped up to the collection network voltage, and the turbines are connected together in

400

Collector platform

To shore

33 kV AC DC

DC

±300 kV DC

From

AC

turbines

DC

AC

3.3 kV

Offshore Wind Farms

Figure 12.6 Conventional AC strings [10].

From other platforms


401

strings. The number of turbines on a string is determined by the current and voltage rating of the available cables, and the rated power of the turbines. Voltage is limited by water-treeing effect with wet insulation cables, while dry-insulation cables with a lead sheath around the insulation would be too expensive. A higher voltage also requires a higher voltage rating for the transformer, which increases the cost and size. Available current ratings are also limited, as the skin depth of the AC current means that conductors with larger areas will be less effective, as the current will not flow in the centre of the cable. For this reason, the cost of AC cables tends to increase quickly with current capacity.

12.5.1.2 DC string An arrangement using DC in the collection strings is shown in Fig. 12.7. In this system, the turbines output a DC voltage, which is then stepped up to the transmission voltage at the collection platform. In most studies, the turbine produces a voltage of around 40e50 kV DC, which requires an ACeDC converter capable of producing such a voltage, featuring many switching devices in series, or a lower voltage ACeDC converter and step-up DCeDC converter. A solution involving a lower-voltage converter and a DC voltage of 5 kV is also possible, which has the advantage of eliminating the turbine transformer and using a conventional 3.3-kV three-level converter. However, the currents in the strings will be extremely high, requiring thick cable and leading to high losses. DC systems are attractive as they could reduce the number of conversion steps between AC and DC, but converters with a voltage boost ratio will require a transformer, requiring conversion to AC and back. As DC cables do not suffer from water-treeing degradation, higher voltages could be used without needing dry-insulation cables, while the current in a DC cable can use the entire surface area of the conductor, so the cable cost will increase linearly with current capacity rather than exponentially as with AC. Because of these factors, it could be possible to implement longer turbine strings much more cheaply with DC than with AC collection. However, this is difficult to quantify as there are no commercially available cables with the required configuration and voltage rating, and previous studies of the cable cost have extrapolated the cost for multicore DC collection cables from the costs for single-core HVDC transmission cables with a significantly higher voltage rating. Another issue with DC collection networks is with fault protection, as the fact that the current does not continually reverse as with AC means that when a circuit breaker opens, the switching arc will not be automatically extinguished when the current reverses. Various DC circuit breaker designs have been proposed, but these become increasingly expensive at higher voltage ratings. DC collection and transmission networks have been designed considering the use of power converters which are capable of stepping down the voltage as well as stepping up, and these can be used to limit the fault current, but at the cost of extra complexity.

402

AC DC

AC

DC

DC

AC DC

±300 kV DC

DC

50 kV AC DC

50 kV Figure 12.7 DC strings connection.

Offshore Wind Farms

3.3 kV


403

12.5.1.3 DC series An alternative DC collection architecture is to use series DC connection of the turbines as shown in Fig. 12.8. Here the DC outputs of the turbines are connected in series, and the turbines connected in a loop. This allows the high collection voltage to be achieved without using high-voltage converters, although the wind generator converter would need to be isolated with respect to ground. An isolation transformer would need to be used, or a generator capable of handling a high-voltage offset. Another option is to use a transformer-isolated converter in the turbine, where the high-voltage side of the converter only consists of a passive diode rectifier, which is much easier to isolate. This arrangement could reduce the cable costs, as it only uses a single-core cable loop, although there is no scope to taper the current rating of the cable. In the event of a turbine fault, the faulty turbine could be bypassed using a mechanical switch, but any cable faults will mean that none of the turbines on the loop would be able to export power. A related idea is to increase the turbine output voltage and the length of the strings, so that the full transmission voltage is produced, eliminating the need for the collection platform. This system has been shown to have the lowest losses due to the high collection network voltage, and the lowest cost due to the elimination of the collection platform. Several strings could be used in parallel to increase fault tolerance. The disadvantage of this system is that the transformer and converter in the turbine must be capable of isolating the full transmission voltage, and high-voltage transformers with a low enough power rating are not commercially available.

12.5.2 Converters on platform 12.5.2.1 AC cluster Research is being conducted on a connection arrangement where turbines with fixed-speed induction generators are connected to a variable-frequency AC collection grid, with strings of turbines being connected through a single converter. This places the converters on the collection platform, allowing them to be more easily repaired in the event of a fault, and a single large converter could potentially be cheaper than several small ones. An AC or DC collection system could be used within the collection platform as shown in Fig. 12.9. The speed of all the turbines in the string can be varied together to track the maximum power point for the current wind speed, but speed control over the individual turbines is lost. The speed of each turbine will be able to vary by a small amount relative to the others, due to the slip of the induction generator, with an increase in turbine speed leading to an increase in slip and an increase in torque. Depending on the number of turbines connected to each converter, this will result in a reduction in the amount of power extracted. This system could also have an impact on the drive-train loads experienced by the turbines, as a turbine experiencing a gust would not be able to speed up to absorb the

404

Collector platform

AC DC

Figure 12.8 Series DC connection.

50 kV DC

±300 kV DC

Offshore Wind Farms

3.3 kV

DC DC

5 kV DC


Collector platform AC DC AC DC

33 kV

33 kV

33 kV

AC DC

AC DC

AC DC

DC AC

AC DC

DC AC

AC DC

DC DC

±300 kV DC

50 kV DC

±300 kV DC AC DC

33 kV

Figure 12.9 Cluster AC connection.

405

406

Offshore Wind Farms

excess power, leading to a high transient torque, putting strain on the drive train and blade roots. Research on the reliability of turbines in service has shown that the move to variable-speed turbines has reduced the level of blade failure compared with fixed-speed turbines.

12.5.2.2 Parallel DC cluster This method, shown in Fig. 12.10, uses a permanent magnet generator and passive rectifier in the turbine, with a DCeDC converter for each string of turbines. The speed of the turbine will be determined by the DC voltage of the string, so the system will behave in a similar way to the AC cluster connection system described previously, with similar issues of drive-train torque transients during gusts. It is considered that the passive rectifier will have considerably greater reliability than an active converter. For a given DC voltage, the amount of possible speed variation of the turbine will depend on the generator inductance, with a higher inductance giving a greater variation in speed. The passive rectifier is unable to supply the generator with reactive power, and if the generator inductance is too high then the maximum torque will be reduced. Inductance is typically much higher in low-speed machines, used in direct-drive turbines, and in these cases capacitors can be used between the generator and rectifier to supply the reactive power requirements. The main advantage of DC over AC clustering is the greater efficiency of the permanent-magnet generator, compared with the induction generator used in the AC system. The greater current and voltage capability of the DC cables could also lead to larger cluster sizes, and a reduction in cable cost, but this could also reduce the power capture. A DC system could also reduce the number of conversion steps and associated losses, increasing efficiency.

12.5.2.3 Series DC cluster A variation of the parallel cluster arrangement is to connect the turbines in series, in a loop, with each loop controlled by a single converter, as shown in Fig. 12.11. In this case, the converter will control the current within the loop, which will determine the generator torque within the turbine, and will be much more analogous to the

DC DC DC DC DC DC

50 kV DC

50 kV DC

Figure 12.10 Parallel DC cluster connection.

DC DC

±300 kV DC


407

DC DC DC DC

DC DC

±300 kV DC

DC DC

50 kV DC

5 kV DC

Figure 12.11 Series DC cluster connection.

conventional turbine control method. As the turbine speeds will be capable of varying individually, transient torque spikes should not be a problem. This emerging connection method is in its initial stage of investigation.

12.5.3 AC collection options: fixed or variable frequency Fixed-frequency AC operation of the offshore network is normal practice and possible with both synchronous (HVAC) and asynchronous (HVDC) connection of the offshore wind farm. AC variable frequency operation at the collection network would be cost-effective only when the wind farm is connected to the grid through an HVDC transmission system [11e14]. This is because the offshore HVDC rectifier can control the offshore frequency independently from the onshore grid, whereas for a synchronous AC transmission link an additional ACeAC or ACeDCeAC conversion system would need to be installed. Examples of wind farm configurations using DC collection as proposed in the literature are described next [15,16].

12.5.3.1 Examples of variable-frequency collection configurations A concept, currently on early research stages, is the use of variable-frequency operation in the collection network with DFIGs and an HVDC transmission link pursuing one of the following objectives [12]: • •

Reduce the rating of DFIG converters; Extend the speed range to maximise the power capture without increasing the rating of the DFIG converters.

A multiterminal configuration based on VSC-HVDC transmission that allows variable-frequency operation in the offshore collection network is under investigation [13]. The proposal in this reference is similar to the collecting network configuration presented in Jovcic and Milanovic [14] but with a multiterminal HVDC link based on current source converters that use force-commutated devices such as IGBTs. In the latter arrangement, the wind turbine generator transformers are not needed to step up the voltage to transmission level, because the current source

408

Offshore Wind Farms

converters are connected in series. In both schemes no additional converters are required at each wind turbine generator (WTG) terminal, and 2-MW permanent magnet generators are synchronized together in a group connected to a centralized converter and all are operating at the same speed. A wind turbine topology that allows variable frequency operation with a squirrelcage induction generator over a wide range of operating conditions is also under investigation [15]. The topology replaces the fully rated back-to-back converter with a single-stage cyclo-converter to decouple the frequency at the offshore substation from the wind turbine side. The scheme has the following features and potential advantages: •

•

• •

The 50-Hz three-phase transformers within each wind turbine generator and at an offshore substation are replaced by medium-frequency (400e500 Hz) single-phase transformers. The transformer insulation must be designed to cope with increased voltage stress, and high dv/dt resulting from step changes in the voltage due to the snubber capacitors of the converter. Since only a single-phase converter is required at the offshore substation instead of a three-phase converter, the number of series-connected devices required decreases significantly resulting in significant reduction in cost, conduction and switching losses. However, as the single-phase converter handles the full power this will also have an impact on cost and losses. Soft switching of the cyclo-converter and offshore converter of the VSC-HVDC reduces the switching losses significantly. The use of thyristors rather than IGBTs in the cyclo-converter reduces the power loss and cost.

12.5.3.2 AC variable-frequency collection evaluation Variable-frequency operation at the offshore collection network in conjunction with an HVDC transmission link and WTGs of the DFIG type can maximise the power extraction and reduce the overall wind farm cost according to reference [12]. However, the consequences of having variable frequency at the offshore network regarding switchgear and protection, transformer operation, voltage and current rating of the equipment located at the offshore network need to be thoroughly investigated. Standard power transformers are designed for a specific frequency of operation (50 Hz or 60 Hz), and the normal tolerance of frequency variation is around 5%, ie, for a 50-Hz unit, limits of 47.5e52.5 Hz. For a lower-frequency design of transformer, a larger core would be needed in order to maintain the required voltage ratio, for a specific current rating and a reasonable flux density avoiding saturation. However, this can be overcome by reducing the voltage in proportion to a reduction in frequency. Another aspect to be investigated is how the reactive power flow through the transformers and cables changes with variable frequency. Operating VSC-HVDC at variable frequency has been demonstrated practically in the gas platform Troll [11]. In that application, the need for variable-frequency operation by the VSC-HVDC inverter is clear, namely to control the speed of an induction motor. For an offshore wind farm on the other hand, variable frequency can be beneficial for the generator rotor only, but further research is required to assess whether it is


409

more economical to achieve this locally at each individual generator rather than at a collection network level.

12.5.4 Evaluation of higher (>33 kV) collection voltage As wind farms and turbines have increased in power capacity, the collection voltage levels have increased from typically 11 to e33 kV nowadays. The use of 48-kV or 66-kV cables to connect wind turbines to the onshore grid via a 48/132-kV or 66/132-kV transformer onshore, instead of using a 132-kV submarine cable and a 33/132-kV transformer offshore is being investigated [17]. The study shows that stepping up the voltage from 33 to 132 kV offshore is most economical for greater distance (>25 km), because the cost of the offshore substation is then less significant compared to the cable costs, and the losses are much more reduced. Another reason for using higher-voltage collection cables is that they can bring the benefit of needing fewer cable strings in collection networks for large offshore wind farms (>300 MW), because each cable has a higher capacity. Especially for everincreasing wind turbine sizes, this may become an attractive solution. For example, with present designs of 33-kV cables it would only be possible to connect up to four 8-MW wind turbines per cable string, whereas in the London Array up to nine 3.6-MW wind turbines are connected to one string. Higher voltages also reduce fault levels for a given MVA generation and ohmic losses would be less, although the overall losses in a particular design must be considered. In addition, since higher voltages offer a longer transmission distance, the collection cables can be longer, so fewer offshore ‘sub-transmission’ platforms may be needed. These were introduced for example in the 400-MW VSC-HVDC linking the BARD Offshore 1 wind farm to the transmission grid in north Germany, where the collection voltage is stepped up from 30 to 155 kV on two separate ‘sub-transmission’ platforms. These are linked via 155-kV cables to the offshore 400-MW HVDC substation [18] to transmit the power via DC over a distance of 203 km to the onshore grid. The challenges of using voltages higher than 33 kV for the collection network presently are: •

•

There is a limited supply of commercially available dry-type transformers rated above 33 kV and also capable of stepping up from a suitable generator voltage, for example 3.3 kV. They are also more expensive than the more widely available 33-kV transformers. Oil-filled transformers at the wind turbines are undesirable because of the risk of a spillage at sea, which poses an environmental threat. Collection cables at 33 kV and below can be of the ‘wet design’, which do not require metallic moisture barriers surrounding the cable as an outer sheathing layer or around the insulated core(s). The sheathing/bedding layer(s) are made from polypropylene (or jute) string/rope. Seawater fills the empty spaces inside the cable making direct contact with the outside of the insulated core(s). Higher-voltage submarine cables presently available are of the ‘dry design’, which use a lead sheath as a water barrier. Their drawbacks are the increased cable capital cost and perhaps somewhat higher installation cost due to the additional weight, and the lead sheath is susceptible to fatigue failure if movement or vibration occur.

410

Offshore Wind Farms

For a specific project, a thorough costebenefit study (ie, cost, supply chain, certification, insurance, etc.), is required to make an informed decision on whether to opt for 33-kV or higher-voltage collection cables because the impact on the overall wind farm design may be profound.

12.6

Wake farm arrangement e wake effects

The arrangements of the turbines within a wind farm depend on the site terrain, wind conditions (velocity and direction) and the size of the turbines. In order to maximize the power output from a wind farm, the wind farm layout needs to be designed in such a way that wake effects will be minimized. To reduce the wake effects and, hence, increase power output from the wind farm, the simplest option is to space the turbines far apart until wake effects are completely negligible. However, this approach will lead to increased interturbine cable cost and land wasting [19,20]. It is therefore important that the turbines are not distributed at unnecessary separations and the economical aspects of site development must balance wake effects and possible loss in energy production. In addition, to the common rectangular layout, some wind farm layout optimization studies have suggested that wind turbines should be arranged in a scattered pattern [21]. In general, in a flat terrain (eg, offshore sites), wind farm layout is mainly based on the prevalent wind direction. If the wind speed is uniform with no dominant wind direction, the distance between wind turbines in rows and columns could be about 5D (where D is the turbine rotor diameter) [21]. However, if there is a predominant wind direction, turbines are generally spaced about 1.5D and 4D apart in the cross-wind direction to the prevailing wind direction, and between 5D and 12D apart in the direction of the prevailing wind. The turbine wake, in general, is characterized by streamwise (axial) velocity deficit, which leads to less power being available for the downstream turbines. It also causes high turbulence levels which can give rise to high fatigue loads. The wake could have significant effects up to a distance 15D downstream of the upstream turbine [22]. The effect of these interactions will have severe implications on the downstream turbines which are located in the wake of the upstream ones. Depending on the distance between the turbines and the arrangement pattern in a wind farm, the power losses due to wake effects can be up to 23% [23] compared to a farm consisting of unobstructed turbines. In fact, these losses can be considerably higher for the first turbines immediately downstream of the most upstream turbine that is exposed to the undisturbed free-stream conditions. Similar effect is experienced on the subsequent downstream turbines but the effect decreases slowly downstream. The increase in fatigue loads on the downstream turbine due to wake interference effects can up to 80% [24] and this may severely shorten the life span of the rotor blades. Turbine wake properties and development depend on many factors which include the wind conditions (speed, direction and turbulence intensity), site topology and surface roughness and, upstream turbines operating conditions. The performance of any turbine operation within the wake of another turbine depends on these parameters as


411

well as the distance between them. The wake of the turbine and some of the factors that can affect its properties are schematically shown in Fig. 12.12.

12.7

Control objectives

The WTG interarrays inside a wind farm (WF) are controlled to fulfil three objectives; (1) high energy efficiency, (2) compliance with grid codes, and (3) high reliability. The control methodology of WTG increases the amount of harvested wind energy as the case of maximum power tracking (MPT). However, controlling the power transmission through the WF interarrays is also significant. As an illustration, the voltage level, losses factor and connection topology have an impact on the WF capacity factor. For example, in some cases, the output power at the default cutoff wind speed of the WTG might be equivalent to the transmission and conversion losses from the turbine to the collection platform particularly in the case of radial topologies. Conversely, star (ie, ring) topologies offer a flexible power transmission, so that in case of poor wind conditions all the power of an array is transmitted through a shorter circuit to mitigate losses. The voltage source controlled DC links offer a wider range of power transfer control [25], such that the voltage level can be tuned within limited ranges (eg, 10% from nominal voltage) to curtail the losses according to WTG production. The second task for controlling the switching and reference values for interarrays is to fulfil the grid code requirements [26,27]. Modern grid codes imply strict requirements on WFs, especially during voltage and frequency events. Voltage events include solid line to line faults where the WTG must ride-through the fault with minimum possible duration of disconnection. In addition it has to provide post-fault reactive power/current support to retain the nominal voltage fast enough.

Performance characteristics

Tip speed ratio Blade pitch angle Blade geometry Yaw angle

Wind conditions

Turbine wake

Site topology/ surface roughness

Figure 12.12 Schematic diagram of wind turbine wake parameters that can affect it.

412

Offshore Wind Farms

The control and manoeuvre of interarrays can play an essential role in limiting the fault currents, and making the WTG sense the fault in an earlier stage, instead of depending on the measurements at the point of common coupling (PCC). In the case of DC links, the stored energy in the cable capacitance could be utilized within a very short duration to provide the reactive power support when the appropriate controllers are integrated. Frequency support is also an important ancillary service to be provided by WFs. Similar to reactive power support, stored energy in cable capacitance can provide an active power surge to mitigate frequency drops [28]. In the case of over-frequency events, suitable manoeuvres (flexibility in interarray disconnection) will reduce the WF production so that the generationeload balance is reached faster. The control of interarrays can also alleviate the wind power fluctuations, especially when the WTGs, inside a WF, are located far from each other. The reliability of WFs is highly related to the connection topology of WTGs. Flexible and expanded control methods provide several alternatives for power transmission which improve the overall WF reliability and curtail the amounts of non-supplied energy. In other words, when the connection topology is changeable through controlling the switchgears; power rerouting is possible when some components fail. This also helps in maintenance procedures, namely for collection platforms so that the platform can split into more than one sector (instead of making the whole platform/WF out of service) to improve the availability of the connected WF. The applied control methods should be not so complicated and accurate at a reasonable sampling rate for the required signals, and average switching frequency for the installed power electronic converters (PECs). The complication of control methods implies longer time delays between setting the reference signals, and the response of controlled components. Moreover, the costs of integrated CPUs and FPGAs will increase, and the equivalent simulation models will require higher computational efforts (and memory storage), and their accuracy will be questionable. Likewise, high switching frequency is a requisite for some complicated control algorithms, which increases the losses and heating problems of PECs. These problems are clearer in the case of DC links rather than AC connections.

12.8

Collector design procedure

The high-level requirements for the electrical array were introduced in Section 12.2. The design procedure has to take into account and address some technical issues for the transmission and collection networks including the following [1]: • •

Connection distance to the onshore grid. This will determine whether AC or DC transmission should be used (subject to a detailed site-specific cost-benefit analysis). The WTG size and WF layout. The WTG size is required in order to be able to design the wind farm array layout. The location of the WTGs and the offshore substation within a defined area must be known in order to be able to design a collection network that is optimized for lowest life cycle costs, determined by cable lengths, losses and availability. Some research suggests that a larger WTG could reduce the cost of the collection network


•

•

413

somewhat, however the collection network with its relatively short lengths of cables bears only a relatively small proportion of the overall cost, and a larger WTG size may not be optimum for the entire wind farm. The optimum WTG output voltage. Whereas for the collection network itself there is no convincing benefit to change from the present standard AC fixed-frequency 50 Hz, a change in this may be beneficial for the WTG. Regarding DC, research has identified that it is currently an unfeasible option for the collection network because of the difficulties in stepping up a DC voltage to a required transmission level, and the market unavailability of DC circuit breakers. Variable frequency may allow omission of power electronic converters at the WTGs, thus reducing its costs, if the required converter control functionalities can be taken up by a VSC-HVDC converter offshore. Evaluating the use of 66 kV instead of 33 kV required an optimum design to be carried out for both voltage levels for a fair cost comparison. The optimal number of transmission links and offshore substations. This will partly depend on whether AC or DC transmission is used. Since high-voltage AC cables are more limited in their maximum power ratings than DC cables, they will naturally require more parallel links to connect a large offshore wind farm and thus have more built-in redundancy.

For a specific project the optimum collection and transmission networks can be designed, after the above issues are resolved. A step-wise procedure for designing the electrical system for the internal collection grid of an offshore wind farm is now introduced [7]. Since the design choices are very dependent on the specific circumstances, including cost data that are difficult to obtain on a generic level, this outline does not intend to provide a recipe for making “best” choices. Instead, it aims to describe which electrical engineering considerations need to be addressed, in which order to proceed, and what the options are. This procedure is intended for present and near-future wind farms, such that the solutions which are considered are either presently available or expected to be available in the near future. Fig. 12.13 shows the flowchart for the design procedure for the electrical design of the interarray of an offshore wind farm. The flow chart is composed by selection, calculation and decision processes. These are represented in the chart by a block with blue line, green line and red line, respectively. The procedure assumes that the capacity and physical layout of the wind farm have been defined previously. This means that wind patterns, wake effects, structural constraints, and all relevant studies concerned with the location are assumed to be available data. In the flowchart, the RAMS stands for reliability, availability, maintainability, and serviceability. In general, high reliability is required from offshore wind farm components due to difficult and expensive access to perform repairs. Overall reliability is usually calculated based on a number of reliability indices specified at the component level. These indices include, for example, failure rate and time-to-repair. Availability refers to the ability of the wind farm to produce the maximum possible amount of power. It is dependent on the reliability, but can be increased by including redundancy. The level of reliability/availability within the wind farm is a trade-off between the extra investment cost and the cost savings from reduced loss of production. The choice affects the grid topology and switchgear and protection system in the wind farm.

414

Offshore Wind Farms

Start

Capacity of the turbines Physical layout

Input data

Wind turbine technology

Select AC or DC

Select voltage level Number of substation and location

Cable optimization

Select topology

Select cable type(s)

Load flow/ short circuit Yes No Availability of cables

Maximum current limit

No

Yes No Minimum power losses?

Yes Switch/protection system RAMS

No

Reliability indices fulfilled?

Yes End

Figure 12.13 Flowchart of procedure for designing the electrical collector of an offshore wind farm [7].


415

Once the reliability indices have been defined, the RAMS assessment is done. Based on this information, the topology can be modified to improve wind farm availability or to make the system less costly. If the topology is modified all the power flow and short-circuit analyses must be done again to ensure the cables and protection systems are correctly calculated.

Abbreviations CAPEX

Capital expenditure

CPU

Central processing unit

D

Diameter

DC

Direct current

DFIG

Doubly fed induction generator

FPGA

Field programmable gate array

GIS

Gas-insulated switchgears

HVAC

High-voltage alternating current

HVDC

High-voltage direct current

IEC

International Electrotechnical Commission

IEEE

Institute of Electrical and Electronics Engineers

LV

Low voltage

MV

Medium voltage

OPEX

Operational expenditure

PCC

Point of common coupling

PEC

Power electronic converter

PMG

Permanent magnet generator

RAMS

Reliability, availability, maintainability and serviceability

SCIG

Squirrel-cage induction generator

SCT

System Operator-Transmission Owner Code

STATCOM

Static compensator

SVC

Static var compensator

VSC

Voltage source converter

WF

Wind farm

WTG

Wind turbine generators

XLPE

Cross-linked polyethylene

416

Offshore Wind Farms

Acknowledgements Some content in Sections 12.3.1, 12.3.2, 12.4 and 12.5 is adopted from that originally published in [2]. The author of this chapter also acknowledges Harald Svendsen, Atsede Endegnanew and Raymundo Torres-Olguin for the material in Sections 12.3.2 and 12.8, which was adapted from their report in Endegnanew et al. [7].

References [1] Aten M, Philip G, Anaya-Lara O. Electrical array system report. Helmwind Project. 2010. [2] Anaya-Lara O, Campos-Gaona D, Moreno-Goytia EL, Adam GP. Offshore wind energy generation: control, protection, and integration to electrical systems. Wiley; May 2014, ISBN 978-1-118-53962-0. [3] E.ON Climate and Renewables. E.ON offshore wind energy factbook. 2012. www.eon. com [last accessed 04.09.13]. [4] Dolan D, Jha A, Gur T, Soyoz S, Alpdogan C, Camp T. Comparative study of offshore wind turbine generators (OWTG) standards. Oakland, CA: MMI Engineering; 2009. [5] The Crown State. A guide to an offshore wind farm. www.thecrownestate.co.uk/media/.../ guide_to_offshore_windfarm.pdf [last accessed 21.10.13]. [6] Bazargan M. Renewables offshore wind: offshore substation. Power Eng 2007;21(3): 26e7. [7] Endegnanew A, Svendsen H, Torres-Olguin R. Design procedure for inter-array electric design (D2.2). EU FP7 EERA-DTOC Project. 2013. [8] Quinonez-Varela G, Ault GW, Anaya-Lara O, McDonald JR. Electrical collector system options for large offshore wind farms. IET Renewable Power Gener 2007;1(2):107e14. [9] Sanino A, Liljestrand L, Breder H, Koldby E. On some aspects of design and operation of large offshore wind parks. In: Sixth international workshop on large-scale integration of wind power and transmission networks for offshore wind farms, Delft, The Netherlands; 2006. p. 85e94. [10] Parker M, Anaya-Lara O. The cost and losses associated with offshore windfarm collection networks which centralise the turbine power electronic converters. 2013. IET-RPG, 2013. [11] Hyttinen M, Bentzen K. Operating experiences with a voltage source converter HVDC-light on the gas platform troll A. ABB Website; 2002. [12] Feltes C, Erlich I. Variable frequency operation of DFIG based wind farms connected to the grid through VSC-HVDC link. In: IEEE Power Engineering Society general meeting, 24e28 June 2007; 2007. [13] Jovcic D. Interconnecting offshore wind farms using multi-terminal VSC-based HVDC. In: IEEE Power Engineering Society general meeting, 2006; 2006. [14] Jovcic D, Milanovic JV. Offshore wind farm based on variable frequency mini-grids with multi-terminal DC interconnection. In: The 8th IEE International Conference on AC-DC Power Transmission (ACDC 2006), London, UK, 28e31 March 2006; 2006. [15] Meyer C, et al. Control and design of DC grids for offshore wind farms. IEEE Trans Ind Appl November/December 2007;43(6):1475e82. [16] Lundberg S. Performance comparison of wind park configurations. Technical Report. Chalmers University of Technology; 2003.


417

[17] McDermott R. Investigation of use of higher AC voltages on offshore wind farms. In: EWEC 2009, Marseille, France 16e19 March 2009; 2009. In: www.ewec2009 proceedings.info/allfiles2/283_EWEC2009presentation.pdf. [18] Stendius L. Nord E.ON 1 400 MW HVDC light offshore wind project. In: EOW 2007, Berlin, 5 December 2007; 2007. In: http://www.eow2007proceedings.info/allfiles/290_ Eow2007presentation.ppt. [19] Anaya-Lara O, Tande JO, Uhlen K, Undeland T, Svensen H. Control challenges and opportunities for large offshore wind farms. Internal Report. SINTEF Energy; December 2012. [20] Mustakerov I, Borissova D. Wind turbines type and number choice using combinatorial optimization. Renewable Energy 2010;35:1887e94. [21] Mardidis G, Lazarou S, Pyrgioti E. Optimal placement of wind turbines in a wind park using Monte Carlo simulation. Renewable Energy 2008;33:1455e60. [22] Højstrup J. Spectral coherence in wind turbine wakes. J Wind Eng Ind Aerodyn 1999;80: 137e46. [23] Dahlberg J-Å, Thor S-E. Power performance and wake effects in the closely spaced Lillgrund wind farm. Proceedings of European offshore wind 2009 conference and exhibition, 14e16 September, Stockholm, Sweden. [24] Sanderse B. Aerodynamics of wind turbine wakes: literature review. Energy Research Centre of the Netherlands (ECN) Report ECN-E-09-016. 2009. [25] Du C. VSC-HVDC for industrial power systems. In: Energy and environment. Chalmers University of Technology; 2007. [26] Anaya-Lara O, Ledesma P. D2.5 Procedure for verification of grid code compliance. University of Strathclyde; 2012. [27] Tsili M, Papathanassiou S. A review of grid code technical requirements for wind farms. Renewable Power Gener, IET 2009;3(3):308e32. [28] Junyent-Ferre A, Pipelzadeh Y, Green T. Blending HVDC-link energy storage and offshore wind turbine inertia for fast frequency response. IEEE Trans Sustainable Energy 2015;6:1e8. PP(99).


Cabling to connect offshore wind turbines to onshore facilities

13

Narakorn Srinil Newcastle University, United Kingdom

13.1

Introduction

Cabling systems play a vital role in integrating an offshore wind farm into a national grid or electricity transmission network by conveying a vast amount of electricity over a long distance to thousands of homes and millions of people. The cabling industry has experienced significant growth over the years as demand for subsea cables is growing steadily. Many countries make commitments to offshore renewable wind energy and interconnect to their national grids. As more projects of offshore wind farms in deeper waters with variable environmental conditions are proposed to improve the financial return, advanced cabling technology is required to develop higher-capacity and larger-diameter power cables. For safe and cost-efficient energy transmission, subsea cable layouts must be optimized, depending on the desired wind turbine locations, offshore and onshore infrastructures, geophysical and geotechnical properties. In light of the international commitments to significantly increase offshore wind capacity, the interarray and export cabling systems have been identified by the offshore wind industry as ones of the key areas where related cost savings should be considered. In practice, the capital expenditures can be reduced by increasing the cable voltage levels. For a specific wind farm, the lifecycle cost reductions, particularly for interarray and export cables, can be further achieved through the cable installation procedure, sequence and failure prevention. These are controlled by several influential factors, including the offshore wind farm and coastline architecture, water depth, wind turbine size, fixed or floating foundation substructure type, environmental sea state and seabed conditions, and the arrangement of construction, transportation and installation (Department for Business Enterprise & Regulatory Reform, 2008). This chapter is devoted to a review of general and practical aspects of offshore wind farm cables, current technical challenges, potential technology developments and innovations to minimize cable installation costs, damage rates and hence insurance costs. Some theoretical background and analysis formulae are provided with attention being paid to the global structural mechanical behaviours of flexible cables rather than their electrical, manufacturing or material counterparts which can be found in other excellent textbooks (Burton et al., 2011; Worzyk, 2009) and recognized offshore standards (DNV, 2012, 2014).


420

13.2

Offshore Wind Farms

Offshore wind farm cables

An offshore wind farm consists of a number of wind turbine generators, substructures and supporting foundations which need to be appropriately positioned to minimize the overall operational costs and the energy loss due to the wind interference effect from upstream or nearby turbines. The impact on the environment must also be thoroughly assessed as was done, for example, for the Beatrice wind farm (2011). The electrical transmission system for offshore wind farm technologies includes interarray, interplatform and export cables of different specifications, offshore collector or substation and converter (if required) infrastructures. The key design mechanical parameters of cables with composite materials include outer diameters, dry and submerged weights, lengths, axial and bending stiffness, and end connections, apart from the dynamic properties such as natural frequencies, modal shapes and viscous damping coefficients. These are chosen variably, depending on the wind farm site physical and environmental characteristics, turbine rating and sizes, distribution network connection and operational voltages, polymeric insulation properties such as cross-linked polyethylene (XLPE) or ethylene propylene rubber (EPR), conductor materials (copper or aluminium) and cross-sectional areas, embedded fibreoptic communication control lines, external sheath protections and offshore routes to mainland facilities (Worzyk, 2009). Fig. 13.1 illustrates different functions of offshore wind farm cabling systems for fixed and floating foundations. Tables 13.1e13.3 demonstrate properties including the approximated outer diameters and weights of some typical interarray (33 kV, AC), export HVAC (132 kV) and export HVDC (300 kV, bipole) cables, respectively.

13.2.1

Interarray cables

Interarray cables link several wind turbines to offshore substation platforms to allow the electricity generated from each turbine to be collected and transformed before being transmitted to onshore facilities. Within the turbine tower, each cable end is terminated onto the high-voltage switchgear, which is able to isolate cable failures and keep the remaining functioning parts of the turbine in operation. Cables between wind turbines are relatively short in length being approximately less than 1500 m, depending on the turbine rotor diameter (DR) and the minimum spacing requirement (typically more than 5DRe7DR) between the two turbine systems. However, those connecting turbines to the offshore substations could be longer and possibly up to 3000 m. Interarray cables are typically three-core copper conductors with steel wire armoured and conductor/insulation screening components (Worzyk, 2009). The nominal operating voltage of the cable system is in the medium voltage range, ie, 33 kV. The 33-kV cable is traditionally the standard for offshore wind distribution, but a higher voltage cable (ie, 66 kV) is under investigation to be considered by the future offshore wind industry for larger wind farms with high wind turbine ratings (Ferguson et al., 2012). The higher voltage 66-kV cable with a smaller crosssection and lower current would allow for greater power capacity and a reduction


421

(a)

Onshore facilities

Export cable

Inter-array cable

(b)

Onshore facilities

Inter-array cable

Inter-platform cable

Export cable

(c)

Onshore facilities

Inter-array cable

Inter-platform cable

Export cable

Figure 13.1 Offshore wind farm cables for fixed (a, b) and floating (c) wind turbines, substations (aec) and converter (b, c) platforms.

Table 13.1

Example of interarray cable properties Nominal cross-sectional area of conductor (mm2)

Details

95

120

185

240

Outer diameter (mm)

104

106

114

119

16

18

21

23

Weight (kg/m)

422

Offshore Wind Farms

Table 13.2

Example of interplatform cable properties Nominal cross-sectional area of conductor (mm2)

Details

300

400

500

800

Outer diameter (mm)

167

168

176

194

48

51

58

74

Weight (kg/m) Table 13.3

Example of export cable properties Nominal cross-sectional area of conductor (mm2)

Details

300

800

2000

3000

Outer diameter (mm)

102

118

140

155

22

33

53

70

Weight (kg/m)

of system power losses. In addition, a fewer offshore substations with significantly lighter transformers would be required with the potential cuts of overall cable component necessities. Together with the optimization of the wind turbine spacing and interarray cable layout, the higher-voltage transmission could offer lower lifecycle cost savings. By using aluminium conductors, costs might be further reduced by about a third. However, due to the lighter weight of aluminium, attention must be paid to the greater dynamic cable movement and a higher bending radius during the installation in water. Recently, the UK Carbon Trust estimated that the use of 66-kV interarray cables could cut the cost of offshore wind energy by about 1.5% (Carbon Trust launch race for next generation of offshore wind cables, 2013). Accordingly, several manufacturers now aim to design, qualify and certify 66-kV cables to be integrated into relevant switchgears and transformers whose specifications are available at 66 kV.

13.2.2

Interplatform cables

A number of offshore collector platforms are typically required to gather power from wind turbine generators through interarray cables (see Fig. 13.1). To minimize the transmission loss versus increasing distance to shore, the step-up transformers located on the collector platforms are used to increase the cable voltage to the higher level (eg, 132 kV and higher). If the HVDC (high-voltage direct-current) export cables are desired for a wind farm located very far from the mainland (more than 60e100 km or 80 km on average), the electricity from the offshore collector platforms will be transmitted to a converter platform via the HVAC (high-voltage alternating-current) interplatform cables. The HVAC is converted into the HVDC with nominal voltage possibly up to 525 kV (ABB launches world’s most powerful cable system for renewables, 2014).


423

Both interarray and interplatform three-core cables are substantially similar, except for the thickness of the insulation layers and thus the resulting outer diameters which may be different (up to 250 mm for interarray cables and 300 mm for interplatform cables) according to the different carried voltages. For HVAC interplatform cabling, a single-core cable design may also be used. Fig. 13.2(a) illustrates a three-core copper XLPE 33-kV interarray cable. Fig. 13.3 presents a schematic cross-sectional structure of the three-core XLPE 66-kV cable proposed for the Fukushima offshore wind farm (Tominaga et al., 2014).

13.2.3 Export cables The key function of offshore export cables is to efficiently transmit the electrical power with minimal loss from the offshore wind farm to the cable connection facility at the landfall point. Export cables may operate with HVDC (single-core) or HVAC (one- or three-core) technology, depending on several balancing factors, particularly the offshore distance to shore, the total generated power of the wind farm, material and operational costs. HVDC cables require both offshore (AC/DC) and onshore (DC/AC) converter stations, and the erection of an offshore converter station is fairly expensive. However, HVDC cables have greater advantages when a larger amount of power is transmitted over a long distance due to the power loss reduction as compared to the HVAC transmission as shown in Fig. 13.4 (Gevorgian and Hedrington, 2010). In addition, the HVDC system reduces cable quantities or Figure 13.2 Example of (a) three-core 33-kV cable and (b) single-core HVDC cable.

424

Offshore Wind Farms

Conductor Conductor screen XLPE insulation

Insulation screen Metallic screen Metallic sheath Inner sheath Filler Bedding Armour Outer sheath Optimal fibre unit

Figure 13.3 Example of three-core XLPE 66-kV cable.

Costs Total AC technology cost

Total DC technology cost

AC losses

DC losses

DC line

AC line

DC terminals

AC terminals Transmission distance Break-even distance ( ∼ 60–70 km)

Figure 13.4 Cost comparison of HVDC and HVAC cabling technologies.


425

constraints on supply and allows the transmission to better handle the occurrence of strong wind gusts. The baseline offshore export cable design assumption is a pair of single-core HVDC cables (typically XLPE insulated and with copper conductor) for each converter platform, with a high positive and negative voltage relative to the earth. The two cables of each HVDC circuit and associated fibreoptic cables or health-monitoring sensors may be bundled together as a single unit and buried in the same subsea trench until they approach the landfall location. Alternatively, they can be installed separately, depending on the cable burial strategy and local seabed conditions. It is noted that the installation cost per km of a single bundle export link is considerably lower than that of individual cables. Fig. 13.2(b) illustrates a single-core HVDC export cable. Based on a recent report summary towards the end of December 2014, there are currently 47 export cables for 21 offshore substations and 1309 wind turbines operating in UK waters (The Crown Estate, 2015a).

13.2.4 Cable layout and spacing For known locations of offshore wind turbines which depend mainly on the prevailing wind direction and geotechnical properties, interarray cables are more constrained than interplatform and export cables in terms of the layout and spacing within a wind farm. From an economical viewpoint, the shortest routes (so-called strings) to the substation platforms for interarray cables should be wisely determined and optimized from the outset of any development plan by also taking into account several factors and risk-based assessments. At the present time, wind turbines in several offshore wind farms are spaced apart for a certain distance which can in turn influence the access of subsequent activities including cable (laying and burial) installation and turbine-substructure maintenance. Since local information and circumstances of different wind farms are site-specific, the cable layout and spacing analysis should be carried out on a case-by-case basis (United States Department of Interior Bureau of Safety and Environmental Enforcement (2014)). Depending on the number of turbines per string, the number of strings, the substation and converter locations, and the export cable orientations within a wind farm, several subsea layouts of interarray cables can be proposed. The main categories may include the radial (open-loop), radial with branching, and ring (closed-loop) designs as shown in Fig. 13.5. With respect to the cable lengths, capital costs and operational power losses, the radial-branching networks may be more optimal than the radial ones for fixed wind turbine locations (Jenkins et al., 2013). In all layout cases, the use of 66-kV cables with possible aluminium conductors will potentially reduce capital costs for offshore cabling if higher bending radius can be accommodated.

426

Offshore Wind Farms

Figure 13.5 Example of subsea layout of offshore wind farm interarray and export cables (similar to the layout of Gwynt y Mor wind farm). ˇ

13.3

Offshore cable installation, protection and challenges

Large offshore wind farms in deep waters experience several technical challenges and unexpected delays related to the installation activities of subsea cables. Recommendations and standardization of cable-laying burial techniques and sequences are needed to reduce the scheduling and operational costs. There are also risks associated with the cable transportation and installation which can affect the cable long-term performances. The financial, technical and insurance-related costs and potential damage to corporate reputations have significant implications. Offshore cables require diligent planning to minimize potential risks at an early design stage. Typical key stages of offshore cable installation include the preinstallation survey, route clearance, cable laying and burial, and postinstallation assessment. To safeguard subsea cables from environmental hazards, offshore works normally take place outside the winter months, based on a 24-h basis and the cables are installed in narrow trenches (typically about 250 mm wide) or buried into the seabed at a predetermined depth depending on the soil conditions. Where cable burial is unfeasible (eg, near the wind turbine foundation) or insufficient (eg, due to the active seabed scour), reliable protection means should be alternatively considered and implemented. These include, for example, the use of controlled rock placements, rock dumping, preformed concrete mattresses, positioning of grout or gravel bags, ducting and use of sleeves (Beindorff and Baalen, 2013). Estimation of the installation time and cost related to cable installations can be found in Kaiser and Snyder (2012).


427

13.3.1 Preinstallation survey Detailed preinstallation survey and mapping of offshore geotechnical, environmental, archaeological and meteorological conditions are essential to the selection of subsea cable routes, corridors, layouts, installation planning and methodologies. They are also useful for the identification of potential hazards and obstacles which can impact on the cable laying and burial operations. These include, for example, large rocks, wrecks, unexploded ordnances, areas near an aquaculture farm, and existing pipeline and cable infrastructures, especially in the North Sea regions. In general, a 500-m Advisory Safety Zone (ASZ) is requested around all installation vessels, while in the wind farm construction area a 50-m ASZ around each installed turbine is required. A 250-m ASZ is required along the export cable route (The Crown Estate, 2012). For geotechnical explorations, the preinstallation frameworks may include the cone penetrometer, boreholes and vibrocore tests of soil samples at a maximum depth of 5 m below the mean seabed level, within a wind farm and along the potential export cable routes (Worzyk, 2009). Cable burial trials may also be conducted in advance to make sure that the desired burial equipment is appropriate for actual seabed conditions while achieving the burial depth requirements. Such trials should be planned and carried out through a certain length (eg, 1 km) covering different soil types. From a geophysical standpoint, the location and mobility of sand waves along the cable route should be determined to assess whether such features could be avoided or, if impossible, whether seabed preparation or clearance is required to ensure the stable seabed character maintaining the burial depth index and without uncovering the buried cables in a short period of time.

13.3.2 Cable laying and burial practice In practice, there are several approaches available for cable installation and the finally selected one would be based on overall data obtained from the preinstallation site investigations. For a large offshore wind farm such as the Dogger Bank Creyke Beck (FOREWIND, 2013), a combination of ploughing, jetting and trenching may be considered for cable installation, depending on the water depth, seabed condition and level of sedimentary variation along cable routes. In practice, the interarray cabling operation would take place before the export one. For export cables, the installation may start from the wind farm towards shore or in the opposite direction. The most common wind farm cable-laying and burial philosophy may be largely classified as follows.

13.3.2.1 Simultaneous lay and burial The laying and burial operations are coupled. Cables are laid by a turntable and buried simultaneously with burial equipment, such as a plough, being towed by a self-propelled large vessel or spread-moored barge, cutting a narrow trench, lifting the soil out of the trench and then backfilling over the laid cables. This continuous operation is typically considered for long export cables because of their lengths and heavy weights.

428

Offshore Wind Farms

The main advantage of this coupled laying-burial method is minimized risks to the cable from exposure. However, this system requires synchronized control of lay and burial factors as well as a large and more reliable weather window such as in the summer, with the laying effectiveness being naturally governed by the burial speed. A slow burial process would hence slow down the entire laying operation and this is particularly costly for a high day-rate, which could be as high as £200,000 a day, of a specialized large cable payload vessel used for the cable deployment. In addition, the ploughing performance achieving the target burial depth is dependent on the plough equipment geometry and seabed conditions with a very stiff clay or dense sand providing greater challenges. For very stiff and bedrock-type seabed layers, modern mechanical trenchers may be alternatively used. Delay in laying subsea cables can lead to rising operational costs, as seen from the Lewis Wind Farm in the Western Isles where delays in cable laying or assessing the cable business case have been blamed, resulting in a withdraw of the project plan (Lewis wind farm backer pulls out over cable delay, 2014).

13.3.2.2 Post-lay burial The laying and burial operations are decoupled. Cables are laid separately in advance by a cable-laying vessel and they are subsequently buried by a separate dynamically positioned vessel using, for example, a high-pressure water jetting combined with a remotely operated vehicle (ROV) tracking system. The post-lay burial method has the advantage of the laying operation being able to be completed more rapidly with the slower burial being performed independently and at a convenient time. This cable burial operation can be done by a lower day-rate vessel. However, hazards may arise when the cable is left unprotected for some time on the seabed. These can lead to cable lateral and vertical movements, the so-called on-bottom stability, or damage due to the penetration of fishing gear and drag anchor during the bottom-trawling and vessel-positioning activities. Therefore, if the post-lay burial solution is used for reducing the installation time and cost, the period of cable exposure should be limited (ie, within a few hours) and a guard vessel should be deployed patrolling the cable route to warn fishing boats. It is crucial that the fishermen are informed of unburied cables which may pose a risk to their fishing activities. As realized from the Greater Gabbard offshore wind farm, problems with unburied cables covered with improperly laid subsea mattresses have been criticized by fishermen (Greater Gabbard OWF’s Cabling Infrastructure Worries Local Fishermen, 2014). An example of the post-lay burial method is the Gwynt y Mor Offshore Wind Farm project. There are four HVAC export cables with the copper conductor, XLPE insulation, 132-kV nominal voltages and an average length of 21.3 km. In addition, there are 161 interarray XLPE copper 33-kV cables with a total length of 148 km (Gwynt y Mor cabling complete, 2015). ˇ

13.3.3

Cable burial depth identification

One of the challenging aspects for offshore wind farm cables is to identify the minimum burial depth which can provide long-term safe protection of subsea cables against any


429

underwater hazards or threats such as trawling, fishing gear and ship anchoring throughout their operational lifetime. This is important because the damage, integrity loss or fault of the cabling system is difficult to repair and the associated energy losses and insurance claims can be significant. Depending on the seabed condition and the level of protection or hazard type, the burial depth of several wind farm cables is typically in the range of 0.5e2 m following a detailed burial risk assessment and the widely recognized ‘Burial Protection Index’ (BPI) based on the real ploughing database and developed originally for fibreoptic telecommunication cables (Mole et al., 1997). Recently, the Carbon Trust’s Offshore Wind Accelerator (OWA) joint industrial programme has published recommendations for a new optimization-based Cable Burial Risk Assessment (CBRA) methodology to determine the cable optimal and minimum ‘Depth of Lowering’ (DoL), which is practically and economically acceptable from a cable protection viewpoint based on the probability of attack (eg, due to an anchor) on the cable (Carbon Trust, 2015). Generally speaking, the CBRA method is intended to reduce the conservatism and uncertainties in the use of a significant burial depth normally required by stakeholders or operators. Depending on the site updated survey data, the CBRA method can be used several times during the wind farm development from design, routing, trenching, laying and decommissioning, and should be used independently for interarray and export cables. Fig. 13.6 shows the definition of the trench minimum versus target DoL suggested by the CBRA method and the cable installer, respectively. It is noted that because of a lower DoL suggested by the CBRA method, the overall cost of cable installation can be reduced.

13.3.4 J-tube and J-tubeless interface Each end of the interarray, interplatform and export cables is typically connected to the wind turbine generators, substation or converter platforms through the so-called J-tube entry and interface system (DNV, 2014). This steel conduit with a nearly ‘J’ shape may be placed inside or outside the foundation substructure (monopile, jacket, gravity base or tripod) with position being fixed (inside and outside the tube) or adjustable with respect

Mean seabed level Recommended minimum depth of lowering

Cover depth Target depth of lowering

Cable

Target trench depth

Figure 13.6 Trench parameters.

430

Offshore Wind Farms

to the optimal height and angle (outside the tube). The J-tube path runs vertically from the cable termination or hang-off point at the tower base above sea level down to the seabed bottom and has an outward opening bellmouth where the cable is pulled through by means of a messenger wire and winch. To guarantee a smooth transition without damage and ripping of the cable during the pull-in operation, the bellmouth radius should be larger than the cable minimum bending radius (MBR). Fig. 13.7 displays the standard J-tube and J-tubeless systems of a monopile and jacket foundation.

(a)

J-tube

(b)

Figure 13.7 External/internal J-tube and J-tubeless systems for (a) monopile and (b) jacket.


431

The number and size of J-tubes required for each wind turbine and substation depend on the wind farm capacity and subsea cable layouts. Two J-tubes are normally found for turbine generators. For cable interlaying between the two end J-tubes, the cable part close to the second end should be laid with a greater slack condition to allow for additional movement or stretching. For a single large J-tube accommodating several cables, for example, to a substation platform, the thermal and heat condition within the crowded tube should be given special attention. To replace the requirement for costly steels and their lifetime corrosions, a J-tube with composite materials may be designed and implemented. To reduce the cable installation time and cost by eliminating the need for protective J-tubes and pull-in operations, a novel concept of a J-tubeless entry system has recently been introduced to modern offshore wind farms. Because the cable suspended from the hang-off location is flexible and susceptible to movement, this J-tubeless system should be internally routed inside the wind turbine substructure to avoid the dynamically environmental impact from waves and/or currents. The feasibility, benefit and risk of a free-hanging cabling design should be carefully explored with respect to the cable dynamics to fully understand the mechanical fatigue versus electrical testing results and determine realistic capital and operational costs versus the potential impact on the wind farm development schedule. For both J-tube and J-tubeless interfaces, the subsea protection of a free-spanning section (typically about 20 m long) outside the wind turbine foundation of an interarray cable is needed. This protection is normally based on the recognized techniques including the use of rock dumping (also for scour protection), concrete mattresses plus sandbags, and the inflatable grout bags, some of which requiring diver intervention or assistance. For a zero-diver operation, the static or dynamic bend restrictors may be implemented by utilizing a heritage of standards from the oil and gas industry. A series of bend restrictors control and protect the cable from the variation of MBR or overbending during the installation and operation.

13.4

Dynamic cables for floating wind turbines and substations

For offshore floating wind turbines, attention must be paid to the development, design and optimization of a robust dynamic power cable which requires a high level of flexibility, compliance with a floater motion and fatigue resistance subject to dynamic environments. The combined effects of waves, currents, seabed touchdown interactions and top-end floater movements on the cable connected to a floater are critical. These need to be fully investigated via a global fluidestructure interaction and multibody coupling model together with the experimental tests. Resembling offshore risers used in the oil and gas applications for transporting fluids from the seabed wells to the platforms (Bai and Bai, 2014), potential configurations for floating turbine or platform cables are displayed in Fig. 13.8 which includes the so-called ‘W’ (Fig. 13.8(a)) and lazy wave (Fig. 13.8(b)) shapes. The latter concept

432

Offshore Wind Farms

(a)

Floating turbine

W-shaped cable

Mooring line

(b)

Floating substation

Buoyancy modules

Lazy-wave cable

Figure 13.8 Dynamic power cable concepts: (a) interarray cable; (b) export cable.

has recently been considered for some floating wind turbines such as the Statoil Hywind (Offshore wind farms: Floating solutions for deep-water cables, 2010) and those applied to the Fukushima offshore wind farm (Tominaga et al., 2014). The W-type cable may be used for connecting between the two floating wind turbines or between the floating substation and wind turbines, whereas the lazy wave cable may be used as an interarray, interplatform and export cable. In any case, there is a requirement for a design of buoyancy modules (typically made of synthetic foams) to


433

be distributed along a certain mid-water part of the cable. Secure fastening and precise positioning of buoyancy modules are ensured by using internal clamps. The use of such distributed buoyancy is mainly aimed at minimizing the overall cable dynamic responses and accommodating the cable dynamic tension or curvature changes in harsh weather conditions, by decoupling the platform motion from the cable touchdown point at the seabed. Since the W or wave cable configuration is significantly controlled by the change of additional buoyancy, designers must accurately estimate the buoyancy particulars such as the diameter, submerged weight (density) and spacing: these should be maintained throughout the lifetime operations by avoiding any buoyancy loss and/or slippage. Fig. 13.9 displays the effect of buoyancy change on the overall configurations of lazy wave cables manifesting different hang-off, arch bend and touchdown catenaries. Depending on the hang-off angle, water depth, platform offset, footprint and soil stiffness, areas of local maximum static tensions typically occur at the hang-off point and the two ends (ie, lift and drag points) of the distributed buoyancy portion, whereas areas of local maximum static bending moments correspond to the sagging, hogging and touchdown areas where maximum curvatures occur. These parameters should be systematically verified through a cable shape sensitivity study. Nevertheless, in practice there are some other effects such as the mean drag force and its direction owing to the combination of wave and current, marine growth and installation imperfection (eg, due to the water absorption or ingress to a buoyancy module) which can lead to variable or uncertain hydrodynamic (normal and tangential) drag coefficients to be used in the design of non-uniform cable sections in comparison with the bare sections. These effects have to be accounted for using a more sophisticated finite element or computational fluid-structural dynamics software. 160

Hang-off catenaries Vertical span along water depth (m)

140

Arch-bend catenaries

120

100

80

60 40

Touch-down catenaries

20

0 0

50

100

150

200

250

300

Horizontal span along seabed (m)

Figure 13.9 Effect of buoyancy distribution (red line) change on the configuration of dynamic cable.

434

Offshore Wind Farms

The free-hanging dynamic cable also requires a special protection at the platform hang-off location and seabed end termination using, for example, a bend stiffener or restrictor to avoid high local bending stresses leading to the increased fatigue damage of the cable (Bai and Bai, 2014).

13.5

Some mechanical aspects of subsea cables

Cable is a one-dimensional structure being inherently flexible, having a high slenderness, high aspect (length-to-diameter) ratio, high axial but lower bending rigidity. Three-dimensional large cable displacements and associated axial or bending stresses can be induced by different environmental loading during operational phases (Srinil, 2004). Therefore, it is of practical importance to understand some fundamental properties and mechanical aspects of cable structures installed at sea and connecting different support structures. To evaluate cable static and dynamic behaviours as well as fatigue, there are several models, formulae and governing (eg, thermal, electrical, mechanical, boundary condition) factors to be parametrically considered during the early stage analysis and design of subsea cables. Herein, some of the key mechanical and global structural analyses are discussed with reference to Fig. 13.10.

13.5.1

Catenary configuration

A single-span cable shape during the laying operation may be described based on a simple catenary theory, as shown in Fig. 13.10, which assumes that the cable configuration only depends on its own effective weight per length in water (w) and axial tension (T). The effects of static deformation, bending moment and drag force are neglected as they may be considered to be of secondary importance. By further

Wave Cable-laying vessel

Cable Current

Bending radius

Seabed

Figure 13.10 Cable configuration being laid by a vessel.


435

assuming a zero tangential slope at the cable seabed touchdown, a horizontal tension component (TH) is thus a constant and equal to the seabed bottom tension. Accordingly, the Cartesian x-y coordinates along the arc length (s) can be derived as (Berteaux, 1976) TH 1 ws sinh x¼ TH w y¼

TH ws 1 cosh TH w

[13.1] [13.2]

13.5.2 Minimum bending radius (MBR) During the cable vertical laying operation, the maximum bending stress normally occurs near the seabed where the minimum radius (maximum curvature) occurs. Using a linear momentecurvature relationship and Eqs. [13.1] and [13.2], the minimum bending radius (MBR) at the seabed may be simply approximated as MBR ¼ TH =w

[13.3]

Alternatively, if the total length (L) and vertical location of the hang-off or laying point (H) of the cable is known, it may also be approximated that MBR ¼

L2 H 2 2H

[13.4]

For a horizontally radial laying operation where a straight-line laying approach is not feasible due to, eg, an obstacle, Worzyk (2009) suggested, by introducing a safety factor of two and the cable-soil lateral friction coefficient m, that MBR ¼

2TH mw

[13.5]

13.5.3 Laying tension and hang-off angle As the cable is being laid vertically from the vessel, the cable tensile force at the laying wheel (TL) may be approximated as TL ¼

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi T 2H þ ðwLÞ2 ¼ TH þ wH

[13.6]

Correspondingly, the departure or hang-off angle with respect to the horizontal level at the laying wheel may be approximated in different ways as f ¼ tan1

wL TH wL ¼ cos1 ¼ sin1 TH TL TL

[13.7]

436

13.5.4

Offshore Wind Farms

On-bottom stability

In shallow water with a high energetic environment, it may be necessary to understand the on-bottom stability of interarray cables being laid on the seabed prior to the post-lay burial operation which is aimed at cable installation cost reduction. For a certain period of time, how much the unburied cable laterally or even vertically displaces is checked when subject to the hydrodynamic lift and drag loading from the combined wave and current against the soil resistance and cable weight in water. In practice, the cable on-bottom stability may be preliminarily verified following the DNV-RP-F109 guidelines (2007) which have been developed specifically for pipeline installations in deep waters. Two relevant approaches are the absolute and generalized lateral stability which allow zero and certain (up to 10 diameters) displacements of the cable, respectively. In DNV-RP-F109, the specific weight of a pipe required for obtaining the stability is very much dependent on the pipe diameter. The design curves for the generalized stability are obtained from a large number of one-dimensional dynamic analyses, ie, on the flat seabed by neglecting bending and axial deformation of the pipe. In comparison with the pipelines (eg, 24e42 in. diameters and 489 kg/m submerged weight for the 24-in. pipe), the diameter and unit weight of the cables are much smaller (eg, up to 11e13/16 in. for export cables and 68 kg/m submerged weight for a 132-kV HVAC cable); thus, cables are more flexible. In addition, as wind farms are presently located in shallow waters, the effects of hydrodynamic loads on subsea cables are very significant. Depending on the weather and soil conditions, spatial and temporal movements of subsea cables are fairly complicated and it may be too conservative to achieve the stability requirement when using the pipeline design code. The dynamic stability method is needed to understand actual physical mechanisms based on the cableesoilefluid interaction modelling and dynamic investigation.

13.5.5

Vortex-induced vibration (VIV)

As flow passes a long cylindrical bluff body such as underwater cable, alternating vortices are naturally created in the wake of the cylinder and make the cylinder oscillate due to a resonance or so-called ‘lock-in’ condition between the vortex-shedding and structural natural frequencies. This lock-in, typically called ‘strumming’ for a cable (The Crown Estate, 2015b), takes place in a wide range of flow velocities, depending on the Reynolds number, structural-fluid mass and damping ratios, cable natural frequencies and mode shapes, amongst other parameters. For free-span static cables near the fixed foundations and free-hanging dynamic cables of floating foundations, the fatigue effect due to VIV can become detrimental due to the associated fluctuating lift and drag forces combined with the amplified mean drag components along the cable span. In practice, it is challenging to predict accurately the long-term quantitative damage effect from VIV due to the associated complicated fluidestructure interaction mechanisms, although several theoretical and experimental studies have been carried out in the past decades (Sarpkaya, 2004). Nevertheless, it is possible to check the likelihood


437

of VIV occurrence based on a certain set of parameters. Theoretically speaking, the vortex-shedding frequency in Hz (fs) can be found from the following relationship fs ¼ St

V D

[13.8]

where St is the Strouhal number (typically assumed to be 0.2), V is the flow velocity and D the cable outer diameter. To check the resonance (fs z fn), the lateral natural frequencies in Hz (fn) of the free-span cable may be approximated by n fn ¼ 2L

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffirffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi T n2 p2 EI 1þ 2 m þ ma L T

[13.9]

where n is the mode number, L the cable length, T the averaged tension, m the cable mass per unit length in water, ma the associated added mass and EI the bending rigidity. Note that Eq. [13.9] is most suitable for a taut cable (string or tensioned beam) with pinnedepinned boundary conditions. For a cable with significant sag-to-span ratio, more than 1/8 and/or different boundary conditions, a hybrid analyticalnumerical approach should be used to find natural frequencies and modes which can be classified as planar (in-plane of curvatures) and non-planar (out-of-plane) shapes (Srinil, 2004). These in-plane and out-of-plane modes are coupled due to combined cross-flow and in-line VIV excitations which would further complicate the fluidestructure interaction phenomena due to the effect of cable curvatures (Srinil, 2010). For the VIV analysis with a given V, a multimode response should be realized with regard to the lock-in condition. Other parameters that can influence the flow physics and cable VIV behaviours, but less understood in the literature, include the effect of oscillatory flows due to waves, current direction and variation along the water depth and cable span (Srinil, 2011), the angle of attack of the flow with respect to the inclined cable axis, the proximity to seabed and free surface, the marine growth and the floating platform motions in different degrees of freedom. In order to understand these individual or combined effects, experimental and computational fluid dynamics studies should be carried out along with the mathematical reduced-order model development tools (Srinil et al., 2009).

13.6

Outlook for offshore wind farm cables

Future offshore wind farm cables will be in deeper waters and remote areas farther from shore with a greater layout of interarray cables and longer distance for export cables to transfer a greater wind generation capacity from the larger (ie, above 5 MW) wind turbines. In particular, both wind turbine and subsea cable layouts will be more complicated and need to be well optimized with respect to the wind, wave and current directions apart from a proper route survey and site selection based on certain geotechnical conditions. Different novel models of floating wind turbine platforms and substations based on the tension leg, spar and semisubmersible concepts are

438

Offshore Wind Farms

potential solutions in deep waters (EWEA, 2015), and the ideal hybrid configurations of dynamic power cables with distributed buoyancy are considered using the design concepts similarly to those of flexible risers in oil and gas applications. Accordingly, there is a need to fully understand the non-linear dynamic behaviours of the free-hanging cables coupled with the moving floaters subject to different environmental wave and current conditions. The offshore wind industry will focus on several areas of potential cost reductions related to the offshore installation, operation and maintenance as well as the avoidance of risk, failure and transmission loss of subsea cables. Innovative offshore cabling technologies will be focussing on. • • • • • • • • • • •

the wet cable design with a cost-efficient and robust sheath prevention of water treeing and protection from extreme stresses, the development of powerful (eg, >525 kV) and cost-efficient HVDC export cables, the development and certification of interarray 66-kV cables as, eg, driven by the UK Carbon Trust’s Offshore Wind Accelerator (OWA) programme, the development of aluminium versus copper conductors for 66-kV cables, the standardized and optimal process of cable construction and installation, the efficient post-lay burial operations for multidirectional interarray cables, the cost-efficient simultaneous laying-burial techniques, the identification and optimization of cable burial depth for different soil properties, the pull-in operation with J-tubeless cable entry systems, the feasibility of using free-hanging interarray cables subject to energetic dynamic environments, and the robust cable protection and maintenance strategies without divers.

Health monitoring and conditioning technologies for subsea cables will also be of practical and industrial significance since any faults e typically unknown or difficult to distinguish e can lead to a breakdown of the cables, make a whole wind farm offline for several weeks (Cable fault shuts 400MW Anholt, 2015) and certainly lead to high insurance and repair costs. A pair of vessels may be needed to conduct the repair activities including the uncovering of the cable from the seabed and the cable lift up to the surface for the repair, which is much more difficult and lengthy than the on-land operation. Like other offshore infrastructure, the offshore repair will require several days of fairly calm weather conditions. Therefore, the reduction in cable repair and reinstallation times will significantly lead to the associated cost savings (The Crown Estate, 2015a).

Abbreviations ASZ

Advisory safety zone

BPI

Burial protection index

CBRA

Cable burial risk assessment

DoL

Depth of lowering


EPR

Ethylene propylene rubber

HVAC

High voltage alternating current

HVDC

High voltage direct current

MBR

Minimum bending radius

OWA

Offshore wind accelerator

ROV

Remotely operated vehicle

VIV

Vortex-induced vibration

XLPE

Cross-linked polyethylene

439

Acknowledgements The author would like to thank researchers at the University of Strathclyde, namely Xu Ji for drawings, Nguyen Dat and Hossein Zanganeh for gathering some information, and Bowen Ma for references.

References ABB launches world’s most powerful cable system for renewables, 2014. From: www. offshorewind.biz (retrieved 09.02.15.). Bai, Q., Bai, Y., 2014. Subsea Pipeline Design, Analysis and Installation. Elsevier. Beatrice Offshore Windfarm, 2011. Beatrice Transmission Works: Environmental Scoping Report Beatrice. Beindorff, R., van Baalen, L.R., 2013. Theory on submarine power cable protection strategies. In: Paper Presented at the EWEA Offshore. Berteaux, H.O., 1976. Buoy Engineering. Wiley. Burton, T., Jenkins, N., Sharpe, D., Bossanyi, E., 2011. Wind Energy Handbook. Wiley. Cable fault shuts 400MW Anholt, 2015. From: www.renews.biz (retrieved 23.02.15.). Carbon Trust, 2015. Offshore Wind Cable Burial Risk Assessment Methodology. Carbon trust launch race for next generation of offshore wind cables, 2013. From: www. carbontrust.com (retrieved 25.01.15.). Department for Business Enterprise & Regulatory Reform, 2008. Review of Cabling Techniques and Environmental Effects Applicable to the Offshore Wind Farm Industry. DNV, 2007. DNV-RP-F109 On-bottom Stability Design of Submarine Pipelines. DNV, 2012. DNV-RP-F401 Electrical Power Cables in Subsea Applications. DNV, 2014. DNV-RP-J301 Subsea Power Cables in Shallow Water Renewable Energy Applications. EWEA, 2015. The European Offshore Wind Industry e Key Trends and Statistics 2014. Ferguson, A., de Villiers, P., Fitzgerald, B., Matthiesen, J., 2012. Benefits in moving the inter-array voltage from 33 kV to 66 kV AC for large offshore wind farms. In: Paper Presented at the EWEA2012. FOREWIND, 2013. Dogger Bank Creyke Beck Wind Farm: Cable Details and Grid Connection Statement.

440

Offshore Wind Farms

Gevorgian, V., Hedrington, C.T., 2010. Submarine power transmission. In: Paper Presented at the U.S. Virgin Islands Clean Energy Workshop. University of the Virgin Islands. Greater Gabbard OWF’s Cabling Infrastructure Worries Local Fishermen, 2014. From: www. offshorewind.biz (retrieved 09.02.15.). Gwynt y Mor cabling complete, 2015. From: www.maritimejournal.com (retrieved 16.12.14.). Jenkins, A.M., Scutariu, M., Smith, K.S., 2013. Offshore wind farm inter-array cable layout. In: Paper Presented at the IEEE, Grenoble. Kaiser, M.J., Snyder, B.F., 2012. Offshore Wind Energy Cost Modeling: Installation and Decommissioning. Springer. Lewis wind farm backer pulls out over cable delay, 2014. From: www.bbc.co.uk (retrieved 23.02.15.). Mole, P., Featherstone, J., Winter, S., 1997. Cable protection e solutions through new installation and burial approaches. Dossier. http://dx.doi.org/10.3845/ree.1997.059. Offshore wind farms: Floating solutions for deep-water cables, 2010. From: www. powerengineeringint.com (retrieved 13.04.15.). Sarpkaya, T., 2004. A critical review of the intrinsic nature of vortex-induced vibrations. Journal of Fluids and Structures 19 (4), 389e447. Srinil, N., 2004. Large-amplitude Three-dimensional Dynamic Analysis of Arbitrarily Inclined Sagged Extensible Cables (Ph.D. dissertation). King Mongkut’s University of Technology Thonburi (KMUTT), Bangkok. Srinil, N., 2010. Multi-mode interactions in vortex-induced vibrations of flexible curved/straight structures with geometric nonlinearities. Journal of Fluids and Structures 26 (7e8), 1098e1122. Srinil, N., 2011. Analysis and prediction of vortex-induced vibrations of variable-tension vertical risers in linearly sheared currents. Applied Ocean Research 33 (1), 41e53. Srinil, N., Wiercigroch, M., O’Brien, P., 2009. Reduced-order modelling of vortex-induced vibration of catenary riser. Ocean Engineering 36 (17e18), 1404e1414. The Crown Estate, 2012. Submarine Cables and Offshore Renewable Energy Installations. The Crown Estate, 2015a. Offshore Wind Operational Report. The Crown Estate, 2015b. PFOW Enabling Actions Project: Sub-sea Cable Lifecycle Study. Tominaga, Y., Nakano, H., Koji, T., et al., 2014. Dynamic cable installation for Fukushima floating offshore wind farm demonstration project. In: Paper Presented at the AORC Technical Meeting. United States Department of Interior Bureau of Safety and Environmental Enforcement, 2014. Offshore Wind Submarine Cable Spacing Guidance. Worzyk, T., 2009. Submarine Power Cables: Design, Installation, Repair, Environmental Aspects. Springer. ˇ

Integration of power from offshore wind turbines into onshore grids

14

O.D. Adeuyi, J. Liang Cardiff University, Cardiff, United Kingdom

14.1

Introduction

In 1991, the first offshore wind farm to become operational was the 4.95-MW Vindeby project, which was located at a grid connection distance of 2.5 km from the shore of Denmark [1,2]. By 2014, around 8045-MW offshore wind capacity had been installed in the North Sea (63.3%), Atlantic Ocean (22.5%) and Baltic Sea (14.2%), and connected to the electricity grids of 11 European countries [3e5]. The installed offshore wind capacity in Europe is expected to increase to 23.5 GW by 2020 [6]. Wind farm collection systems gather the power generated from wind turbines and submarine power transmission systems transfer the electricity generated from offshore to onshore grids. This chapter describes the technologies of wind farm collection systems and explains the basic principles of submarine power transmission based on high-voltage direct-current (HVDC) technologies.

14.2

Wind farm collection systems

Wind farm collection systems use array cables to connect different wind turbines together, forming a string. Wind farm collection systems can be based on either alternating-current (AC) or direct-current (DC) technologies. AC collection systems are mature, while DC collection systems are a more recent development which are intended to occupy less space than AC collection systems.

14.2.1 AC collection systems Fig. 14.1 shows an AC collection system with array cables to collect the electricity produced from an offshore wind farm with permanent magnet synchronous generators (PMSGs). In this arrangement, the power output of each wind turbine generator is first rectified from AC to DC and then inverted from DC back to AC, using fully rated converters.


442

Offshore Wind Farms

Wind farm String 1 690 V AC

String n

WTG m

1.2 kV DC AC

String 2

690V / 33 kV DC

PMSG

DC

AC

Fully rated converters Wind turbine generator (WTG)

WTG 2

MVAC array cables

WTG 1

MVAC busbar

33 kV AC

To transmission system

Figure 14.1 AC collection system.

The array cables operate at a medium voltage (MV) of 33 or 66 kV AC and connect the wind turbine generators to an MVAC busbar system. The rated power, Pstring, for one string is around 35 MW for a 33-kV circuit. The number of strings, n, to be connected to a busbar is limited by the MV cable capacity, SMVA [7]: SMVA ¼

pffiffiffi 3Irated VLLrms

[14.1]

At a voltage, VLLrms of 33 kV, the rated current, Irated, is 2500 A and the downstream circuit capacity SMVA is: SMVA ¼

pffiffiffi 3 2500 A 33 kV ¼ 142 MVA

The number of strings, n, to be connected to the 33-kV busbar system is: n¼

SMVA 142 MVA z4 ¼ 35 MW Pstring

[14.2]

The number of wind turbines per string, m, depends on the rated power, Pwt, of individual wind turbines. If each wind turbine has a rated power of 5 MW, then the total number of wind turbines, m, in a string is: m¼

Pstring 35 MW ¼7 ¼ 5 MW Pwt

[14.3]

The total number of wind turbines, Tmn, connected to a busbar of the wind farm collection system is: Tmn ¼ m n ¼ 28

[14.4]


443

Increasing the voltage of the array cables from 33 to 66 kV can reduce electrical losses, increase possible transmission distance, increase the power transfer capacity for a given conductor size and result in fewer offshore AC substations. The 66-kV array cables would preserve the number of wind turbines in each string, m, of the collection system as the turbine ratings increase [8].

14.2.2 DC collection system A DC collection system is a concept that is intended to use direct-current array cables to gather the power generated from wind turbines. Each DC cable is to connect different wind turbines together in order to form a string. DC array cable systems can be built in medium voltages and low voltages [9,10].

14.2.2.1 Medium voltage Fig. 14.2 shows a medium-voltage DC (MVDC) array cable system, where the array cables operate at a medium voltage of 33 kV DC. In this arrangement, the generator output is first rectified using an AC/DC converter and then stepped up to a higher DC voltage, using a turbine DCeDC converter [10]. The DCeDC converters are to eliminate the need for offshore AC substations and platforms in the wind farm collection system and reduce project costs [11e13]. The turbine DCeDC converter in Fig. 14.2 uses three substages e an inverter, a transformer and a rectifier e to achieve a transformation ratio of 1.2 kV/33 kV. This will result in higher complexity and lower efficiency [10]. In the UK, it is expected that medium-voltage DC array cable demand will surpass MVAC array cable demand by 2020 [11].

14.2.2.2 Low voltage Fig. 14.3 shows a low-voltage DC (LVDC) array cable system, where the array cables operate at a low voltage of 1.2-kV DC. In this arrangement the output of the generator Wind farm String 1 690 V AC

1.2 kV / 33 kV

String 2

String n

WTG m

DC

AC PMSG

DC

DC

Wind turbine generator (WTG) DC

Turbine DC/DC converter

AC

AC

WTG 2

MVDC array cables

DC

Medium frequency transformer (1 kHz)

WTG 1

MVDC busbar

33 kV DC


Figure 14.2 Medium-voltage DC array cable system.

444

Offshore Wind Farms

Wind farm String 1 690 V AC

1.2 kV DC

String 2

String n

WTG m

AC PMSG

DC

Wind turbine generator (WTG)

WTG 2

LVDC array cables

WTG 1

1.2 kV DC

MVDC converter platform

DC DC

DC DC

MVDC busbar

33 kV DC


Figure 14.3 Low-voltage DC array cable system.

is rectified using one power conversion stage. LVDC strings are to connect the wind turbine generators to MVDC converters, which step up the collection voltage to 33 kV. The MVDC converters would be installed on offshore platforms.

14.3

Offshore wind power transmission systems

Offshore wind power transmission systems use submarine cables to transmit the electricity generated from offshore wind farms to land and interconnect AC grids of different countries. Fig. 14.4 shows three key offshore wind power transmission technologies. These are: medium-voltage alternating-current (MVAC), high-voltage alternating-current (HVAC) and high-voltage direct-current (HVDC). The transmission technologies are described in this section.

14.3.1

MVAC transmission

Fig. 14.4(a) shows an MVAC transmission system, in which medium-voltage export cables are used to connect wind farms, located at a transmission distance of up to 20 km, to onshore substations. The export cables operate at a medium voltage of 33 kV. The onshore substations use step up transformers to transform the medium voltage to the onshore grid voltage. It is the simplest arrangement for offshore wind power transmission [7].

14.3.2

HVAC transmission

HVAC technologies are mature and suitable for submarine power transmission at distances between 20 and 70 km. Fig. 14.4(b) shows an HVAC transmission system,


445

(c) Wind farm 3 String n WTG m

(b) Wind farm 2 String n WTG m

WTG 2 WTG 1

(a) Wind farm 1 String n WTG m

33 kV AC MVAC cables AC platform 3

WTG 2 WTG 1

HVAC cables

33 kV AC MVAC cables

WTG 2

AC platform 2

WTG 1 33 kV AC MVAC cables

HVAC cables < 20 km

HVDC platform 3

HVDC cables > 70 km

20–70 km

To onshore AC grid

Figure 14.4 Submarine power transmission technologies. (a) Medium-voltage alternating-current (MVAC). (b) High-voltage alternating-current (HVAC). (c) High-voltage direct-current (HVDC).

where AC array cables connect offshore wind turbines to offshore substations. The transformers of the offshore substation step up the collection voltage from 33 kV to a high voltage of 132 kV or above. HVAC export cables connect the offshore substations to onshore substations and transmit the total power generated to shore. They can operate at a high voltage of up to 245 kV (although more usually at around 150 kV) and use three-core crosslinked polyethylene (XLPE) cables. At transmission distances beyond 70e80 km and at a voltage of 150 kV and above, HVAC is not practical due to the capacitance and hence charging current of submarine cables.

14.3.3 HVDC transmission Fig. 14.4(c) shows an HVDC transmission system, where HVAC subsea export cables connect offshore AC substations to HVDC networks. Three key components of the HVDC networks are offshore converter platforms, submarine power cables and onshore converter stations. HVDC submarine power cables can also interconnect the grids of two or more countries, thereby creating an offshore grid.

14.3.3.1 Offshore converter platforms Two main components of an offshore converter platform are the topside and the foundation support structure. Topsides house the offshore HVDC converter stations.

446

Offshore Wind Farms

Foundation support structures host the topsides. There are three possible foundation support structures: fixed, mobile jack-up and gravity-base. Fixed platforms use jacket support structures which are attached to the seabed through piles. The topsides and jackets are installed by lifting from a barge using a heavy-lift crane vessel. A mobile jack-up platform has a self-installing topside which is mounted on a substructure. These topsides house offshore converter platforms which have an embedded jack-up system. The gravity-base platform consists of a topside welded to a gravity-base support (GBS) structure. These GBS platforms are constructed onshore, towed into position and secured to the seabed by their own weight and ballasting.

14.3.3.2 HVDC submarine export cables HVDC submarine export cables connect offshore converter platforms and onshore converter stations of the HVDC networks. HVDC submarine cables have a sheathed and armoured layer for protection against harsh conditions associated with offshore installation and service [14]. The two designs of HVDC subsea cables available on commercial terms are mass impregnated (MI) cables and extruded XLPE plastic cables. The insulation of MI paper cables consists of clean paper impregnated with a high-viscosity compound based on mineral oil. The next generation of MI paper cables would use paper polypropylene laminate as insulation to achieve ratings of 650 kV and 1500 MW per cable by 2020. A single-core MI paper cable could have conductor size up to 2500 mm2 and weigh about 37 kg/m [14]. XLPE cables rated up to 500 kV and 700 MW per cable were installed in the Skagerrak 4 project between Norway and Denmark in 2014. The 500-kV XLPE cables are to achieve a power rating of 1000 MW per cable by 2020 [15,16].

14.3.3.3 Onshore converter stations There are two main HVDC converter technologies: line-commutated converter (LCC) and self-commutated voltage source converter (VSC). LCC-HVDC is a mature technology and is suitable for long-distance bulk power transfers. The main power electronic device used in the LCC-HVDC scheme is the thyristor. LCC requires strong AC grids to be connected to both ends of the converter stations with large footprints for relatively low-order harmonic filters, hence they are not suitable for offshore wind power transmission [17]. VSC-HVDC is a more recent development and has independent control of active and reactive power, improved black start capability, and occupies less space than LCC-HVDC. Therefore, VSC-HVDC is the key technology for offshore wind power transmission.

14.4

Voltage source converters

This section describes the physical structure, operating characteristics and topologies of VSC-HVDC schemes.


447

14.4.1 Physical structure Fig. 14.5 shows the schematic diagram of a VSC-HVDC transmission scheme. The main components of a VSC scheme are the converter stations, phase reactors, AC filters and transformers.

14.4.1.1 Converter station VSCs use insulated gate bipolar transistors (IGBTs) to transform electricity from AC to DC at a transmitting station (rectifier) and from DC to AC at the receiving station (inverter) [18]. The IGBT is a three-terminal power semiconductor device which is controlled by a voltage applied to its gate. It is a fully controllable device and allows power flow in the ON state and stops power flow in the OFF state. Many IGBT cells can be connected in series to form an IGBT valve, increase the blocking voltage capability of the converter and increase the DC bus voltage level of the HVDC system [17e19]. The DC capacitors of the VSCs shown in Fig. 14.5 store energy, enable the control of power flow, provide a low inductance path for the turned-off current and reduce DC voltage ripple [17,19e21]. The DC side of the converter station A and B may be connected through a combination of DC subsea cables, underground cables and overhead lines [18,22]. Each converter station has a cooling system, auxiliary system and control system [22]. The rectifier station A is mounted on an offshore converter platform.

14.4.1.2 Phase reactors Phase reactors are connected in series between the converter bridge and the transformers of the VSC scheme as shown in Fig. 14.5. They create a voltage difference between the output of the converter bridge and the AC system. The alternating current flowing through the phase reactor controls active and reactive power of the VSCs [17,19,23]. Phase reactors also reduce high-frequency harmonic components of the alternating current.

14.4.1.3 AC filters VSCs can operate at a high switching frequency of 1 kHz and above and create high-frequency harmonic components in their output voltage. AC filters are connected

Offshore wind Phase farm reactor

AC filters

Converter station A

DC cables and overhead lines

Converter station B Transformer

DC capacitor

Figure 14.5 An offshore wind farm connected through a VSC-HVDC system.

Onshore AC grid

448

Offshore Wind Farms

in parallel between the phase reactors and the grid transformers to eliminate the high-frequency harmonic contents of the output voltage of the VSCs.

14.4.1.4 Transformers Transformers interface the AC system to the AC filters, phase reactors and converter stations and regulate the voltage of the AC system to a value that is suitable for the HVDC system [17,18,21].

14.4.2

VSC operating characteristics

VSCs create an AC voltage waveform on their output in order to exchange active and reactive power with another AC system. Fig. 14.6 shows the schematic diagram and phasor diagram of two AC voltage sources connected through a reactor. The AC voltage, Vout, at the sending end is generated by a VSC and the voltage, Vac, at the receiving end is the voltage of the AC system. Assuming that there are no power losses in the reactor, XL, shown in Fig. 14.6(a), and that the AC system connected to the AC filter is ideal, then the active power, P, the reactive power, Q, and apparent power, S, transferred through the VSC are: P¼

Vout sin d Vac XL

[14.5]

Q¼

Vout cos d Vac Vac XL

[14.6]

S¼

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi P2 þ Q2

[14.7]

where d is the phase angle between the voltage phasor Vout and Vac (in Fig. 14.6(b)) at the fundamental frequency, and is it usually called power angle.

14.4.3

VSC topologies

VSCs can use two-level, three-level and multilevel topologies. These VSC topologies are described in this subsection.

(a)

(b)

Receiving end

ΔV Vac

Vout

IL

Sending end

Imaginary part

XL

Vout ΔV 0

Vac Real part I

Figure 14.6 Two AC voltage sources connected through an ideal reactor (a) Schematic diagram (b) Phasor diagram.


449

14.4.3.1 Two-level Two-level VSCs use IGBTs valves to switch between the positive polarity and negative polarity of a charged DC capacitor [20,24]. Fig. 14.7 shows the circuit for one phase of a two-level VSC with the DC capacitor grounded at the midpoint. The two-level VSC is capable of generating an output voltage with two levels 1 /2 Vdc and 1/2 Vdc between the midpoint of the DC capacitor and the point ‘a’ shown in Fig. 14.7. Two-level VSCs use a pulse width modulation technique to control the magnitude and phase angle of their output AC voltage. They operate at a high switching frequency of 1 kHz and above, produce high-frequency harmonic components, have high switching losses and require AC filters at their output. Two-level VSCs use a special converter transformer with the capability to withstand high-voltage stresses due to the large DC voltage steps at the converter output. The total power losses of a two-level converter are about 1.6% at its rated transmission capacity [25].

14.4.3.2 Three-level The four different designs of three-level voltage source converters are neutral point clamped, T-type, active neutral point clamped and hybrid neutral point clamped [26]. Fig. 14.8 shows the circuit of one phase of a neutral point clamped converter. Three-level VSCs have the capability to generate an output voltage with three different voltage levels (1/2 Vdc, 0 and 1/2 Vdc) per phase between the point ‘a’ and a neutral point ‘0’ as shown in Fig. 14.8. The switching signals of their IGBT valves are generated using the PWM technique. They operate at a reduced switching frequency, have lower switching losses, and their transformers are exposed to lesser voltage stresses than the two-level converters.

14.4.3.3 Multilevel Multilevel converter designs are a more recent development with lower switching frequency, reduced power losses and reduced harmonic components than the two-level and three-level topologies of VSCs. The two types of multilevel converters Idc

Interface transformer

1 2 Vdc

Phase reactor

ΔV Iac

AC filter

Vac

IL

Vdc

a Vout

–1 V 2 dc

Figure 14.7 One phase of a two-level VSC.

IGBT valve

List of symbols Vdc : DC voltage with respect to ground Vout : AC voltage across IGBT stack ΔV : Voltage drop across phase reactor Vac : Voltage across AC filter Idc : Current through DC circuit IL : Current through phase reactor Iac : Current through AC filter a : Interface point between phase reactor and IGBT valves

450

Offshore Wind Farms

+

Phase reactor

a

1 V 2 dc

0

Diode valve

Vdc

+

–1 2

Vdc

IGBT valve

Figure 14.8 One phase of a three-level neutral point clamped VSC.

available on commercial terms are the modular multilevel converter (MMC) [27,28], and the cascaded two level (CTL) [29e31]. Fig. 14.9 shows the schematic diagram of a MMC. Each multivalve arm of the MMC consists of multiple submodules connected in series with an arm reactor. A submodule is formed by a DC capacitor, IGBTs and diodes. Each submodule is capable of producing a voltage step at its output. The submodules in each phase arm (shown in Fig. 14.9(b)e(d)) are switched in the correct sequence to synthesize a sinusoidal AC voltage at the converter output. The IGBTs of the submodules are turned on once every cycle during steady-state operation. The transformers of MMCs are not exposed to DC voltage stresses and can utilize a simple two-winding transformer (with star/delta connection) [20]. The arm reactors of the MMCs smooth the phase currents and limit the inrush current during capacitor voltage balancing and circulating currents between the phase arms during unbalanced operation [23].

14.4.3.4 Submodule circuits The three main types of switching circuits in the submodules of the MMCs are half-bridge, full-bridge and clamp double. The half-bridge circuit is the simplest design. It consists of two IGBTs with antiparallel diodes and a DC capacitor as shown in Fig. 14.9(b). The output voltage of the half-bridge circuit is either 0 or the DC capacitor voltage (Vc) [32] and current flows through only one IGBT during steady-state


+ +

Upper-arm voltage

SM1, a

SM1, b

SM2, a

SM2, b

SMlevel, phase

451

Types of submodule circuits

–

SMN, a

Vc + C –

+

SM2, c

VSM –

SMN, b

S1

SMN, c

S2

(b) Half-bridge

S1 Vc Vb

ic

ib

Va i a +

Lower-arm voltage

Arm reactors

Vdc

+ VSM

SMN+1, a

SMN+1, b

SMN+2, a

SMN+2, b

SMlevel+1, phase

–

SMN+2, c

Vc

SM2N, b

SM2N, a

Phase arm

SM2N, c

C – S4

S2

(c) Full-bridge

S1 Vc

–

S3 +

Idc

– VSM

+ C – S5

Vc

S3 + –

C

S4

S2

Multivalve arm Submodule (SM)

(a) Three-phase topology

(d) Clamp double

Figure 14.9 Schematic diagram of an MMC-HVDC Scheme (a) Three-phase Topology (b) Half-bridge submodule (c) Full-bridge submodule (d) Clamp double submodule.

operation. The half-bridge circuit has the lowest cost and the least conduction losses [17,20]. The full-bridge circuit has four IGBTs with antiparallel diodes and a DC capacitor as shown in Fig. 14.9(c) [20,32,33]. The voltage output of the full-bridge circuit is þVc, 0 or Vc and the current flows through two devices during steady-state operation. It has higher capital costs and increased conduction losses than the half-bridge circuit [34]. The clamp double circuit consists of two half-bridge designs connected in series. The positive terminal of one half-bridge is connected to the negative terminal of the other as shown in Fig. 14.9(d) [23,32,34]. It has five IGBTs with antiparallel diodes, two DC capacitors and two additional diodes. The voltage output of the clamp double circuit is 0, Vc or 2Vc and the current flows through three IGBTs during steady state operation [32e34]. The switch S5 is always in the ON state during normal operation and contributes only to conduction losses. The clamp double circuit has improved efficiency compared to the full-bridge circuit and has higher conduction losses than the half-bridge circuit [33,34].

14.4.3.5 Examples of VSC-HVDC projects Table 14.1 outlines some examples of existing VSC-HVDC transmission schemes with their converter topologies, ratings, application and commissioning year.

452

Table 14.1

Examples of existing VSC-HVDC schemes Ratings per converter

Project name (country)

Converter topology

Capacity (MW)

Voltage (kV)

Application

Year

Estlink (EstoniaeFinland)

Two-level

350

150

Electricity interconnection and grid reinforcement

2006

Borwin 1 (Germany)

Two-level

400

150

Connection of offshore wind farms

2009

Cross Sound (USA)

Three-level

330

150


2002

Murray Link (Australia)

Three-level

220

150


2002

Trans Bay (USA)

Modular multilevel

400

200


2010

Borwin 2 (Germany)

Modular multilevel

800

300


2013

Dolwin 1 (Germany)

Cascaded two-level

800

320


2015

Offshore Wind Farms

Information taken from C.-C. Liu, L. He, S. Finney, G.P. Adam, J.-B. Curis, O. Despouys, T. Prevost, C. Moreira, Y. Phulpin, B. Silva, Preliminary Analysis of HVDC Networks for Off-shore Wind Farms and Their Coordinated Protection (Online). Available: http://www.twenties-project.eu/node/18, 2011 (accessed 19.04.15); G. Justin, Siemens Debuts HVDC PLUS with San Francisco’s Trans Bay Cable, Living Energy (Online). Available: http://www.energy.siemens.com/hq/pool/hq/energy-topics/publications/living-energy/pdf/issue-05/Living-Energy-5-HVDCSan-Francisco-Trans-Bay-Cable.pdf, 2011 (accessed 18.11.14); D. Das, J. Pan, S. Bala, HVDC light for large offshore wind farm integration, in: 2012 IEEE Power Electronics and Machines in Wind Applications (PEMWA), 2012, pp. 1e7; ABB, DolWin1 (Online). Available: http://new.abb.com/systems/hvdc/references/dolwin1, 2013 (accessed 18.08.14).


14.5

453

Development of future submarine power transmission schemes

This section describes the development of offshore wind power transmission schemes. The key technologies for future submarine power transmission systems are low-frequency AC transmission, diode rectifiers with VSC inverters, multiterminal VSC-HVDC schemes and the Supernode concept.

14.5.1 Low-frequency AC transmission Low-frequency AC transmission is a concept that decreases the frequency of AC systems, reduces the cable charging current and extends the distance at which HVAC systems can be cost-effective [26,33]. This could increase the power transfer capacity of a given cable and reduce the number of subsea export cables. At the onshore station, frequency converters would be required to transform the low-frequency AC supply back to the frequency of the onshore grid [26]. There is a need to further research, develop and demonstrate this concept for submarine power transmission systems.

14.5.2 Diode rectifier and VSC inverter Fig. 14.10 shows an HVDC transmission scheme with a diode rectifier and a VSC inverter for grid connection of offshore wind farms [36e38]. The offshore diode-rectifier platform will have reduced costs, reduced power losses and occupy less space than existing offshore VSC platforms [38]. During start-up operation, an MVAC cable can be used to connect the offshore wind farm to an onshore AC grid, in order to create an AC voltage at the offshore grid. The cable will be disconnected during normal operation and is usually called an umbilical cable. Multiple diode cells are connected in series to increase the voltage withstanding capability of a converter. The diode valves are arranged into a bridge to form an uncontrolled rectifier. The HVDC transmission scheme shown in Fig. 14.10 consists of an offshore 12-pulse diode-rectifier and an onshore VSC station. It is expected that this new offshore converter design will be available on commercial terms by 2016 [38]. Offshore wind farm

VSC inverter

Diode rectifier HVDC cable

HVDC cable MVAC cable

Figure 14.10 Diode rectifier connected to a VSC inverter.

Onshore AC grid

454

Offshore Wind Farms

14.5.3

Multiterminal VSC-HVDC schemes

Multiterminal VSC-HVDC schemes are intended to facilitate the transfer of electricity generated from offshore wind farms to land, supply electricity to offshore oil and gas installations and interconnect the grids of adjacent countries. The polarity of a VSC does not change when the direction of power flow changes, hence, multiple VSCs can be connected to a DC bus with fixed polarity to form a multiterminal HVDC (MTDC) system [39e41]. The operation of MTDC grids requires at least one converter terminal to regulate the DC voltage [41]. Onshore converters will connect AC grids and other energy storage plants to DC grids. The onshore converter control will use a DC voltage active power droop control system to maintain the DC voltage and share equal power flow to the converters connected to the DC grid [39]. Offshore converter stations create an AC voltage with a fixed magnitude, frequency, and phase angle at the offshore grid, in order to absorb the active power generated by the wind turbines. However, the reliable operation of MTDC schemes will require high-power DC circuit breakers and direct current flow control devices, which are still being developed. Manufacturers of DC circuit breakers have announced the results of prototype tests in which direct current exceeding 3 kA was interrupted in less than 3 ms [42,43]. The next step is to deploy such DC breakers into real HVDC networks.

14.5.4

Supernode concept

The Supernode is a concept that is intended to eliminate the requirement for DC circuit breakers in multiterminal HVDC transmission. Fig. 14.11 shows a Supernode for offshore wind power transmission. It consists of an islanded AC network with multiple AC/DC converters. The converters of the Supernode would be required to have fault ride through capabilities and regulate the frequency and AC voltage of =

HVDC

+ – 320 kV

2 × 500 MW

Converter station HVAC

2 × 500 MW

1 GW = 1 GW + – 320 kV

=

400 kV AC hub

=

+ – 320 kV

1 GW

= 2 × 500 MW

1 GW 2 × 500 MW

+ – 320 kV

Figure 14.11 The Supernode concept.


455

the AC island [16]. Additional offshore converter platforms would be required to connect new HVDC circuits to the Supernode and this could result in high grid expansion costs and increased power losses.

14.6

Conclusions

The installed capacity of offshore wind farms in Europe is expected to increase from 8 GW in 2014 to about 23.5 GW in 2020. This chapter described the key technologies required for wind power collection and grid connection of offshore wind farms to onshore grids. Offshore wind farms will use both HVAC and VSC-HVDC submarine power transmission technologies to transfer the electricity generated from offshore wind farms to onshore grids. There is a need to further develop low-frequency AC transmission systems, DC circuit breakers for DC grids and the scheme comprising diode rectifiers and VSC inverters for grid connection of offshore wind farms.

References [1] EWEA, Offshore Statistics January 2008, 2009 (Online). Available: http://www. ewea.org/fileadmin/ewea_documents/documents/statistics/Offshore_Wind_Farms_2008.pdf (accessed 07.10.15). [2] B. Snyder, M.J. Kaiser, A comparison of offshore wind power development in Europe and the U.S.: patterns and drivers of development, Appl. Energy 86 (10) (October 2009) 1845e1856. [3] EWEA, Offshore Statistics 2014, 2015 (Online). Available: http://www.ewea.org/ statistics/offshore-statistics/ (accessed 07.10.15). [4] Global Wind Energy Council (GWEC), Global Offshore, 2015 (Online). Available: http:// www.gwec.net/global-figures/global-offshore/ (accessed 07.10.15). [5] EWEA, The European Offshore Wind Industry e Key Trends and Statistics 2014, 2015 (Online). Available: http://www.ewea.org/fileadmin/files/library/publications/statistics/ EWEA-European-Offshore-Statistics-2014.pdf (accessed 07.10.15). [6] Ernst & Young, Offshore Wind in Europe Walking the Tightrope to Success, 2015 (Online). Available: http://www.ewea.org/fileadmin/files/library/publications/reports/EYOffshore-Wind-in-Europe.pdf (accessed 07.10.15). [7] CIGRE TB 610, Offshore Generation Cable Connections, 2015. Paris. [8] BVG Associates, Offshore Wind: A 2013 Supply Chain Health Check, 2013 (Online). Available: http://www.bvgassociates.co.uk/Portals/0/publications/BVGATCE Offshore Wind SC Health Check 1311.pdf (accessed 28.03.15). [9] P. Lakshmanan, J. Liang, N. Jenkins, Assessment of collection systems for HVDC connected offshore wind farms, Electr. Power Syst. Res. 129 (December 2015) 75e82. [10] J. Pan, S. Bala, M. Callavick, P. Sandeberg, DC connection of offshore wind power plants without platform, in: 13th Wind Integration Workshop, 2014. [11] BVG Associates, Building an Industry 2013: Updated Scenarios for Industrial Development, 2013 (Online). Available: http://www.renewableuk.com/en/publications/index.cfm/ BAI2013 (accessed 28.03.15).

456

Offshore Wind Farms

[12] BVG Associates, UK offshore Wind Supply Chain: Capabilities and Opportunities, 2014 (Online). Available: https://www.gov.uk/government/uploads/system/uploads/ attachment_data/file/277798/bis-14-578-offshore-wind-supply-chain-capabilities-andopportunities.pdf (accessed 28.03.15). [13] W. Chen, A.Q. Huang, C. Li, G. Wang, W. Gu, Analysis and comparison of medium voltage high power DC/DC converters for offshore wind energy systems, IEEE Trans. Power Electron. 28 (4) (April 2013) 2014e2023. [14] ABB, HVDC MI Cables, 2014 (Online). Available: http://new.abb.com/systems/ high-voltage-cables/cables/hvdc-mi-cables (accessed 03.09.14). [15] National Grid, Electricity Ten Year Statement, 2012 (Online). Available: http://www. nationalgrid.com/NR/rdonlyres/DF56DC3B-13D7-4B19-9DFB-6E1B971C43F6/57770/ 10761_NG_ElectricityTenYearStatement_LR.pdf (accessed 08.06.15). [16] FOSG, Roadmap to the Supergrid Technologies, Final Report, March 2011 (Online). Available: http://mainstream-downloads.opendebate.co.uk/downloads/FOSG_ WG2_Final-report.pdf (accessed 31.10.12). [17] SKM, Review of Worldwide Experience of Voltage Source Converter (VSC) High Voltage Direct Current Technology (HVDC) Installations, 2013 (Online). Available: https://www.ofgem.gov.uk/ofgem-publications/52726/skmreviewofvschvdc.pdf (accessed 18.02.15). [18] R. Rudervall, J.P. Charpentier, R. Sharma, High voltage direct current (HVDC) transmission systems technology review paper, Energy Week 2000 (2000) (Online). Available: http:// www2.internetcad.com/pub/energy/technology_abb.pdf (accessed 20.04.15). [19] N. Flourentzou, V.G. Agelidis, G.D. Demetriades, VSC-based HVDC power transmission systems: an overview, IEEE Trans. Power Electron. 24 (3) (March 2009) 592e602. [20] N. MacLeod, Enhancing System Performance by Means of HVDC Transmission (Online). Available: http://www.pbworld.com/pdfs/pb_in_the_news/gcc_power_paper_a095.pdf (accessed 22.02.14). [21] C. Du, The Control of VSC-HVDC and Its Use for Large Industrial Power Systems, 2003 (Online). Available: http://publications.lib.chalmers.se/records/fulltext/11875.pdf (accessed 08.04.15). [22] Alstom Grid, HVDC e Connecting to the Future, First, Alstom Grid, Levallois-Perret Cedex, 2010. [23] C.-C. Liu, L. He, S. Finney, G.P. Adam, J.-B. Curis, O. Despouys, T. Prevost, C. Moreira, Y. Phulpin, B. Silva, Preliminary Analysis of HVDC Networks for Off-shore Wind Farms and Their Coordinated Protection, 2011 (Online). Available: http://www.twenties-project. eu/node/18 (accessed 19.04.15). [24] G. Justin, Siemens Debuts HVDC PLUS with San Francisco’s Trans Bay Cable, Living Energy, 2011 (Online). Available: http://www.energy.siemens.com/hq/pool/hq/ energy-topics/publications/living-energy/pdf/issue-05/Living-Energy-5-HVDC-San-FranciscoTrans-Bay-Cable.pdf (accessed 18.11.14). [25] ERDF, Irish Scottish Links on Energy Study (ISLES) Technology Roadmap Report, 2012 (Online). Available: http://www.scotland.gov.uk/resource/0039/00395552.pdf (accessed 18.02.15). [26] T.B. Soeiro, J.W. Kolar, The new high-efficiency hybrid neutral-Point-clamped converter, IEEE Trans. Ind. Electron. 60 (5) (May 2013) 1919e1935. [27] Siemens, High Voltage Direct Current Transmission e Proven Technology for Power Exchange, 2015 (Online). Available: http://www.siemens.com/about/sustainability/pool/ en/environmental-portfolio/products-solutions/power-transmission-distribution/hvdc_proven_ technology.pdf (accessed 31.0315).


457

[28] Siemens, The Sustainable Way: Grid Access Solutions Form Siemens, 2011 (Online). Available: http://www.energy.siemens.com/co/pool/hq/power-transmission/grid-accesssolutions/landingpage/Grid-Access-The-sustainable-way.pdf (accessed 21.04.15). [29] B. Jacobson, P. Karlsson, G. Asplund, L. Harnefors, T. Jonsson, VSC-HVDC transmission with cascaded two-level converters, in: Cigré 2010 Session, 2010 (accessed 20.04.15). [30] D. Das, J. Pan, S. Bala, HVDC light for large offshore wind farm integration, in: 2012 IEEE Power Electronics and Machines in Wind Applications (PEMWA), 2012, pp. 1e7. [31] ABB, Skagerrak, 2014 (Online). Available: http://new.abb.com/systems/hvdc/references/ skagerrak (accessed 18.08.14). [32] N. Ahmed, A. Haider, D. Van Hertem, L. Zhang, H.-P. Nee, Prospects and challenges of future HVDC SuperGrids with modular multilevel converters, in: Proceedings of the 14th European Conference on Power Electronics and Applications (EPE 2011), 2011, pp. 1e10. [33] J. Dorn, H. Gambach, J. Strauss, T. Westerweller, J. Alligan, Trans Bay Cable e a breakthrough of VSC multilevel converters in HVDC transmission, in: Cigré Colloquium e HVDC and Power Electronic Systems for Overhead Line and Insulated Cable Applications, 2012. [34] T. Modeer, H.-P. Nee, S. Norrga, Loss comparison of different sub-module implementations for modular multilevel converters in HVDC applications, in: Proceedings of the 14th European Conference on Power Electronics and Applications (EPE 2011), 2011, pp. 1e7. [35] ABB, DolWin1, 2013 (Online). Available: http://new.abb.com/systems/hvdc/references/ dolwin1 (accessed 18.08.14). [36] R. Blasco-Gimenez, S. A~no-Villalba, J. Rodriguez-D’Derlée, F. Morant, S. Bernal-Perez, Distributed voltage and frequency control of offshore wind farms connected with a diode-based HVdc link, IEEE Trans. Power Electron. 25 (12) (December 2010) 3095e3105. [37] S. Bernal-Perez, S. Ano-Villalba, R. Blasco-Gimenez, J. Rodriguez-D’Derlee, Efficiency and fault ride-through performance of a diode-rectifier- and VSC-inverter-based HVDC link for offshore wind farms, IEEE Trans. Ind. Electron. 60 (6) (June 2013) 2401e2409. [38] P. Menke, New grid access solution for offshore wind farms, in: European Wind Energy Association (EWEA) Conference, 2015 (Online). Available: http://www.ewea.org/ offshore2015/conference/allposters/PO208.pdf (accessed 20.04.15). [39] Friends of the Supergrid (FOSG), Roadmap to Supergrid Technologies e Update Report, 2014 (Online). Available: http://www.friendsofthesupergrid.eu/wp-content/uploads/2014/ 06/WG2_Supergrid-Technological-Roadmap_20140622_final.pdf (accessed 21.08.14). [40] N. H€orle, M. Asmund, K. Eriksson, T. Nestli, Electrical supply for offshore installations made possible by use of VSC technology, in: Cigré 2002 Conference, 2002. [41] T.M. Haileselassie, K. Uhlen, Power system security in a meshed North sea HVDC grid, Proc. IEEE 101 (4) (2013) 978e990. [42] Alstom, Alstom Takes World Leadership in a Key Technology for the Future of Very High Voltage Direct Current Grids, 2013 (Online). Available: http://www.alstom.com/ press-centre/2013/2/alstom-takes-world-leadership-in-a-key-technology-for-the-future-ofvery-high-voltage-direct-current-grids/ (accessed 22.04.13). [43] ABB, The High Voltage DC Breaker e The Power Grid Revolution, 2012 (Online). Available: http://www04.abb.com/global/seitp/seitp202.nsf/c71c66c1f02e6575c125711f00 4660e6/afefc067cd5a69c3c1257aae00543c03/$FILE/HVþHybridþDCþBreaker.pdf (accessed 22.04.13). ́


Energy storage for offshore wind farms

15

D.A. Katsaprakakis Technological Educational Institute of Crete, Heraklion, Greece

15.1

Introduction

15.1.1 The necessity of energy storage Storing goods is a common practice applied in several human and natural activities. It helps towards the optimum management of the required supplies for the implementation of a specific action or for regular daily activities. For example, the invention and the use of refrigerators enabled the storage and conservation of foods for longer periods, contributing considerably to the time and economic savings of our daily living. Energy storage is inherently present in animal and human bodies, which is critical for survival in harsh conditions. Energy storage is designed in manmade systems as well. The filling of the tanks of vehicles or central heaters enables the covering of long distances and the heating up of buildings for long time periods. The simplified examples reveal the importance of energy storage in technical and natural worlds. In electrical systems, the hugely important role of storage is predominantly provided in the form of reserves in power plants. The storage of electricity constitutes a significant procedure, especially in cases of high power production from renewable energy source (RES) power plants, such as wind parks. The nonguaranteed power production from wind parks and the wind turbines’ technical specifications can introduce significant problems (contingencies), especially in small autonomous systems or those of high wind power penetration. The wind turbines’ low tolerances in system events often leads to their tripping (power production cut) in case of a slight variation of the system’s frequency or voltage amplitude. This vulnerability can be the beginning of further generators’ sequential tripping, leading ultimately to a total blackout, especially in cases of weak and non-interconnected systems. To avoid this negative effect of wind parks in systems’ dynamic security, the wind power penetration is restricted up to a maximum percentage of the power demand. This percentage depends on the size of the system, the available spinning reserve of the thermal generators, the weather conditions and it is usually configured around 30% of the power demand [1e6]. The excess wind power is rejected. Moreover, the stochastic power production from wind parks cannot follow the power demand variation by itself adequately. This stochastic nature prevents the maximization of wind parks’ power production in small autonomous power systems, such as the insular ones. In these cases, although the low power demand can be totally covered from wind parks, this possibility is not feasible due to the non-guaranteed wind power availability. Offshore Wind Farms. http://dx.doi.org/10.1016/B978-0-08-100779-2.00015-5 Copyright © 2016 Elsevier Ltd. All rights reserved.

460

Offshore Wind Farms

The above-mentioned inadequacies can be handled with the introduction of storage power plants and their cooperation with the wind parks and other RES power plants. It has been proved through a number of electrical system simulations, that the dynamic security of the system is ensured with the support of a storage device, such as conventional electrochemical batteries, hydro turbines of a pumped storage system or the air turbines of a compressed air energy storage system [3,6e9]. Additionally, with the support of storage power plants, the stochastic power production from the wind parks can be adapted to the power demand, through the sequential charging and discharging procedure, whenever there are wind power surpluses or shortage exhibits, respectively. Actually, the combined operation of an RES power plant and a storage device, commonly described as a ‘hybrid power plant’, converts the stochastic power availability from the RES into guaranteed power production, enabling the power plant to follow adequately the varying power demands and approach high annual energy production penetration [10e14]. Even in large interconnected systems, the installation of high wind parks and the corresponding penetration of high non-guaranteed power can cause serious dynamic security problems. With the wind parks being the leading RES technology regarding the development of RES power projects in interconnected systems, the total installed RES power increases rapidly [15]. The importance of the introduction of electricity storage power plants in interconnected systems in turn rises. With the continuously reducing reserves of conventional fossil fuels and the simultaneous increase in the global energy consumption, the shift to the alternative RES seems to be an inevitable necessity. The combined introduction of wind parks and storage power plants in interconnected systems features as the unique solution to the target of the secure maximization of RES penetration in electrical systems and the eventual total substitution of fossil fuel consumption for power production.

15.1.2

A quick glance at former studies

A large number of former studies investigate the issue of the combined operation of RES power plants with storage power plants. The most famous among the investigated technologies seems to be the wind-powered pumped storage systems (WP-PSS). These systems aim to exploit local, renewable and environmentally friendly wind energy by improving the stability of the system and reducing the use of thermal power plants, minimising the consumption of fossil fuels, reducing the cost of electricity and boosting local economies. The most popular topic examined in existing papers is the introduction of a PSS in remote islands, to recover otherwise-rejected wind energy due to restrictions imposed for the systems’ stability and dynamic security [16e20]. The PSS, using a single penstock, produces power only during the power demand peak hours and helps to reduce amount of wind power rejected. A second approach extensively examined in the past combines the operation of wind parks and a PSS to produce guaranteed power during the power demand peak hours [10e14,21]. The guaranteed power is produced exclusively by the hydro


461

Frequency (Hz)

turbines. To maximise the wind energy stored by the PSS, a double penstock is used. The economic feasibility of these WP-PSS is strongly dependent on the feed-in tariff. A more revolutionary evolution of the above approaches is the introduction of WP-PSS in insular systems with high wind potential, aiming to maximise wind energy penetration [9,14,22,23]. Power production from these WP-PSS is not restricted to power demand peak hours but it is extended for the whole day. The high wind potential leads to annual wind energy penetration that can exceed 80%. Due to the large quantities of the produced electricity, the corresponding investments are very attractive, with lower sensitivity to the feed-in tariff. The introduction of WP-PSS has also been studied for interconnected power systems in Greece [24], Ireland [25] and Turkey [26]. A common conclusion of these studies is that these systems are necessary to achieve high wind energy penetration. Their operation can also be combined with base thermal generators leading to significant power demand peak shaving. The economic feasibility of such systems is more likely to be guaranteed, given the large quantities of the electricity demand in the interconnected systems. Finally, there were a number of studies that examined the contribution of a PSS to the improvement of the dynamic security of electrical systems [3,6e9]. Indicatively, the results from the simulation of the dynamic behaviour of the autonomous electricity system of Crete are presented in Figs 15.1 and 15.2. In these figures, the frequency variations after the loss of 80 MW of wind power are presented. The power demand at the time of the contingency is assumed to be 250 MW. In Fig. 15.1, the power production is supported with adequate spinning reserve from thermal generators, while in Fig. 15.2 the system is supported by the synchronized hydro turbines of a PSS. It can be seen that with the support of the hydro turbines the system recovers faster and exhibits lower permanent frequency deviation after the contingency. Consequently 50.1 50.0 49.9 49.8 49.7 49.6 49.5 49.4 49.3 49.2 49.1 49.0 48.9 48.8 48.7 48.6 0

10

20

30

40

50

60

70

80

90

100

110

Time (s)

Figure 15.1 Frequency variation after the loss of 80 MW of wind power in the existing power production system in Crete, with adequate spinning reserve from thermal generators.

462

Offshore Wind Farms

50.2 50.1

Frequency (Hz)

50.0 49.9 49.8 49.7 49.6 49.5 49.4 49.3 49.2 49.1 49.0 0

10

20

30

40

50

60

70

80

90

100

Time (s)

Figure 15.2 Frequency variation after the loss of 80 MW of wind power in the power production system in Crete, with the support of the synchronized hydro turbines of a pumped storage system.

it clearly exhibits improved behaviour compared to the one with support from the thermal generators’ spinning reserve. In addition to WP-PSS, other proposed technologies include windephotovoltaicbattery systems and windecompressed air systems (WCAES). The windephotovoltaic systems are usually proposed for small- or medium-sized decentralized production (eg small settlements [27e29]). The conventional lead batteries are generally not proposed for large-sized systems (power demand higher than 1 MW), taking into account their short life, the low storage capacities that can be achieved and the important environmental impacts from their disposal. Alternative battery technologies, such as the floating batteries, although they exhibit strongly improved technical features compared to the lead batteries, in terms of capacity and life time period, they still exhibit quite high procurement costs. The introduction of WCAES has also been widely studied. Several alternatives have been examined, such as the introduction of conventional [30,31] or adiabatic [32] CAES, the cooperation with flywheel [31] or pumped storage systems [33], etc. The thermodynamic simulation of such systems, the accurate calculation of their efficiency is another favourite issue of research [32]. Another research objective often met is the optimization of the WCAES in order to achieve the highest possible wind energy penetration [34,35] with the lowest possible cost [36]. A major drawback of the WCAES is the consumption of fossil fuels, which affects considerably both the environmental aspect of these systems and the economic efficiency of the required investments. For the second reason, the introduction of WCAES seems to be feasible in places with low-price fossil fuels (eg Canada [37], Australia [38] or the USA). In such cases the WCAES exhibit quite attractive economic features.


15.2

463

The storage technologies

Offshore wind parks are always power plants of some tens or hundreds of MWs of installed power. The installation of high nominal power is the only way to compensate for the increased set-up cost of the offshore wind parks, compared to onshore installations. The storage power plants required for such electricity quantities must exhibit a charging/discharging ability approximately equal to the wind park’s nominal power and a total energy capacity which can be between 1% and 3% of the total annual electricity production of the wind park, depending on the size of the wind park and the system that it is connected to, as well as the operational algorithm of the wind parkestorage plant station. This means that for an offshore wind park with a nominal power of 50 MW and a capacity factor of 30%, a storage capacity of about 1300 MWh is required. This, in turn, implies an effective capacity for the upper reservoir of a pumped storage system of about 1,700,000 m3 with a net head of 300 m. Although theoretically there may be several different storage technologies, the alternatives for such large storage power plants are rather restricted. Practically there are two available storage technologies suitable to manage the large electricity quantities produced from offshore wind parks: • •

Compressed air energy storage systems (CAESs) Pumped storage systems (PSSs).

These technologies are presented in the next subsections.

15.2.1 Compressed air energy storage systems CAESs are a method of energy storage through the compression of air. CAES are distinguished into two alternatives, conventional and adiabatic. For the time being, there are two conventional CAES systems operating, one in Neuen Huntorf, Germany and one in McIntosh, USA [39]. As far as adiabatic compressed air energy storage systems (AA-CAESs) are concerned, industrial applications were expected in approximately 2015 [40].

15.2.1.1 Conventional CAES The operation of a conventional compressed air energy storage system is presented in Fig. 15.3. Specifically, in this figure the operating algorithm of the existing CAES storage plant in Neuen Huntorf, Germany [41] is presented. Any potential electricity surplus is provided for a two-stage compressor with intercooling, that compresses ambient air up to 40e70 bar. The compressed air is then led to an after-cooler to keep its temperature close to ambient. Finally, the compressed and cool air is stored in an underground storage reservoir. When power is needed, the compressed air is heated-up by a combustion chamber in order to obtain increased power during the expansion process (expansion with reheating). In new CAES systems, the stored compressed air is preheated by a recuperator before it enters the combustion chamber. With the use of the recuperator, the total

464

Offshore Wind Farms

Ambient air Compressors LP

Turbines

Motor/generator

Heat

HP

M

HP

LP

Combustion chambers

Heat Compressed air storage

Exhaust

Natural gas

Figure 15.3 The structure of the existing conventional CAES storage plant in Neuen Huntorf, Germany.

efficiency of the storageeproduction cycle can be increased by 10%. On the other hand, a significant disadvantage of this alternative integration is the large size of the recuperator, which implies a considerable increase in investment [42]. The operating principle of this alternative, presented in Fig. 15.4, has been applied in the second existing CAES storage plant in McIntosh, Alabama, USA.

Ambient air Compressors

HP

LP

Heat

Turbines

Motor/generator M

HP

LP

Combustion chambers

Heat Compressed air storage

Exhaust

Natural gas

Recuperator

Figure 15.4 The structure of the existing conventional CAES storage plant in McIntosh, Alabama, USA, with the introduction of a recuperator.


465

Table 15.1 Fundamental technical specifications of the existing conventional CAES power plants Technical/economic data

Neuen Huntorf, Germany

McIntosh, Alabama, USA

Power (MW)

321

110

1160

2640

Cavern volume (m )

310,000 (2 caverns)

560,000

Storage maximum pressure (bar)

70

75

Turbines’ mass flow (kg/s)

416

154

Compressors’ mass flow (kg/s)

104

96

Set-up cost ($)

167,000.000

65,000,000

Set-up specific cost ($/kWh)

143.966

24.621

Set-up specific cost ($/kW)

520.25

590.91

Storage capacity (MWh) 3

In Table 15.1 the fundamental technical specifications of the existing two CAES systems are presented. For the compression of the incoming air, either axial compressors, achieving a pressure ratio of about 20 and a flow rate of 1.4 Mm3/h, or radial compressors, with flow rates up to 100,000 m3/h and a maximum compression pressure up to 1000 bar, can be used. With the currently available technology, air compression is executed in two stages with intercooling at temperatures from 40 to 200 C [43]. The arisen high-pressure airefuel mixture is expanded in air turbines with pressure ratios up to 22 and with a maximum inlet temperature of 1230 C. The storing of the compressed air at near-ambient temperature conditions allows higher density of the stored medium, reducing the required size of the storage reservoirs. For the storage of the compressed air, the aquifer, underground caverns made of high-quality rocks, depleted natural gas storage caves and salt domes are most commonly used, with storage capacities from 300,000 to 600,000 m3. Another feasible alternative is storage in underground, high-pressure pipes (20e100 bar).

15.2.1.2 Adiabatic CAES In an AA-CAES the heat released during air compression is stored in a separate heat storage reservoir. This is the main difference from a conventional CAES. With an AA-CAES the consumption of fossil fuels for the compresses is eliminated and this is one of the main reasons for the development of AA-CAES. The structure of an AA-CAES system is shown in Fig. 15.5. When a power surplus exists, air is compressed without intercooling and releases its heat in a separate heat storage reservoir before being stored. At discharge periods, compressed air is heated up to the appropriate turbine inlet temperature (600 C) by

466

Offshore Wind Farms

Ambient air Compressor

Motor/generator

Turbine

M

Exhaust

Heat Heat storage

Compressed air storage

Figure 15.5 Operating principle of an AA-CAES.

regaining the heat from the heat storage reservoir. Overall efficiency rates of adiabatic compressed air storage plants are expected to reach values of up to 70% [40,41,43], approaching the corresponding efficiency of a PSS. In two-stage AA-CAES systems, heat released in the low-pressure (LP) and high-pressure (HP) compressors is stored in separate heat tanks. At discharge periods, heat from the HP and LP heat tanks is regained before the inlet to the HP and LP turbines, respectively. Two-stage AA-CAES systems achieve higher energy storage density, which compensates for the increased complexity of the plant (two heat storage tanks and piping). Important advantages of the AA-CAES technology are the elimination of the fuel added before expansion in the turbine and of the concomitant CO2 emissions, as well as the compression of air without intercooling that allows for higher outlet temperatures from the compressor and, thus, higher amounts of heat stored in the heat tank. However, major components of the plant need to be redesigned as conventional components cannot be utilized. Specifically, heat storage tanks with capacities of 120e1800 MWhth need special design to achieve sufficiently high heat transfer rates and constant outlet temperature. Minimization of heat losses during charging and discharging of the heat reservoir is another point of consideration [44e46]. Regarding the compressor of the plant, in AA-CAES systems adiabatic compression is preferred to isothermal, which are adopted in conventional CAES plants. However, conventional


467

compressors cannot reach the high pressures and temperatures required for adiabatic compression (100 bar/620 C for single-stage and 160 bar/450 C for two-stage AA-CAES plants) and, along with the need for low response times and high isentropic efficiency, the design of novel compressors for AA-CAES systems becomes a necessity. Recent studies converge that to illustrate that meeting these requirements is best achieved by constructing a compressor consisting of three parts: (1) an axial or a radial compressor, as a low-pressure compressor in case of high or low air flow rates, respectively, and single-shaft radial compressors for the (2) intermediate- and (3) high-pressure sectors. The turbine sector needs to be redesigned to achieve increased turbine inlet temperatures, air flow rates and efficiency. In order to satisfy these requirements, a novel non-conventional regulation stage with lower losses should be designed for improved handling of pressure and flow rate fluctuations. Preheating of the turbine is also proved to be desirable in order to achieve temperature profiles, which will enable low response times [44e46].

15.2.2 Pumped storage systems 15.2.2.1 Basic concepts PSSs are the technically most mature and economically most competitive electricity storage technology for large power plants. With tens of PSS projects already constructed and operating worldwide under considerably different conditions, covering a power production range from 5 MW to 2 GW, huge experience has been gained regarding their technical specifications and operating procedures. The fundamental operating principle of a PSS is presented in Fig. 15.6. Two water reservoirs are constructed in two neighbouring geographical positions, with adequate altitude difference between them, usually some hundreds of metres. The water reservoir capacities can vary from some hundreds of thousands of cubic metres

Transformers and transmission lines

te Wa

Powerhouse Lower reservoir

te Wa

rw

he

n

r

e wh

r sto

nr

ing

e

ele

a

rg ne

g sin

y

Hydro turbines/pumps

Figure 15.6 Basic structure of a pumped storage system.

en

erg

y

Upper reservoir Penstocks

468

Offshore Wind Farms

to some millions of cubic metres. Water can be transferred between the two reservoirs with either a single or a double penstock. The lower edges of the penstocks are connected to a pump station and a hydro power plant. When there is a power generation surplus that must be stored, water is pumped from the lower reservoir and stored in the upper reservoir. In this way the available energy surplus is stored in the form of the gravitational energy. When power is needed, water is released from the upper reservoir and passes through the hydro power plant, thus providing the requested power. For the time being, tens of PSSs have been installed worldwide, combined with large thermal power plants, aiming at the so-called ‘power peak shaving’. For power peak shaving, power is stored during low power demand periods (usually night-time periods) in order to be available during power demand peak periods. In this way, cheap electricity produced during low power demand periods is stored, instead of being rejected, while the use of expensive generators (most commonly gas turbines) during power demand peak periods is avoided. Power peak shaving with the use of PSSs is usually applied in systems with large thermal or nuclear power plants, where the reduction of the total produced power during low demand periods from the large operating steam turbines is often not possible. For power peak shaving applications, the employed PSS is equipped with a single penstock, since simultaneous water pumping and falling, namely simultaneous power storage and production, is not sensible.

15.2.2.2 Wind-powered pumped storage systems The combined operation of a PSS exclusively with a wind park, although widely studied in earlier articles, has been applied in practice in only two cases. The first was on the island of El Hierro, in Canaria Archipelago, while the second one was on the Greek island of Ikaria, in the Aegean Sea. The WP-PSS is introduced in both cases aiming at the penetration maximization of the primary energy source, namely the wind energy, into the annual electricity production. The fundamental technical specifications of the introduced WP-PSS in both islands are presented in Table 15.2. The main scope of WP-PSS stations, usually named ‘hybrid power plants’, which is to maximize the wind energy annual penetration, introduces several peculiarities in the system’s design and operation. A major design innovation is the installation of a double penstock, enabling simultaneous water pumping and falling [23]. The necessity for the use of a double penstock arises from the fact that there is a maximum direct penetration percentage for the wind energy, for dynamic security reasons. The installation of a double penstock enables the storage of the wind power that cannot penetrate directly in the network, while the rest of the power demand can be covered by the hydro turbines at the same time, employing the water falling penstock. The double penstock also improves the system’s flexibility and its ability to react in cases of system contingencies. For example, in case of a sudden loss of wind power production, the exclusive falling penstock enables direct power production from hydro turbines, even if water was pumped when the contingency occurred.


469

Table 15.2 Fundamental technical specifications of the wind-powered pumped storage systems in El Hierro and Ikaria Technical/economic data

El Hierro, Spain

Ikaria, Greece

Power demand annual peak (MW)

13.3

7.8

Annual electricity consumption (MWh)

41,000

27,600

Wind park (units/power)

5 2.3 MW ¼ 11.5 MW

4 600 kW ¼ 2.4 MW

Upper reservoir effective capacity (m3)

380,000

900,000 (1st tank) 80,000 (2nd tank)

Lower reservoir effective capacity (m3)

150,000

80,000

Gross head (m)

655

724 (1st tank) 555 (2nd tank)

Storage capacity (MWh)

580

1500

Pump station (units/power)

2 1500 kW þ 6 500 kW ¼ 6 MW

8 250 kW ¼ 2 MW

Hydro power plant (units/power)

4 Pelton 2830 kW ¼ 11,32 MW

2 1550 kW þ 1050 kW ¼ 4.15 MW

Total set-up cost (V)

64,700,000

26,000,000

PSS set-up cost (V)

50,000,000

23,000,000

PSS set-up specific cost (V/kWh)

86.21

15.33

PSS storageeproduction cycle efficiency (%)

65

69

Annual wind energy penetration percentage (%)

80.0

50.0

The operating algorithm of a WP-PSS hybrid power plant is also completely different from that of a conventional PSS employed for power peak shaving. This philosophy is presented in Fig. 15.7 [23]. The power demand Pd is provided with power Pw by the wind park, at a certain time point. The wind park direct penetration is always restricted to a maximum value Pwp ¼ a$Pd (0 < a < 1), in order to ensure the system’s dynamic security. This is achieved with the introduction of pump loads for the excess wind power. Two cases are distinguished: 1. If the PSS upper reservoir is empty, the remaining power demand is covered by the existing diesel engine generators, that produce power equal to Pt ¼ Pd Pwp. The hydro turbines do not produce power, Ph ¼ 0. The PSS pumps are provided with the wind power surplus Pp ¼ Pw Pwp, in order for water to be stored in the PSS upper reservoir.

470

Offshore Wind Farms

Power demand Pp

Pd

Pwp

EMS Ph Pt

Wind parks

Diesel engines

Detail Pumps

Pw

P

T

Hydro turbines Detail

Figure 15.7 The operating philosophy of a wind-powered pumped storage system.

2. When the PSS upper reservoir isn’t empty, power demand is covered by the hydro turbines (Ph ¼ Pd Pwp). At the same time, any possible wind power surplus Pw e Pwp is stored through the water pumping penstock on the condition that the upper reservoir isn’t full. If this is not the case, no more wind power can be stored; this energy can be used for other applications such as hydrogen production or desalination, etc. Power from the diesel generators is null: Pt ¼ 0.

As it is revealed from the above analysis, thermal generators are only used as reserve units. The main power production unit is the wind park. The PSS is the storage unit. In case the scope of the WP-PSS is the power demand peak shaving, the guaranteed power production is restricted to the demand peak hours. The rest of the above operational algorithm remains the same. The electricity storage from offshore wind parks has not yet been studied widely. However, the basic principles do not differ from the ones met in onshore wind parks. A major difference between the two cases is the higher procurement and installation cost of the underwater connection cable in offshore installations. But this only affects the total efficiency of the overall investment, without any effect on the storage power plant technical considerations.

15.2.2.3 Seawater-pumped storage systems An alternative of distinguished importance is the use of seawater directly in the PSS and the utilization of sea as the lower reservoir. At the time of writing there was only one commercial S-PSS (Seawater PSS) constructed worldwide e in Okinawa, Japan e used for power peak shaving. Already operating for more than 10 years, the Okinawa S-PSS is an important source of experience for similar stations [47e49]. Seawater PSSs provide a valuable solution in cases of geographical territories with low annual rainfalls, since it guarantees the adequacy of the working means


471

in the PSS without affecting the freshwater reserves. These cases are often met in small islands or in geographical regions close to the equator. Since sea water is utilized as the PSS lower reservoir, such storage power plants must be installed on the coastline [50]. The land morphology close to the coastline highly affects the technical feasibility of the S-PSS, as well as the total set-up cost of the project. Small mountains or hills with absolute altitudes higher than 200 m and lower than 600 m are considered ideal for the installation of the PSS upper reservoir. A mild land morphology on the coastline (lack of cliffs) helps towards the elimination of the required earth works and minimization of the set-up cost for the hydro power plant and the pump station installation. Finally, the intensive slopes of the mountains or hills can perhaps impose the underground installation of the penstocks, with the construction of underground tunnels. Such works raise the total set-up cost significantly and can negatively affect the economic feasibility of the investment, especially in cases of small S-PSS power plants. In such cases, the existence of mountains or hills that are not steep also constitutes a fundamental prerequisite to ensure the economic feasibility of the hybrid power plant. The proximity of the S-PSS installation sites to the sea makes them an excellent prospect for storing electricity produced from offshore wind parks. The two power plants (offshore wind park and S-PSS) can be connected to a common substation, built near the coastline, constraining, on the one hand, the network connection cost and enabling, on the other hand, flexible operation of the overall hybrid station, without any interference to the rest of the utility network or other existing power plants. In the following section, two characteristic case studies of wind-powered S-PSSs will be presented, one large and one small.

15.3

Indicative case studies: S-PSSs in Rhodes and Astypalaia

In this section the studies of two WP-PSSs on the islands of Rhodes and Astypalaia are presented. Each presented system aims at the maximization of wind power penetration in isolated insular power systems. Each one of the presented power production systems consists of a wind park (an offshore and an onshore one) and a PSS. The PSS reservoirs are connected to each other with a double penstock. The construction of two penstocks, one exclusively for water fall and one exclusively for pumping water, enables flexible operation of the PSS and maximizes the station’s contribution to the maximization of the wind energy penetration, as well as to the system’s stability and dynamic security. The presented PSSs utilize the sea as the lower reservoir and work on seawater so that the water supply is guaranteed even for areas with relatively low annual rainfalls. The proximity of the PSSs to the sea seems also to be an ideal choice for storing electricity from offshore wind parks.

472

Offshore Wind Farms

15.3.1

Siting of the S-PSSs

The installation sites for the examined S-PSSs in Rhodes and Astypalaia were selected after thorough examinations in several areas that should satisfy certain morphological prerequisites, such as: • • • • • • •

Proximity to the coastline; An appropriate area for the upper reservoir, namely a flat area or physical cavities with adequate land (depending on the reservoirs’ capacity) and at least 150-m altitude above sea level; A gradual slope from the upper reservoir position to the coastline where the penstock will run (no cliffs, gaps, gorges or canyons must be met along the penstock routes); The ratio of the penstock length (L) over the absolute altitude difference between the penstock’s ends (H) must not exceed the value of 5; Adequate land (about 30,000 m2) next to the coast for the installation of the hydrodynamic stations; The absence of other activities in the area (eg tourism) that could raise negative reactions from the local community against the construction and the operation of the S-PSS; Site accessibility by land and sea.

Satisfying certain criteria in the above areas will help to justify the technical and economic feasibility of the S-PSS as the technical works required are minimized and, consequently, the investment for set-up is lower. Apart from the above, the selection of the appropriate sites also depends on the size of the S-PSS which, in turn, depends on the size of the non-interconnected electricity system. In Table 15.3, the characteristic power demands on the islands of Rhodes and Astypalaia are presented. From Table 15.3 it can be seen that Rhodes is a large isolated power system, while the system on Astypalaia is a small one. The proposed S-PSSs will cooperate with wind parks, one offshore and one onshore, aiming at maximizing wind power penetration. Therefore, the size of the S-PSS is defined by the power demand scale. The size of the S-PSS is also defined by the necessity to improve the economic feasibility of the project (the investment’s economic indices improve with the size of the S-PSS). The required energy-storing capacity arises from the daily power production required and it determines the capacity of the upper reservoir (taking into account

Table 15.3 Main electricity demand features in the islands of Rhodes and Astypalaia Rhodes (2011)

Astypalaia (2013)

Maximum annual power demand (MW)

176.40

2.25

Minimum annual power demand (MW)

38.70

0.31

Total annual energy consumption (MWh)

789,168.37

6670.11

Mean daily energy consumption (MWh)

2162.11

18.27


473

the absolute altitude of the reservoir). An initial estimate, based on experience, is an upper reservoir of 4,500,000 m3 at 160 m altitude and of 300,000 m3 at 300 m altitude for the S-PSS in Rhodes and Astypalaia, respectively. Following the above parameters, the overall positioning of the S-PSS, including the upper reservoir, the penstock route and the hydrodynamic stations (pump station and hydro power-plant), is presented in Figs 15.8 and 15.9.

3,983,000

3,982,500

3,982,000

838,500

838,000

837,500

837,000

836,500

836,000

3,981,500

Figure 15.8 The PSS upper tank position and the penstock route in Rhodes S-PSS on the Greek Geodetic Coordinate Reference System. 4,054,000

Isobath 20m

N W

E

Hydro power plant & pump station

S

4,053,500

55.0%

Upper reservoir

On-shore wind park

Penstock route

4,053,000

S2

S1

S3

S4 Existing road New road 719,500

719,000

718,500

718,000

717,500

4,052,500

Figure 15.9 The PSS upper tank position and the penstock route in Astypalaia S-PSS on the Greek Geodetic Coordinate Reference System.

474

Offshore Wind Farms

On the right-hand side of Fig. 15.8, particularly, the position of the PSS upper reservoir in the southwest part of the island of Rhodes is presented. On the same map, the siting of the wind turbines of an offshore wind parks is also presented. However, thorough description of this task is beyond the scope of this chapter. On the left-hand side of Fig. 15.8, detailed siting and design of the upper reservoir is depicted. Analytical description of this particular drawing is provided in Section 15.3.2. The S-PSS characteristic features, defined by the land morphology at the installation sites, are presented in Table 15.4. Observing Figs 15.8 and 15.9 and the data presented in Table 15.4, leads to the fundamental prerequisites set for the selection of the installation sites being fulfilled on both occasions. Once the sites for the S-PSS components have been selected, the positioning of the components and the calculation of the required works are performed. They are presented in the following sections.

15.3.2

The design of the reservoirs

Topographical views of the reservoirs are presented in Figs 15.8 and 15.9. Some important issues regarding the two reservoirs of the S-PSS stations in Rhodes and Astypalaia are presented below: 1. The reservoir site in Astypalaia is a hilltop, while the reservoir site in Rhodes is a physical cavity. In both areas there is enough land available to satisfy the reservoirs’ required capacities. 2. The reservoir basin in Astypalaia will be created with excavation works that lend the form of an inverted truncated cone. The limestone formations are tectonic with joint and fissures systems, karstified at places, making the excavations in the upper area of the artificial reservoir relatively easy, through the use of pile drivers and/or explosives. 3. In Rhodes, the area selected for the PSS upper tank is a valley which can be made into a reservoir by constructing two dams. No additional digging will be required. 4. A disadvantage of the site selected in Rhodes is its low absolute altitude (160 m). The PSS’s head height is maximized by using the sea as the PSS’s lower reservoir. 5. The reservoirs in Rhodes and Astypalaia follow the contours of 160 and 340 m, respectively, above sea level, which is the outer contour found in the flat terrain of the site; the maximum available flat area is occupied. The incline dip is set at 3:1, following the relevant bibliography for artificial dam and tank construction [51,52].

Characteristic features of the S-PSS, defined by the land morphology

Table 15.4

Island

Rhodes

Astypalaia

160

350

Available land for the upper reservoir (m )

450,000

40,000

Distance of upper reservoir from the coast L (m)

750

1110

Ratio L/H

4.69

3.17

22.68

20.48

Upper reservoirs’ sites absolute altitude H (m) 2

Average inclination of penstock route

( )


475

Topographic maps of the reservoir installation sites, provided by the Hellenic Military Geographical Service, are digitized and the reservoirs are designed thoroughly on a digitized terrain of contours with 4-m height difference. The volume of the required excavation work, the reservoir capacities and depth, etc., are calculated and the results are presented in Table 15.5. It can be seen from Table 15.5 that the physical cavity found in Rhodes reduces the volume of excavation works required. This is not the case for the reservoir in Astypalaia, where the reservoir will be formed by excavating, raising the S-PSS set-up cost. In both occasions the required reservoir capacities are achieved. In fact, the capacity of the resulting reservoir in Astypalaia is much higher than that initially required. This was later found to be a major benefit since it allows for the contribution of the WP-PSS in controlling power production and the improvement of the system’s security [1,2,6,7,9]. Furthermore, the excess energy storage capacity enables the combined operation of the WP-PSS with a desalination plant for potable

Table 15.5

Characteristic features of the PSSs’ upper tanks

Island

Rhodes

Astypalaia

5,107,924

269,238

4,554,257

255,648

1787

219

553,667

13,590

414,997

30,503

Bottom area (m )

424,540

31,676

Upper surface altitude (m)

160

348

Bottom altitude (m)

145

333

Maximum depth (m)

15

15

1:3

1:3

0

484,980

81,975

e

SE dam’s volume (m )

304,064

e

NW dam’s total length (m)

220

e

SE dam’s length (m)

387

e

NW dam’s maximum height (m)

20

e

SE dam’s maximum height (m)

40

e

Reservoirs’ overall characteristics Gross capacity (m3) 3

Effective capacity (m ) Energy storage capacity (MWh) 3

Minimum water volume in reservoir (m ) 2

Upper surface area (m ) 2

Inner incline slope 3

Total digging volume (m ) Dams NW dam’s volume (m3) 3

476

Offshore Wind Farms

water production. This perspective is of crucial importance for areas with low annual rainfalls, such as the islands under examination. In both cases, higher reservoir capacities will be justified by the impending interconnection of the islands with the mainland power system. In that case, the hybrid performance of the projects will not be used; the wind farms will be directly connected to the grid, while the hydro part will perform as a simple pumped storage system. To prevent leakage of seawater from the upper reservoir, the technology applied in the Okinawa S-PSS was adopted [47e49]. This technology is presented in Fig. 15.10. An ethylene propylene diene monomer (EPDM) rubber sheet has been adopted for the lining of the upper pond. The EPDM has been proven, via a number of tests, to exhibit excellent material properties and weather-resistance characteristics. For the lining structure, a drainage layer will be constructed using gravel materials (20 mm or less). The layer will be 50 cm thick across the entire surface of the pond. On this layer, a cushioning material to prevent damage from the angular parts of gravels will be laid using a non-woven spun bonded fabric of polyester. Then, an EPDM sheet of 2.0-mm thick will be installed as a surface material. The span of sheet anchors was set at a standard of 8.5 m (on slopes). If damage to the sheet occurs, seawater leakage will be detected by seawater sensors and pressure gauges, both of which are installed in the pipes connected to the drainage layer in each zone. The detector will emit an alarm to indicate seawater leakage, and at the same time the pump will recharge leaked seawater to the upper pond. This system prevents seawater from leaking into the neighbouring environment. Furthermore, since the rubber sheet is the top layer of the lining structure, it can be repaired easily. The water intake from the reservoir will flow through an intake tower at the bottom of the reservoir, as shown in Fig. 15.11. The dimensions shown in Fig. 15.11 refer to the upper reservoir in Astypalaia. The tower’s entrance will be covered with a filter grid, to prevent debris from entering the penstock. The tower’s height is selected to ensure that a minimum water volume will always be stored in the upper reservoir.

Lining structure using rubber sheets Cover sheet (EPDM) Rubber sheet (EPDM)

Filling concrete

Cushion fabric

Drainage laver Precast sheet anchor

Drainage pipe

Inspection gallery

Figure 15.10 The technology applied in the upper reservoir to prevent seawater leakage.


477

348 m

3m 335 m

Water intake tower from the upper reservoir to the underground tunnel

1.5 m

333 m

Figure 15.11 The water intake from the upper reservoir.

This is to prolong the lifetime of the reservoir’s bottom sealing, by avoiding the direct exposure of the sealing materials to the solar radiation. Finally, the soil and dredging spoil from the excavation works will be used to raise an embankment around the reservoirs, to prevent the spread of seawater from the wind to the surrounding environment. The height of the embankment must be at least 2 m.

15.3.3 Construction of the penstock The routes of the penstocks were chosen with the following criteria: • • •

the minimization of the penstock lengths; the avoidance of intensive slopes and cliffs; the land’s morphology where the penstock reaches the sea must be gradual.

For both S-PSSs, an underground tunnel from the water intake position in the reservoir to the other side must be dug. This necessity arises from the morphological conditions created by the construction of the upper reservoir in flat or concave areas and the formation of the reservoir with excavation works. The above can be seen in Fig. 15.12, where the sectional vertical view of the penstock’s route from the reservoir to the sea is presented for the S-PSS in Astypalaia. In Fig. 15.12, the penstock follows the underground tunnel for, approximately, the first 164 m from the reservoir. Excluding that first part, the rest of the penstock is positioned on the surface of the mountain-side. The same configuration applied for the Rhodes S-PSS, where the length of the tunnel is 237 m. The major issue with the penstock construction is the selection of a corrosionresistant material, suitable for the transportation of seawater. An excellent material is glass reinforced polyester (GRP). The chemical structure of this material is not affected by seawater. Moreover, it exhibits a very low flow losses coefficient (approximately 0.030). It is lighter and cheaper than steel, hence it can be transported and

478

Offshore Wind Farms

Absolute altitude (m)

400 A. Water supply

Underground toute B. End of tunnel

300

Overground route

200 C. Hydrodynamic machines station

100

0

0

100

200

300

400

500

600 700 800 900 1000 Distance in horizontal level (m)

1100

1200

1300

1400

1500

Figure 15.12 Sectional vertical view of the penstock’s route from the reservoir to the sea for the S-PPS in Astypalaia.

installed more easily than steel pipes, significantly reducing the project set-up cost. On the other hand, the ability of GRP to withstand hydrostatic pressure restricts the use of GRP tubes (nominal pressure of GRP tubes decreases with increasing diameter). In the case of Astypalaia, the required nominal diameter is 1.50 m. For that diameter, GRP tubes are constructed with nominal pressures lower than 32 bar. As the head height is 348 m, it implies that the penstock can be constructed with GRP tubes for altitudes between 40 and 348 m. For altitudes below 40 m the penstock will be constructed from St52, exhibiting a yield point of 330 MPa. In order to protect the steel tubes from seawater corrosion, a thick film of mixed phenol and epoxy resins, without dissolver, will be applied in the inner surface of the tubes. In Rhodes, the water flow required for the PSS’s operation was calculated as 111.23 m3/s for water fall and 66.69 m3/s for water pumping. Given that the maximum diameter of commercially available steel pipes in Greece is 2540 mm (100 in.), two sets of 20 parallel pipelines are used for the water fall and pumping penstocks. The use of GRP tubes in this case is not possible, because of the large required diameters. The maximum pressure in the penstocks is 27.59 bars (16.1 bars hydrostatic pressure plus 11.49 bars) due to the hydraulic hammer effect for instant flow cut. The minimum wall thickness of 70 steel tubes/100 in. diameter is given as 12.70 mm with a corresponding nominal pressure of 44 bars, adequate for the specific installation. The construction of the penstock is designed as presented in Tables 15.6 and 15.7 for Rhodes and Astypalaia, respectively. The selection of the different pipe thicknesses (except for the necessity to withstand hydrostatic pressure and hydraulic hammer) also aims at the minimization of cost. Special digging works and techniques (eg cut and cover) are required in cases of abrupt changes in the landscape along the route of the penstock as well as close to the coast line, aiming at creating secure passage and grounding of the penstock and its protection against the corrosive conditions caused by the combination of seawater and high winds. Presented in Fig. 15.13 is a 3D view of the mountain slope with the penstock route for Astypalaia S-PSS. The figure shows the length of the different penstock sections, namely different nominal pressures. The underwater suction pipeline and the position of the hydrodynamic stations are also shown.


479

Table 15.6 The analysis of the construction of the PSS penstocks in Astypalaia Maximum hydrostatic pressure (bar)

Nominal pressure (bar)

Route’s length (m)

Total tubes’ length (m)


Material

348e300

GRP

4.8

6

323

646

300e240

GRP

10.8

12

382

764

240e200

GRP

14.8

16

188

376

200e160

GRP

18.8

20

157

315

160e120

GRP

22.8

25

138

276

120e40

GRP

30.8

32

201

402

40e0

Steel X70

34.8

44

95

190

1485

2969

Total route’s and tube’s length (m)

A similar 3D view of the penstock route in Rhodes is presented in Fig. 15.14. The ratio of the head height, H, over the penstock length, L, is calculated at 4.24 for Astypalaia and 5.48. These values (lower than or close to 5) are quite satisfactory, positively affecting several parameters regarding the construction of the penstock and the operation of the PSS, such as the pipeline diameter and thickness required, the water flow linear losses, the PSS total efficiency and finally the penstock set-up cost and economic indices of the investment.

15.3.4 The hydrodynamic machine stations and the suction pipeline The hydrodynamic stations, namely the pump station and the hydro power plant are going to be built onshore, next to the sea. The fundamental prerequisites that the sites must fulfil are: • • •

buildings must be protected against the sea as waves may reach several metres height during the winter period, taking into account the strong winds blowing in the Aegean Sea; the absolute altitude of the hydro power plant must be as low as possible, to maximise the head height; the pump suction level must be below sea level to allow natural water flow from the sea to the pump suction side.

To meet the above requirements, the hydro turbines and the pumps will be installed in two different buildings. Three-dimensional views of the final positioning in the islands of Astypalaia and Rhodes are shown in Figs 15.13 and 15.14, respectively.

480

Table 15.7

The analysis of the construction of the PSS penstocks in Rhodes Material

Nominal (external) diameter (mm)

Wall thickness (mm)

Nominal pressure (bar)

Route’s length (m)

Total tubes’ length (m)

144e8

Steel X70

2540

12.70

44

856

17,120

144e1

Steel X70

2540

12.70

44

877

17,540


Falling penstock Pumping penstock

Penstock

Offshore Wind Farms


481

Legend 6 bar GRP 12 bar GRP 16 bar GRP 20 bar GRP 25 bar GRP 32 bar GRP 44 bar St X70

333 m

320 m

280 m

240 m

200 m

160 m

120 m 80 m 40 m 0m

Figure 15.13 A 3D view of the penstock route for the Astypalaia S-PSS. a. 144

m

b. 144 m

c. 120 m

d. 80 m e. 40 m

Hydro power plant Pumps’ station

Figure 15.14 A 3D view of the penstock route for the Rhodes S-PSS.

In Rhodes, a flat coastal area of adequate land is found where the penstock reaches the coast (Fig. 15.14). The construction of the pump station and hydro power house including the accompanied works for the coastline is straightforward. On the contrary, in Astypalaia, where the penstock reaches the coastline, the land is steep and is exposed to the erosive marine environment so it is more susceptible to collapses and land slips (see Fig. 15.13).

482

Offshore Wind Farms

Water suction from the sea to the pump station can be accomplished in two ways: •

•

With the construction of a breakwater structure using precast concrete blocks. This technique was adopted in Okinawa S-PSS [47e49]. The main disadvantages of this method are the high construction cost and the visible changes to the natural landscape from the technical works. An alternative is the installation of a long pipeline along the sea bed, starting from the pump station and ending where the sea depth is 15e20 m. The pump station is constructed below sea level to ensure natural water flow through the underwater pipeline. This technique exhibits much lower set-up cost than the first and the visible changes to the natural landscape are minimal.

The second method was selected for both S-PSS. The underwater pipeline extends into the sea until depths greater than 15 m are reached (see Fig. 15.15 for Astypalaia). At these depths, stresses to the water suction structure attributed to waves on the surface are negligible. Moreover, the seawater remains relatively clear, free of any underwater debris or waste (eg sand, algae, small stones) as these are swept away by the underwater streams, reducing the probability of such objects entering the pipeline. The underwater suction pipelines will be buried 0.5e1.0 m under the sea bed. The pipeline water inlet will be covered with filter grids, to prevent objects from entering the water inflow. In both S-PSSs examined, GRP tubes with nominal pressure of 6 bars will be used for the pipeline. In Astypalaia’s S-PSS, a single suction pipeline of 1.50-m inner diameter is required, while in Rhodes’ S-PSS, 20 parallel suction pipelines of 2.00-m inner diameter are required. The length of the underwater pipelines is determined by the seabed morphology to ensure that water suction takes places at depths greater than 15 m for the reasons mentioned above. In Astypalaia’s S-PSS, the 20-m isobath is found at 92 m from the coast, while in Rhodes, the 20-m isobath is found at 350 m from the coast.

0m

–18 m

1.50

–20 m

Figure 15.15 The beginning of the underwater suction pipeline in Crete S-PSS.


483

As mentioned earlier, the pump suction level must be below sea level to ensure the natural inflow of water from the sea. By applying Bernoulli’s law and taking into account the pipelines’ length and inner diameter, the suction geostatic height (20 m in both cases), the required water flows (3.33 m3/s for each pipeline in Rhodes and 0.69 m3/s in Astypalaia) and the GRP material flow losses coefficient ( f ¼ 0.029), the suction level in both pump stations is calculated to be 1 m below sea level. A sectional view of the pump station in Astypalaia is shown in Fig. 15.16. The pump station building will be 15 m from the coastline, to protect it from the waves. The hydro power plant building will be constructed next to the pump station. A sectional view from the hydro power plant building in Astypalaia is shown in Fig. 15.17. At both sites, the positioning of the power plant is 10 m from the coastline to protect the building from the waves. This determines the absolute altitude above sea level of the hydro turbines and, consequently, the total geostatic head height for power production from the S-PSS. A water disposal canal of reinforced concrete will lead the water to the sea after its passage through the hydro turbines.

15.0

15.0

5.00 0m –1.0 m 6.00 1.50

Figure 15.16 Vertical sectional view of the pump station in Astypalaia S-PSS. 15.00 17.1 m

6.80 10.0 m

10.00 1.80 6.5 m

2.00

17.80

1.80

0m

Figure 15.17 Vertical sectional view of the hydroelectric power plant in Astypalaia S-PSS.

484

15.3.5

Offshore Wind Farms

Hydrodynamic machines

Pelton hydro turbines and multistage pump models are selected for the operation of the examined S-PSS. The Pelton model is selected because it exhibits constant and high efficiency for 90% of the output power range, low cost, robust construction and it allows an increase in power production within a few seconds. The last feature is very important regarding the power system’s stability and dynamic security. Single-staged pump models were selected for both PSSs. The pump’s power regulation will be performed with a cyclo-converter based on thyristors, to follow the available variable power production from the wind park, as well as to avoid the weak system’s security and stability events caused by abrupt variation of the pumps’ load. Modern voltage source converters (VSC) based on IGBTs can also be used in some other systems in the MW range, and they provide the additional capability of reactive power control for the grid network, in addition to the control functionalities for starting and variable speed operation. Two parallel horizontal-axis Pelton hydro turbines, each of nominal 2 MW, will be installed in Astypalaia’s S-PSS, giving a total hydro power of 4 MW. Taking into account that the maximum power demand in 2013 was 2.55 MW, there will always be a hydro power surplus of 1.45 MW. This power surplus can be exploited as power production spinning reserve, contributing to the system’s frequency regulation and to the improvement of the dynamic security. The guaranteed power will be produced 24 h/day. Similarly in Rhodes, 20 parallel horizontal-axis Pelton hydro turbines, each of nominal power 8 MW, will be installed, giving a total hydro power of 160 MW. Taking into account that the maximum power demanded from the hydro turbine generators is 113.90 MW, there will be a hydro power surplus of 46.1 MW. This surplus capacity can also be used as a power production spinning reserve, contributing to the system’s frequency regulation and to an improvement of the dynamic security. The maximum water flow of each hydro turbine is 5.56 m3/s. For the S-PSS in Astypalaia, the required pumped water flow can be provided by a combination of four parallel pump units, each of 842 kW nominal axis power. The selected pump model is a single-stage pump with horizontal axis. The total pumped water flow required (66.69 m3/s) can be provided by combining 134 parallel single-stage pump units, each of 1074 kW nominal motor power (absorbed power 934 kW). The total maximum demand for electric power is 143.9 MW. The nominal hydrodynamic efficiency is given 85.0%. The runners of both Pelton models are constructed of stainless steel grade G-X5CrNi13.4Mo. The needle tips and the nozzle tip wearing rings are replaceable, and are also constructed from stainless steel. In both pump stations, the selected pump models are developed for reverse osmosis desalination plants. The pump’s shaft, the impeller, the suction stage, the casing, the diffuser and the pressure enclosure are constructed from duplex steel.


485

15.3.6 Wind parks A peculiarity observed in Rhodes is the relatively low wind potential met onshore, giving a capacity factor usually lower than 25% while the capacity factor in other Aegean islands is higher (>40%). Contrary to the above, higher wind potential is found offshore from Rhodes, especially off the southwest coast of the island (capacity factors of w35%). Although an offshore wind park requires higher set-up costs, the wind conditions are improved and planning restrictions are reduced, and as a result the payback time for the investment can be reduced. For these reasons, in Rhodes the wind park will be installed offshore. A wind turbine of high nominal power needs to be selected due to the following reasons: • •

limited space available for the installation of the wind park due to high sea depths in Rhodes; approximately 150e200 MW of wind power is required, as predicted by the maximum annual power demand in 2011, presented in Table 15.3, and the capacity factor at the installation site (the final value just lower than 35%).

A wind turbine model of 5 MW nominal power has been selected. The positioning of the wind turbines (rotor diameter of 126 m) southwest of Rhodes is presented in Fig. 15.8. The wind turbines are positioned in-line perpendicular to the main wind direction. Downwind turbines are placed half way between the two wind turbines in front of it. The distance between two wind turbines on the same line is 5D ¼ 630 m. The distance between two lines of wind turbines is 7D ¼ 882 m. The minimum distance from the coast is 300 m. The maximum depth of installation is 60 m, to avoid higher installation costs related to increasing foundation costs. Following these guidelines, 35 wind turbines are positioned, providing 175 MW. With the restrictions discussed above, the limited offshore area available and the requirement for power defined the density of the wind turbine installation. This results in significant shading losses which vary between 1.36% and 15.56%. In Astypalaia the small size of the power demand decreases the requirements for the installed wind power. A small wind park of four 900-kW nominal power wind turbines, giving a total nominal power of 3.6 MW, is proved to be adequate for the introduced WP-PSS. The wind park is installed onshore, on a hill close to the S-PSS installation site. The siting of the wind turbines, presented in Fig. 15.9, is almost perpendicular to the predominant wind direction, giving low shading losses (from 0.11% to 2.11%).

15.3.7 Annual energy productions and storage The dimensioning of the examined systems is presented in Table 15.8. Given the size of the WP-PSS, as presented in Table 15.8 and by simulating the systems’ annual

486

Table 15.8

Offshore Wind Farms

Results of the systems’ dimensioning Rhodes

Astypalaia

Wind park nominal power (MW)

175.00

3.60

Hydro turbines nominal power (MW)

160.00

4.00

Pumps nominal power (MW)

143.92

3.54

111.23

0.85

Maximum pumping flow (m /s)

66.69

0.77

Falling penstock minimum diameter (m)

7.20

1.50

Pumping penstock minimum diameter (m)

5.60

1.50

Number of parallel falling pipelines/nominal diameter (mm)

20/2540

1/1500

Number of parallel pumping pipelines/nominal diameter (mm)

20/2540

1/1500

3

Maximum falling flow (m /s) 3

operation, the annual energy produced and stored can be calculated; the results are shown in Table 15.9. The percentage of the direct wind power penetration to the guaranteed power production, defined in Section 15.2.2.2 is set equal to 0 in Astypalaia and 50% in Rhodes.

Table 15.9 Annual energy production and storing e wind park capacity factors Rhodes

Astypalaia

Wind park energy penetration (MWh)

223,501.56

0.00

Hydro turbines energy production (MWh)

177,886.96

6076.08

Total RES energy production (MWh)

401,388.52

6076.08

Thermal generators energy production (MWh)

387,779.88

594.03

Total stored energy (MWh)

271,880.75

11,551.22

Wind park rejected energy (MWh)

24,250.67

530.22

Wind park total produced energy (MWh)

519,632.98

12,081.44

Wind park rejected energy percentage (%)

4.67

4.39

Wind park’s final capacity factor (%)

33.90

38.30

PSS total annual efficiency (%)

65.43

57.56

Annual wind energy penetration (%)

50.86

91.09


487

Furthermore, Table 15.9 presents additional information from the operation of the WP-PSS as follows: •

percentage of wind energy rejected Ewr over the total annual wind energy production Ew

wr ¼ •

Ewr Ew

wind park capacity factor, calculated as the annual wind energy penetration Ewp over Pw$T where Pw is the nominal power of the wind park and T the annual time period

cf w ¼ •

Ewp Pw $T

[15.2]

PSS annual efficiency, calculated as Eh the annual electricity production from the hydro turbines over Est the annual energy stored

nPSS ¼ •

[15.1]

Eh Est

[15.3]

the annual RES penetration to the electricity production, calculated as the total wind energy penetration plus hydro energy generation over Ed the total annual electricity consumption.

pRES ¼

Ewp þ Eh Ed

[15.4]

The small size of the power demand in Astypalaia and the remarkable available wind potential enables an annual RES penetration to the electricity production of higher than 91%. In Rhodes, the high power demand, compared to the lower wind potential and the limited available space for the wind park installation, gives an annual RES penetration of up to 50%.

15.3.8 Economic results This section presents some fundamental economic results regarding the corresponding investments for the construction of the S-PSS in Rhodes and Astypalaia. The S-PSS construction cost, presented in Table 15.10, is calculated based on the systems’ dimensioning and site-specific positioning. The total construction cost per kilowatt is calculated over the PSS guaranteed power (160 MW in Rhodes and 4 MW in Astypalaia). The effect of the project’s size on the cost per kilowatt is clearly shown. Following the Greek legislation regarding the operation of WP-PSSs, the investments’ economic evaluation is performed, giving the economic indices over the equities as presented in Table 15.11. The annual revenues of the projects are based on the vending of the produced electricity from the WP-PSS, namely either the hydro

488

Offshore Wind Farms

Table 15.10

WP-PSS set-up cost calculation Set-up cost (V)

No

Set-up cost component

Rhodes

Astypalaia

1

Wind park

350,000,000

3,960,000

2

Hydro power plant

60,000,000

2,200,000

3

Pumps station

90,000,000

1,700,000

4

Upper reservoir

18,500,000

3,500,000

5

Penstocks

45,000,000

1,300,000

6

New roads construction

1,000,000

2,100,000

7

New utility network

30,000,000

400,000

8

Several infrastructure works

5,000,000

500,000

9

Secondary electromechanical equipment

5,000,000

500,000

10

Consultants fees

2,000,000

500,000

11

Several other costs

5,000,000

500,000

Total set-up cost

611,500,000

17,160,000

Total PSS set-up cost

213,500,000

8,700,000

Total set-up specific PSS cost (V/kW)

1334

2175

Total set-up specific PSS cost (V/kWh)

119.47

39.73

Table 15.11

Economic indices calculated over the investments’ equity Rhodes

Astypalaia

Net present value e N.P.V. (V)

356,052,917

10,298,045

Internal rate of return e I.R.R. (%)

13.73

31.42

Payback period (years)

6.35

3.10

Discounted payback period (years)

7.89

3.65

Return on equity e R.O.E. (%)

83.23

441.05

Return on investment e R.O.I. (%)

332.90

110.26

turbines or the wind energy direct penetration. The electricity vending prices, according to the existing legislation, are configured on the basis of the electricity production cost of existing thermal power plants of the insular systems. These prices are set as equal to 0.25 V/kWh for Rhodes (existing electricity production cost 0.26 V/kWh) and 0.37 in Astypalaia (existing electricity production cost 0.40 V/kWh).


489

The indices presented in Table 15.11 prove the economic feasibility of the examined investments.

15.4

Conclusions

The storing of energy is a procedure of ultimate importance for the normal and cost-effective operation of power systems. In the case of non-guaranteed power plants, such as wind parks, the importance of storage increases further due to the necessity to adapt the stochastic nature of the power production to the inelastic power demand. The necessity of wind power storage becomes more intensive in cases of non-interconnected systems, imposed by the sensitive dynamic security requirements and the restricted wind penetration possibilities often found in such systems. Offshore wind parks constitute a special category of wind park, characterized of increased set-up cost compared to the onshore installations. To compensate this disadvantage, offshore wind parks with high nominal power are usually designed and developed, in order to increase the annual electricity production and the corresponding revenues of the investment. The large quantities of electricity production from offshore wind parks imply the introduction of respectively adequate storage power plants. The available technologies for large power storage plants are the PSSs and the CAESs. PSSs are the only power storage technology with tens of different installations around the world. The operation of such power plants since the early 1960s provides considerable experience regarding their construction and operation. PSSs are already considered a technically matured and well-established technology. PSSs exhibit high efficiency of the storageeproduction cycle, usually higher than 65%. The specific set-up cost of PSSs varies from 1300 to 2500 V/kW of guaranteed power or from 40 to 120 V/kWh of storage capacity, depending on the size of the storage plant. In most cases PSSs constitute an economically competitive technology, capable of providing the initially stored electricity with prices lower than the existing specific power production system. CAES is a rather new technology, although the first CAES plant was installed in Germany in the late 1970s (1978). However, since then only one more CAES plant has been installed in the USA in the early 1990s (1991), leaving considerably less experience about the operation of such systems. Most probably, the alternative storage technology, ie PSS, with better economic and efficiency characteristics, was the main obstacle to the development of more CAES storage plants. CAES power plants exhibit lower efficiency than the PSSs, and the efficiency is currently around 40e50%. The set-up cost of the two existing CAES plants is estimated at around 750e800 V/kW of guaranteed power or from 30e200 V/kWh of storage capacity. The alternative of adiabatic CAES (AA-CAES) is estimated to cause a 10% increase to the overall efficiency of the CAES, with a related increase in the set-up cost. A special category of PSS is the ones that use seawater. At time of writing there is only one seawater PSS (S-PSS) constructed worldwide, which is in Okinawa, Japan,

490

Offshore Wind Farms

used for power peak shaving. Already operating for more than 10 years, the Okinawa S-PSS is an important source of experience for similar stations. Two case studies were presented, one small and one large, for the cooperation of seawater PSS with an offshore and an onshore wind park. Seawater can be pumped directly from the sea, thus construction of a lower reservoir is avoided, compensating for the higher costs arising from the use of corrosion-resistant materials for certain components. The proximity of the S-PSS installation sites to the sea makes them an excellent prospect for storing electricity produced from offshore wind parks. The two power plants (offshore wind park and S-PSS) can be connected to a common onshore substation, reducing on the one hand the network connection cost and enabling, and on the other hand allowing more flexible operation of the overall hybrid station, without any interference of the rest of the utility network or other existing power plants. Special issues regarding the use of seawater in PSSs, such as the use of materials for the construction of the penstock, the construction of the upper reservoir, placing the pump station and the hydro power plant on the coast and the selection of pump and hydro turbine models are presented thoroughly. The presented cases studies proved the economic and technical feasibility of seawater PSSs cooperating with wind parks, from small power plants (power demand in the range of 2.5 MW) to large ones. A fundamental parameter affecting the economic feasibility of the storage plant is the price of the produced electricity, which must be defined according to the existing specific electricity production cost of the existing power production system. If this prerequisite is fulfilled, the economic feasibility of the PSS is usually guaranteed.

Abbreviations AA-CAES

Adiabatic compressed air energy storage systems

CAES

Compressed air energy storage systems

EPDM

Ethylene propylene diene monomer

GRP

Glass reinforced polyester

HP

High pressure

LP

Low pressure

PSS

Pumped storage systems

RES

Renewable energy sources

S-PSS

Seawater pumped storage systems

WCAES

Wind-compressed air systems

WP-PSS

Wind-powered pumped storage systems


491

References [1] Katsaprakakis DAl, Papadakis N, Christakis DG, Zervos A. On the wind power rejection in the islands of Crete and Rhodes. Wind Energy 2007;10:415e34. [2] Daoutis LG, Dialynas EN. Impact of hybrid wind and hydroelectric power generation on the operational performance of isolated power systems. Electr Power Syst Res 2009;79: 1360e73. [3] Slootweg JG, Kling WL. The impact of large scale wind power generation on power system oscillations. Electr Power Syst Res 2003;67:9e20. [4] Papathanassiou SA, Boulaxis NG. Power limitations and energy yield evaluation for wind farms operating in island systems. Renewable Energy 2006;31:457e79. [5] Hatziargyriou N, Papadopoulos M. Consequences of high wind power penetration in large symposium, September 18e19, 1998, autonomous power systems. Proceedings of CIGRE Neptum, Romania. [6] Katsaprakakis DAl. Maximisation of wind power penetration in non-interconnected power systems [Doctoral thesis]. National Technical University of Athens; March 2007 [in Greek]. [7] Karapidakis E. Contribution of artificial intelligence to the estimation of the dynamic security of autonomous electricity systems in real time [Doctoral thesis]. National Technical University of Athens; 2003. Gallachoir BP, McKeogh EJ. Techno-economic review of existing and new [8] Deane JP, O pumped hydro energy storage plant. Renewable Sustainable Energy Rev 2010;14: 1293e302. [9] Katsaprakakis DAl, Christakis DG. Maximisation of RES penetration in Greek insular isolated power systems with the introduction of pumped storage systems. In: European wind energy conference and exhibition; 2009. p. 4918e30. EWEC 2009 7. [10] Caralis G, Rados K, Zervos A. On the market of wind with hydro-pumped storage systems in autonomous Greek islands. Renewable Sustainable Energy Rev 2010;14:2221e6. [11] Kapsali M, Anagnostopoulos JS, Kaldellis JK. Wind powered pumped-hydro storage systems for remote islands: a complete sensitivity analysis based on economic perspectives. Appl Energy 2012;99:430e44. [12] Ding H, Hu Z, Song Y. Stochastic optimization of the daily operation of wind farm and pumped-hydro-storage plant. Renewable Energy 2012;48:571e8. [13] Dinglin L, Yingjie C, Kun Z, Ming Z. Economic evaluation of wind-powered pumped storage system. Syst Eng Procedia 2012;4:107e15. [14] Katsaprakakis DAl, Christakis DG. A wind parks, pumped storage and diesel engines power system for the electric power production in Astypalaia. In: European wind energy conference and exhibition; 2006. p. 621e36. EWEC 2006 1. [15] REN21. Renewables 2014 global status report. Paris: REN21 Secretariat; 2014, ISBN 978-3-9815934-2-6. [16] Kaldellis JK, Kapsali M, Kavadias KA. Energy balance analysis of wind-based pumped hydro storage systems in remote island electrical networks. Appl Energy 2010;87: 2427e37. [17] Kapsali M, Kaldellis JK. Combining hydro and variable wind power generation by means of pumped-storage under economically viable terms. Appl Energy 2010;87:3475e85. [18] Anagnostopoulos JS, Papantonis DE. Simulation and size optimization of a pumpede storage power plant for the recovery of wind-farms rejected energy. Renewable Energy 2008;33:1685e94.

492

Offshore Wind Farms

[19] Bueno C, Carta JA. Wind powered pumped hydro storage systems, a means of increasing the penetration of renewable energy in the Canary Islands. Renewable Sustainable Energy Rev 2006;10:312e40. [20] Islam SM. Increasing wind energy penetration level using pumped hydro storage in island micro-grid system. Int J Energy Environ Eng 2012;3:1e12. [21] Katsaprakakis DAl, Christakis DG, Pavlopoylos K, Stamataki S, Dimitrelou I, Stefanakis I, Spanos P. Introduction of a wind powered pumped storage system in the isolated insular power system of KarpathoseKasos. Appl Energy 2012;97:38e48. [22] Katsaprakakis DAl, Christakis DG, Voumvoulakis E, Zervos A, Papantonis D, Voutsinas S. The introduction of wind powered pumped storage systems in isolated power systems with high wind potential. Int J Distrib Energy Resour 2007;3:83e112. [23] Katsaprakakis DAl, Christakis DG. Seawater pumped storage systems and offshore wind parks in islands with low onshore wind potential. A fundamental case study. Energy 2014; 66:470e86. [24] Anagnostopoulos JS, Papantonis DE. Study of pumped storage schemes to support high RES penetration in the electric power system of Greece. Energy 2012;45:416e23. [25] Tuohy A, O’Malley M. Pumped storage in systems with very high wind penetration. Energy Policy 2011;39:1965e74. [26] Dursuna B, Alboyaci B. The contribution of wind-hydro pumped storage systems in meeting Turkey’s electric energy demand. Renewable Sustainable Energy Rev 2010;14: 1979e88. [27] Hessami M-A, Campbell H, Sanguinetti C. A feasibility study of hybrid wind power systems for remote communities. Energy Policy 2011;39:877e86. [28] Rehman S, Alam MdM, Meyer JP, Al-Hadhrami LM. Feasibility study of a windepve diesel hybrid power system for a village. Renewable Energy 2012;38:258e68. [29] Underwood CP, Ramachandran J, Giddings RD, Alwan Z. Renewable-energy clusters for remote communities. Appl Energy 2007;84:579e98. [30] Mason JE, Archer CL. Baseload electricity from wind via compressed air energy storage (CAES). Renewable Sustainable Energy Rev 2012;16:1099e109. [31] Zafirakis D, Kaldellis JK. Autonomous dual-mode CAES systems for maximum wind energy contribution in remote island networks. Energy Convers Manage 2010;51: 2150e61. [32] Zhao P, Dai Y, Wang J. Design and thermodynamic analysis of a hybrid energy storage system based on A-CAES (adiabatic compressed air energy storage) and FESS (flywheel energy storage system) for wind power application. Energy 2014;70:674e84. [33] Kim YM, Shin DG, Favrat D. Operating characteristics of constant-pressure compressed air energy storage (CAES) system combined with pumped hydro storage based on energy and exergy analysis. Energy 2011;36:6220e33. [34] Abbaspour M, Satkin M, Mohammadi-Ivatloo B, Hoseinzadeh Lotfi F, Noorollahi Y. Optimal operation scheduling of wind power integrated with compressed air energy storage (CAES). Renewable Energy 2013;51:53e9. [35] Wang SY, Yu JL. Optimal sizing of the CAES system in a power system with high wind power penetration. Int J Electr Power Energy Syst 2012;37:117e25. [36] Madlener R, Latz J. Economics of centralized and decentralized compressed air energy storage for enhanced grid integration of wind power. Appl Energy 2013;101:299e309. [37] Ibrahim H, Younes R, Ilinca A, Dimitrova M, Perron J. Study and design of a hybrid wind-diesel-compressed air energy storage system for remote areas. Appl Energy 2010;87: 1749e62.


493

[38] Hessami M-A, Bowly DR. Economic feasibility and optimisation of an energy storage system for Portland Wind Farm (Victoria, Australia). Appl Energy 2011;88:2755e63. [39] Karellas S, Tzouganatos N. Comparison of the performance of compressed-air and hydrogen energy storage systems: Karpathos island case study. Renewable Sustainable Energy Rev 2014;29:865e82. [40] Meyer F. Integration von regenerativen Stromerzeugern. 05/2007. Druckluft-Speicherkraftwerke, Projektinfo. [41] Crotogino F. Compressed air storage. In: Internationale Konferenz “Energieautonomie durch Speicherung Erneuerbarer Energien”. Hannover: KBB Underground Technologies GmbH; Oktober 2006. p. 30e1. [42] Ibrahim H, Ilinca A, Perron J. Energy storage systems e characteristics and comparisons. Renewable Sustainable Energy Rev 2008;12:1221e50. [43] N€olke M. Compressed Air Energy Storage (CAES) e Eine sinnvolle Erg€anzung zur Energieversorgung? Promotionsvortrag 2006. [44] Jakiel C. Entwicklung von Großdampfturbinen, W€armespeichern und Hochtemperatur e Kompressoren f€ur adiabate Druckluftspeicherkraftwerke. 5. dena-EnergieForum. Berlin: Druckluftspeicherkraftwerke; September 8, 2005. [45] Zunft S, Tamme R, Nowi A, Jakiel C. Adiabate Druckluftspeicherkraftwerke: Ein Element zur netzkonformen Integration von Windenergie. Energiewirtschaf e tliche Tagesfragen. 55 Jg 2005. Heft 7. [46] Nowi A, Jakiel C, Moser P, Zunft S. Adiabate Druckluftspeicherkraftwerke zur netzvert€aglichen Windstrominegration. VDI-GET Fachtagung “Fortschrittliche Energiewandlung und-anwendung. Strom e und W€armeerzeugung. Kommunale und industrielle Energieanwendungen”, Leverkusen, 09 e 10 Mai 2006. [47] Hiratsuka A, Arai T, Yoshimura T. Seawater pumped-storage power plant in Okinawa island, Japan. Eng Geol 1993;35:237e46. [48] Japan Commission on Large Dams, http://web.archive.org/web/20030430004611/http:// www.jcold.or.jp/Eng/Seawater/Summary.htm [last accessed 02.11.14.]. [49] Fujihara T, Imano H, Oshima K. Development of pump turbine for seawater pumped e storage power plant. Hitachi Rev 1997;47(5). [50] Katsaprakakis DAl, Christakis DG, Stefanakis I, Spanos P, Stefanakis N. Technical details regarding the design, the construction and the operation of seawater pumped storage systems. Energy 2013;55:619e30. [51] FAO Irrigation and Drainage Paper 64. Manual on small earth dams. A guide to siting, design and construction. ISSN 0254-5284, http://www.fao.org/docrep/012/i1531e/ i1531e00.pdf [last accessed 02.11.14.]. [52] Gilbert Gedeon P.E, Slope stability. Continuing Education and Development Engineering Inc. Course No: G04eG001. Credit: 4 PDH, http://www.cedengineering.com/upload/ Slope%20Stability.pdf [last accessed 02.11.14.].


Hydropower flexibility and transmission expansion to support integration of offshore wind

16

N.A. Cutululis Technical University of Denmark, Roskilde, Denmark H. Farahmand Norwegian University of Science and Technology (NTNU), Trondheim, Norway S. Jaehnert Sintef Energy Research, Trondheim, Norway N. Detlefsen Danish District Heating Association Merkurvej 7, Denmark I.P. Byriel Energinet.dk Tonne Kjærsvej 65 Fredericia P. Sørensen Technical University of Denmark, Roskilde, Denmark

16.1

Introduction

As weather-driven generation, wind power varies naturally with fluctuations in wind pattern. The variations are driven on longer timescales by atmospheric pressure gradients associated with weather patterns and on shorter scales by turbulent eddies. This variability is one of the key issues, along transmission, in integrating offshore wind power. While the geographical cross-correlation of wind generation generally decreases with distance, and therefore tends to be smoothed with an increasing number of turbines spread over wider areas, in the case of massive offshore wind power deployment, as planned for the North Sea in Europe, the smoothening is significantly less due to the concentration of wind power in a relatively small geographical area. Offshore wind grew over 50% annually in 2013, reaching a total installed capacity of almost 7 GW worldwide (BTM, 2014). Hydro power is one of the fastresponding sources of electricity, thus power systems with considerable amounts of flexible hydro power can potentially offer easier integration of offshore wind power. The objective of this chapter is to present the interaction of offshore wind and hydro power and the chapter also analyzes how the flexibility of hydro generation can match the variability of offshore wind, allowing for larger shares of variable generation to be integrated in the power systems without decreasing its stability. Offshore Wind Farms. http://dx.doi.org/10.1016/B978-0-08-100779-2.00016-7 Copyright © 2016 Elsevier Ltd. All rights reserved.

496

Offshore Wind Farms

16.2

Technologies

16.2.1

Offshore wind power

Offshore application is one of the main drivers in the development of large, modern wind turbines. Almost all offshore wind turbines operating today have a rated power of 2 MW or above. Already, in 2013, the first demonstration offshore wind farm with 6-MW wind turbines was put in operation.1 It is generally agreed to divide wind turbines into four types (1e4 or AeD) according to the main electrical design (Hansen, 2012). For offshore applications, the most commonly used wind turbine uses the type 4 configuration shown in Fig. 16.1. It is characterised by the full-scale back-to-back converters (machine side converter (MSC) and grid side converter (GSC)). The use of the full-scale converter makes it possible to use a multipole generator and thus omit or reduce the gearbox, which has caused significant maintenance costs in some offshore wind power plants. The power collection grid in an offshore wind power plant collects the power from the individual turbines and transmits it to a transformer station, which is usually placed on an offshore platform, but can also be onshore if the distance to shore is short and/or the wind power plant is not very large. The vast majority of existing and planned offshore wind power plants use a power collection grid with a voltage level of 33e36 kV, and none of the wind power plants with more than 50-MW capacity has used a voltage level above 36 kV. However, several studies have shown that higher voltage levels for power collection are economically advantageous, especially in future GW-size wind power plants (Mc Dermott, 2009; Saez et al., 2012).

∼ GB

=

= MSC

∼ GSC

CB TR

SG or AG WTR

Figure 16.1 Main electromechanical components of a type 4 wind turbine. WTR, wind turbine rotor; GB, gearbox; SG, synchronous generator; AG, asynchronous generator; MSC, machine side converter; GSC, grid side converter; CB, circuit breaker; TR, transformer. 1

Gunfleet Sands 3 project: http://www.gunfleetsands.co.uk/en/demonstration-project.


497

A major advantage of using higher voltage is the reduction in investment costs. Another advantage is that it is possible to construct much larger arrays (Mc Dermott, 2009). Also, the reliability can be improved because lower short-circuit levels make it possible to design the collection grid with a more reliable and cost-effective ring topology (Saez et al., 2012). Finally, there is a potential for reducing the losses, but this will depend on the final design. Still, wind turbine manufacturers have been reluctant to offer wind turbines with voltage levels above 36 kV. Among the reasons are that dry transformers are not commercially available above 36 kV, which is an issue of installation cost, reliability and safety and environment concerns (Mc Dermott, 2009). Other issues are the switchgear inside the turbine, which will increase in size as the voltage increases, and the need for higher qualification of the maintenance staff. The possibility to use DC in wind power plant collection grids has been studied in a number of research projects (Max, 2007; Roshanfekr, 2013). The idea is to omit the grid side converter in Fig. 16.1, and connect the machine side converter directly to a DC grid. This is particularly interesting if it is combined with high-voltage direct current (HVDC) transmission described later. The use of DC power collection has a longer-term perspective compared to the use of higher AC voltages. A key bottleneck here is the commercial development of reliable and efficient DC/DC converters to step up the voltage from the individual wind turbine to the collection grid, and from the collection grid to the HVDC transmission line.

16.2.2 Hydro power Hydro generation has ideal characteristics for providing balancing resources. This is because of its high regulation speed and relatively low operational cost. Hence, hydro generation can add flexibility and reserve to the power system in order to compensate for the uncertainty introduced by variable renewable power generation.

16.2.2.1 The existing flexibility of the Nordic hydro system The existing flexibility in the Nordic system is based on import from adjacent countries in periods with low prices and export in periods with high prices such that the net energy exchange on average is approximately zero. Since there is hardly any pumping capacity in the Nordic system, the import to the region is solely balanced by reducing the hydropower production and storing the water. The actual available flexibility will vary a lot from winter to summer, from working-day to weekend and from peak hours to off-peak hours. Due to such uncertainties, the “flexibility” of a system is rather difficult to quantify, but some underlying values are possible to provide, see Table 16.1. The electricity production of hydropower plants connected to reservoirs can be controlled rather well and they are able to provide flexibility. The total installed capacity of these types of power stations adds up to about 35 GW. Furthermore, the available hydro storage sums up to about 125 TWh, whereas the discharge time for some of the biggest reservoirs is several years. The actual flexibility is determined by the concurrent demand level in the Nordic area. In addition the operation of the hydropower system is limited by discharge constraints as

498

Table 16.1

Offshore Wind Farms

Nordic hydropower system Norway

Sweden

Finland

Reservoir

23.5

10.4

w1

Run-of-river

6.2

5.8

w2

Sum

28e25 max av.

16

3

Accumulated reservoir (in TWh)

85

34

5

Hydropower production (in TWh)

125

65

13

Generation capacity (in GW)

well as maximum and minimum reservoirs levels due to environmental regulations. Furthermore, these flexibility estimations can be limited by the bottlenecks in the transmission system within the Nordic region or between the Nordic region and other countries.

16.2.2.2 The future flexibility of the hydro system In the future, the flexibility of the Nordic hydro system is expected to change as a result of various factors: 1. Changes in the hydropower system, such as increase in production capacity or installation of pumping. 2. Stricter rules on hydro operation due to environmental regulations. 3. Changes in the amount of inflow/inflow pattern because of climate change.

The increase in flexibility is primarily expected to be provided by the increase in generation capacity. NVE2 estimates a possible increase of 16.5 GW in western and southern Norway (NVE, 2011). The Centre for Environmental Design of Renewable Energy (CEDREN)3 estimates a potential increase of up to 18.2 GW in southern Norway, without violating current environmental constraints (Solvang et al., 2012). These increases solely include the construction of new power plants, pumps as well as water ways, but no new reservoirs.

16.2.3

Transmission system

Bringing the electricity produced by offshore wind farms to the onshore grid can be done in two ways, based on the technology used: high-voltage alternating current (HVAC) and HVDC, with the latter being further split into line commutated converter (LCC) and voltage source converter (VSC). The transmission infrastructure typically consists of (Fig. 16.2): (a) an offshore transformer platform that collects the power from the wind farm and increases the voltage; (b) a sea cable that transports the power ashore; and (c) a transformer station onshore. For HVDC connections, the transformer stations also accommodate the AC/DC converter. 2 3

Norwegian Water Resources and Energy Directorate (www.nve.no). http://www.cedren.no/.


499

Wind turbine cluster Shoreline

Offshore transformer substation

Submarine cables (AC or DC)

To other wind turbine clusters HVDC converter

Figure 16.2 Typical offshore wind farm connection diagram; the HVDC converter is for the connection that uses high-voltage direct current technology.

The main limiting factor for the HVAC cable is the production of large amounts of reactive power, which can partially be overcome using compensation at one or both ends of the cable. In general, for distances up to 50e100 km from shore, the better solution e at least from power losses point of view e is HVAC (Barberis Negra et al., 2006). In AC connections, the power losses are proportional to the square of the current, while the transferred power is proportional to the square of the voltage. Therefore, it is desirable to have as high a voltage as possible. The main limiting factor in increasing the voltage of the transmission circuit is the size of the transformer. LCC-HVDC is a proven technology used for many years for bulk transport of power. The main advantage of this technology is the proven track record e although this is only for onshore applications and not with wind power e and the relatively high overall conversion efficiency, ie, in the range of 97e98%, while the main disadvantages are the requirement for a relatively strong network, both offshore and onshore, and the comparatively large offshore substation converter (Ackerman et al., 2012). VSC-HVDC is a much newer technology that is considered to have several significant advantages over LCC-HVDC. The main advantage would be that it does not require a strong offshore or onshore AC network and it can control independently the active and reactive power supply (Ackerman et al., 2012). The main disadvantage e besides the lack of track record and field experience, especially regarding offshore wind power e is represented by the comparatively larger losses in the converters, due to the high switching frequency and high on-state voltage drop of the power electronics semiconductor elements. Nevertheless, the performance of VSC-HVDC is constantly improving, with the losses in the converters being reduced to as low as 1% per converter station (Callavik et al., 2011).

500

16.3 16.3.1

Offshore Wind Farms

Summary e case study Models

The analysis includes two interrelated models e a market model and a flow-based model. The market model optimises the strategic utilisation of hydro energy in the day-ahead market. It includes a detailed modelling of water courses and hydro production. Since there is no significant operation cost for hydro power, but a limited amount of water stored in the reservoirs and seasonal inflow patterns, an optimal and robust strategy for hydropower production has to be determined (Flatabø et al., 1998; Belsnes et al., 2009). The results of the market model are verified by a flow-based simulation, including a more detailed grid description (Korpås et al., 2007; Farahmand et al., 2014). This simulation step computes the optimal generation dispatch and flow along lines consistent with the linear approximation of the power flow equations (also called DC power flow) (Wood and Wollenberg, 1996). A detailed analysis on the effect of different offshore grid structures and onshore grid constraints is considered.

16.3.1.1 Strategic exploitation of hydro power using a transport model and area-based representation The first step in the analysis focuses on the optimal strategy of the hydropower production, given the stochastic inflow to the reservoirs as well as the intermittent production from other renewable energy sources. The market model includes different power system characteristics, considering the distinguishing features of the thermal dominated system in Continental Europe and the hydrothermal system in the Nordic area. The model is a fundamental optimisation model for the mid- and long-term simulation of hydrothermal power systems. The market model has its strength in the strategic exploitation of hydro energy stored in reservoirs. Stochastic dynamic programming (SDP) is employed to calculate water values. The optimisation is stochastic due to the natural variation in climate variables such as temperature, wind speed and inflow, and dynamic since the utilisation of hydro reservoirs couples dispatch decisions in time. The water values reflect the expected marginal value of the water stored in the reservoirs, when substituting hydro power with other production sources (Wolfgang et al., 2009). To represent the seasonal as well as annual variation of renewable energy sources (RES), several climatic years (75 years in this case) are taken into account. At the same time a high temporal resolution is used in the model in order to account for the short-term variation of RES. Fig. 16.3 shows the aggregated inflow of all 75 climatic years to the Norwegian hydropower system. In addition to the renewable energy sources, thermal power production is modelled by individual power plants. The inputs to this model include the generation capacity and marginal cost for thermal production, as well as wind production, solar production, electricity consumption, transmission capacity, and information about historical climate variables such as inflow, wind speed, solar radiation and temperature.


501

14 12

Inflow (TWh)

10 8 6 4 2 0

5

10

15

20

25 30 Week

35

40

45

50

Figure 16.3 Inflow scenarios in the Norwegian hydropower system. Farahmand, H., Jaehnert, S., Aigner, T., Huertas-Hernando, D., 2014. Nordic hydropower flexibility and transmission expansion to support integration of North European wind power. Wind Energy 18 (article in press).

The model is divided into several market areas, based on a transport model. The exchange between areas is modelled as transport corridors with a capacity equal to the net transfer capacities (NTCs) between areas. A more detailed description of the model can be found in Wolfgang et al. (2009), while the underlying dataset is described in Jaehnert and Doorman (2014).

16.3.1.2 Detailed grid impact on dispatch of wind and hydro using power flow analysis and nodal representation In order to achieve efficient utilisation of all the available energy resources, especially RES located far away from load centres, and access flexible resources, transmission systems play a pivotal role. Therefore, the next simulation step is related to detailed techno-economic assessment of transmission bottleneck impacts on wind penetration and usage of hydropower flexibility. The simulations are based on flow-based power market simulations using DC optimal power flow (DCOPF) with a detailed grid model. The tool used in this simulation step is SINTEF’s Power System Simulation Tool (PSST) (Korpås et al., 2007; Farahmand, 2012). PSST considers a detailed grid model and computes the optimal generation dispatch and flow along transmission lines for each hour of the simulation year. The optimal solution for each hour is found by minimising the operating cost that essentially expresses the cost of generation based on different marginal generation costs of available power stations. The scenarios and data for generation portfolio and demand in PSST are consistent with those of the EMPS model.

502

Offshore Wind Farms

The distinguishing feature that makes the methodology different from the existing DCOPF algorithm is the modelling of hydro generation and reservoirs. Despite the constant marginal cost of thermal generators, the marginal cost of hydro units (water values) is dependent upon the reservoir level and thus on the amount of water that is available for energy production. The hydro generation is coupled in time by the use of water values and reservoir hydro levels, and PSST is run for the entire year (8760 consecutive hours). This enables the simulation to capture the effect of inflow variation and reservoir hydro level variation related to the hydro generators. At each simulation step, the hydro levels of reservoirs are updated according to the generation and pumping level in the last period and inflow scenario in that time period. The water values are determined based upon the reservoir level and the time of the year. As explained in EMPS model, water values are imported as exogenous inputs from the EMPS model simulation, which is targeted to strategic utilisation of available hydro resources.

16.4 16.4.1

Scenarios Geographical area (EMPS e the Northern European system and PSST Continental Europe with focus on Northern Europe)

16.4.1.1 EMPS model overview The Northern European power market model implemented in EMPS includes a detailed system description for Norway, Sweden, Finland, Denmark, Germany, the Netherlands, Belgium and Great Britain. Furthermore, the exchange to neighbouring countries is considered in the simulations. Within the model, Norway, Sweden, Denmark, Germany and Great Britain are split into several areas, accounting for water courses (specifically in the Nordic countries) and bottlenecks in the transmission system. The resulting model area division is shown in Fig. 16.4.

16.4.1.2 PSST model overview The PSST model covers the Continental European grid. The European transmission grid model consists of five synchronous regions e the Nordic region, RG4 Continental Europe, Great Britain, Ireland and the Baltic region. The total size of the model for each synchronous area, as they are implemented in the model, is shown in Table 16.2. The model encompasses many countries in Continental Europe listed in Table 16.3. 4

Regional groups within ENTSO-E (the Regional Group Continental Europe comprises the TSOs of the former UCTE synchronous area).


503

N11

N10

S1 S2 N9

S3 S4

N8 N7 N6

Fl

N3 N4

N1 S5

N2

N5 U3

S6 D1

D2

U2

G2 NL

U1

G1 G3

BE

G5

G6

G4

Figure 16.4 Geographic overview of the EMPS dataset. Farahmand, H., Jaehnert, S., Aigner, T., Huertas-Hernando, D., 2013. TWENTIES Task 16.3. Nordic Hydro Power Generation Flexibility and Transmission Capacity Expansion to Support North European Wind Power: 2020 and 2030 Case Studies. SINTEF Energy Research. D 16.3. Table 16.2

Size of PSST simulation model

Synchronous regions

# Nodes

# Generators

# Branches

RG Continental Europe

3815

1235

6758

451

841

774

1377

316

2071

Ireland

2

6

1

Baltic

6

12

7

SUM

5651

2410

9611

Nordic Great Britain

504

Table 16.3

Offshore Wind Farms

Countries included in TWENTIES study

Albania

Denmark

Hungary

Netherlands

Slovenia

Austria

Estonia

Ireland

Norway

Spain

Belgium

Finland

Italy

Poland

Sweden

Bosnia-Herzegovina

France

Latvia

Portugal

Switzerland

Bulgaria

Germany

Lithuania

Romania

Ukraine

Croatia

Great Britain

Macedonia

Serbia

Czech Republic

Greece

Montenegro

Slovak Republic

16.4.1.3 Generation mix (2030) The following case study uses the projected state of the European power system in 2030. The according generation mix for the 2030 scenario is mainly based on the outlook of the EU energy trends to 2030 (EU, 2009) and assumptions made in the IEE-EU OffshoreGrid project (Woyte et al., 2011). The existing generation portfolio is adapted by the decommissioning of aging generators and the installation of new generators in order to match the predicted generation mix. In this future scenario for 2030, fuel costs are assumed to be constant, while the CO2 emission price is expected to increase from 13 V/tonne CO2 (2010) up to 44 V/tonne CO2 (2030) according to the assumptions made in the IEE-EU OffshoreGrid project (Woyte et al., 2011). In addition, the development of nuclear capacity is assumed in accordance with current policies in various countries, specifically regarding the nuclear phase-out in Germany (2011). In order to achieve representative simulation results, a benchmark scenario (2010) is calibrated to represent the generation mix reported from ENTSO-E.5 A comparison of the generation mix from the calibrated model and the reported data is shown in Fig. 16.5. The calibration is performed by adjusting the availability factors for the different technologies of thermal power plants. See Jaehnert and Doorman (2014) for further explanations of the calibration process. Apart from Germany, the figures are rather consistent. The difference in Germany is due to the utilisation of the power production on all grid levels in the market model, while ENTSO-E solely reports that for the transmission grid level.

16.4.1.4 Offshore wind power per area Offshore wind power has experienced considerable growth in the last years, reaching 5111 MW installed capacity at the end of 2012, the vast majority of which is in Northern Europe (BTM, 2014). This growth is expected to increase even further, reaching estimated installed capacities between 40 and 56 GW (depending on the scenario) 5

https://www.entsoe.eu.


(a)

700 Hydro Wind Solar Biomass Coal gas oil Nuclear

600

500 Generation mix (TWh)

505

400

300

200

100

0

DE

UK

SE

NO

NL

FI

BE

DK

(b) 700 Hydro Wind Solar Biomass Coal gas oil Nuclear

Generation mix (TWh)

600 500 400 300 200 100 0

DE

UK

SE

NO

NL

FI

BE

DK

Figure 16.5 Generation mix from the market model results for Northern Europe. (a) EMPS data and (b) ENTSO-E data. Reproduced from: Farahmand, H., Jaehnert, S., Aigner, T., Huertas-Hernando, D., 2015. Nordic hydropower flexibility and transmission expansion to support integration of North European wind power. Wind Energy 18 (6), 1075e1103.

506

Offshore Wind Farms

by 2020 (Moccia and Arapogianni, 2011; Cutululis et al., 2012b) and between 110 and 141 GW in 2030 (Cutululis et al., 2012a). The vast majority of the future offshore wind power will be located in the North and Baltic Seas, hence concentrated in a relatively small geographical area. An illustration is given in Fig. 16.6, including the wind farms planned by 2030 (blue). The installed capacities, per country, are given in Table 16.4 (Cutululis et al., 2012a)

16.4.1.5 Hydro expansion scenario Stronger interconnections combined with large-scale development of wind power production in neighbouring countries that are connected to Norway increase the potential for flexible hydropower production, eg, to provide balancing services. Increased hydro generation capacity will have a serious impact on the local ecosystem due to increased variations of the hydro reservoir level and river regulations. Tougher restrictions on the operation of hydropower stations, based upon local environmental impacts, may lead to more limited utilisation of the hydro reservoirs. In this context CEDREN has conducted a study of potential hydro expansion (Solvang et al., 2012). The CEDREN report comprises several case studies for the expansion of hydropower production capacity as well as pumping capacity in southern Norway. It includes 19 specific power plants in southern Norway, which are combined in three different scenarios. 0º

15º E

60º N

Figure 16.6 Offshore wind farms locations: 2020 (red) and 2030 (blue).


Table 16.4

507

Offshore wind power development scenarios per country MW installed by the end 2020

Country

Baseline

Belgium

2156

2156

3956

3956

Denmark

2811

3211

4611

5811

Estonia

0

0

1695

1695

Finland

846

1446

3833

4933

France

3275

3935

5650

7035

Germany

8805

12,999

24,063

31,702

Ireland

1155

2119

3480

4219

Latvia

0

0

1100

1100

Lithuania

0

0

1000

1000

5298

6298

13,294

16,794

Norway

415

1020

3215

5540

Poland

500

500

500

500

Russia

0

0

500

500

1699

3129

6865

8215

UK

13,711

19,381

39,901

48,071

TOTAL

40,671

56,194

113,663

141,071

Netherlands

Sweden

High

MW installed by the end 2030 Baseline

High

These scenarios range from a capacity increase of about 11.2 GW up to an increase of 18.2 GW. These hydro expansion scenarios do not require new regulated reservoirs, and they do not violate the existing environmental restrictions including the limits for the highest and the lowest regulated water level. In the CEDREN study, it is assumed that no new reservoir is built, whereas the new power stations are constructed with new tunnels between existing reservoirs. Scenario 1 (11 GW) includes 12 new power stations, where five are “pump storage power stations” with a combined installed capacity of 5.2 GW and the remainder are “hydro storage power stations.” The pump storage power stations have reversible pump turbines, pumping water between two reservoirs, while hydro storage power stations are not equipped with such pump turbines. One of these potential expansions is located in Tonstad in Norway, where a new pump storage power station is expected. Two cases were analysed in connection with Tonstad as follows: • •

A1 Tonstad pumped storage power station (Homstølvatn e Sirdalsvatn) A2 Tonstad pumped storage power station (Nesjen e Sirdalsvatn)

508

Offshore Wind Farms

The new tunnels in these two cases and associated reservoirs are shown in Fig. 16.7. The detailed results for the water level regulation in the upstream and downstream reservoirs as well as all other cases are explained in Solvang et al. (2012). For the analysed power system scenario of 2030, scenario 1, with an increase of 11.2 GW of hydro generation capacity, is chosen. It is distinguished between new power generation and pumping capacity. The expansion in the Norwegian hydropower system is solely an expansion of generation and pumping capacity, with no additional inflow. However, the hydro expansion in Sweden is expected to be in the form of small-scale hydropower facilities. The hydro generation capacity in Sweden is expected to increase by 1 GW. No expansion of hydropower production is assumed for Finland or Great Britain. The PSST model likewise distinguishes between the expansion of installed production capacity and the expansion of pumping capacity. When the pumping capacity is increased, the hydro production capacity is increased by the identical capacity. The expanded capacity is proportionally divided among generators in one water course based on their capacity. The overview of capacity expansion is presented in Table 16.5 (Solvang et al., 2012).

Gravatn 340 mill. m3 HRV = 660 LRV = 625 Ousdalsvatn 12 mill. m3 HRV = 498 LRV = 482

Nesjen 275 mill. m3 HRV = 715 LRV = 677

Homstølvatn 55 mill. m3 HRV = 498 LRV = 471

Tjørom

Solhom

120 MW

200 MW

A1

Sirdalsvatn 32 mill. m3 HRV = 50 LRV = 48

Kvinen

80 MW

A2

960 MW Tonstad Pumpe

Figure 16.7 Tonstad cases. Reproduced from: Solvang, E., Harby, A., Killingtveit, Å., 2012. Increasing Balance Power Capacity in Norwegian Hydroelectric Power Stations (A Preliminary Study of Specific Cases in Southern Norway). SINTEF Energy Research, CEDREN Project, Project No. 12X757.


509

Hydro power expansion and pumping in southern Norway (11.2 GW) in Solvang et al. (2012)

Table 16.5

CEDREN case

Station name

Power station

A2

Tonstad

Pumped storage

1400

B3

Holen

Pumped storage

700

B6a

Kvilldal

Pumped storage

1400

B7a

Jøsenfjorden

Hydro storage

1400

C1

Tinnsjø

Pumped storage

1000

D1

Lysebotn

Hydro storage

1400

E1

Mauranger

Hydro storage

400

E2

Oksla

Hydro storage

700

E3

Tysso

Pumped storage

700

F1

Sy-Sima

Hydro storage

700

G1

Aurland

Hydro storage

700

G2

Tyin

Hydro storage

700

Total new power generation capacity

16.5

Capacity(MW)

11,200

Results

16.5.1 Transmission expansion and the impacts on the provision of hydropower flexibility Fig. 16.8 illustrates the geographical distribution of the new hydropower stations throughout southern Norway, which coincide with the current largest installed hydropower plants. Solid arrows in Fig. 16.8 indicate power stations that can, in principle, be linked directly to HVDC cables, since they are located close to the sea. As shown, most of the new stations can be directly connected to HVDC cables without the need for reinforcement of the central transmission grid within Norway. However, in order to maintain the exchange with the international grid in the case when one or some of the external links are out of service, it is necessary to upgrade the internal grid. The dotted lines in Fig. 16.8 between the new stations can be the part of the internal grid in Norway that requires to be upgraded. This will call for the construction of new 420-kV links and upgrading to 420 kV at several points on the relevant routes. According to Statnett’s grid development plan, the Norwegian south-west corridor reaching from Aurland down to the southern tip is planned to be upgraded. The existing plans for new links and voltage upgrades in the central transmission grid are described in Statnett (2013).

510

Offshore Wind Farms

Tyin

Aurland

Hol Sima Nore

Mauranger/Oksla/Tysso Tinnsjø

Kvilldal Jøsenfjorden

Holen

Lysebotn

Tonstad

Figure 16.8 Geographical distribution of the new power stations. Solvang, E., Harby, A., Killingtveit, Å., 2012. Increasing Balance Power Capacity in Norwegian Hydroelectric Power Stations (A Preliminary Study of Specific Cases in Southern Norway). SINTEF Energy Research, CEDREN Project, Project No. 12X757.

The power station installation at Tinnsjø may represent the greatest challenge with regard to grid capacity since the distance to appropriate international links is the longest here. The size of the installation will also determine which alternatives are possible. There will be several connection points in the central transmission grid which are appropriate for a new or upgraded 420-kV link from the Tinnsjø area. These may be located to the north towards Nore or to the east and south-east towards Flesaker and Rød/Hasle. The south-west link is a possible exchange route linking onwards to Europe. In periods when power from Tinnsjø can relieve (take over) the power transmission between western Norway and the south-eastern region, the need for grid expansion will be less than if this power is being supplied in addition to this transmission from


511

west to east. However, it is not certain that this fact will be of significance when assessing grid capacity requirements. The above discussion shows that for a provision of flexibility of the Norwegian hydro system to Continental Europe, not only hydro generation and pumping capacity is necessary, but also appropriate transmission capacity.

16.5.2 EMPS results 16.5.2.1 Strategy for using hydro power Since there is a limited amount of water stored in the hydro reservoirs, its long-term utilisation has to be optimised. The dual values of the weekly reservoir balance constraint are regarded as the water values. The water values reflect the expected marginal value of water stored in the reservoirs, when substituting hydro power with other production sources (Wolfgang et al., 2009). The water values are calculated in the market model and are an exogenous input to power-flow model. The resulting water values matrix is a function of the reservoir level and the week within a year. Fig. 16.9 shows the water value matrix for a reservoir in southern Norway. The water values are used as the marginal production cost for hydro power. Given these production costs as the optimal strategy, the hydropower system is used to balance fluctuating energy inflow, in the short term as well as in the long term.

16.5.2.2 Hydro production and reservoir handling The large share of hydropower production in the Nordic area is considered to provide an essential share of production flexibility to the continental power system in order to

Water value (€/MWh)

100 80 60 40 20 0 60 40 20 Week #

0

0

20

40

60

80

100

Depletion (%)

Figure 16.9 Calculated water values for a reservoir in southern Norway (Farahmand et al., 2014). Farahmand, H., Jaehnert, S., Aigner, T., Huertas-Hernando, D., 2014. Nordic hydropower flexibility and transmission expansion to support integration of North European wind power. Wind Energy 18 (article in press).

512

Offshore Wind Farms

be able to integrate large amounts of variable power production from RES into the power system. Figs 16.10 and 16.11 show the accumulated reservoir handling (reservoir trajectories) and hydropower production in Norway for all hours in all 75 climatic years. Plotted are the percentiles for all 75 climatic years in order to display the annual variability. In the 2010 scenario, the reservoir handling (Fig. 16.10a) is characteristic for the Nordic countries, with depletion during the winter and early spring as well as filling during the rest of the year. Assessing the reservoir handling for the 2030 scenario (Fig. 16.11a) shows minor differences. It can be observed that the reservoir levels become higher in general, while the seasonal reservoir storage capability is utilised less. This means that percentiles of the reservoir handling become more spreadout and flatter. It also indicates a change of the reservoir utilisation from a longer-term perspective as compared to shorter term. The aggregated hydropower production for Norway illustrates that significant changes in the hydro production pattern can be expected. In 2010 (Fig. 16.10b) there is a rather stable seasonal production trend, with higher production during winter and lower production during the summer, according to the demand in the Nordic area. In addition there is a diurnal pattern, resulting from the differences in demand during day and night as well as the weekend. The stable seasonal pattern vanishes and more volatile hydropower production occurs in the 2030 scenario (Fig. 16.11b). These changes are due to the significant integration of wind power production in the future power system.

16.5.2.3 Electricity prices and transmission The following figures (Figs 16.12 and 16.13) give a geographic overview of the simulation results for the benchmark case (2010) and the 2030 scenario. Shown are average area prices as well as the marginal profit (congestion rent) of the transmission corridors, which indicate the congestions in the system. The area prices are given as the average over all 75 climatic years. The marginal profit is the annual average over all 75 climatic years as well. In 2010 (Fig. 16.12), low prices in the north and high prices in the south are observed, with two minor exceptions, southern Sweden and eastern Germany. The congestion rent is highest on the interconnectors between the Nordic area and Continental Europe, especially on the NorNed cable (about 60 V/kW per annum). Rather low area prices can be observed in the UK, which is due to a high generation capacity of base load power plants. In the 2030 scenario (Fig. 16.13), prices increase significantly in Continental Europe, while there is a certain price decrease in the Nordic area and the UK. This price development is due to the decommissioning of a high share of thermal generation capacity in Continental Europe. On the other side large capacities of wind power productions are expected to be commissioned in the Nordic area (Sweden) as well as the UK (Scotland).

(a)

9

× 104

8

Reservoir level (Mm3)

7 6 5 4 3

100% 85%

2

50% 15%

1

0% 0

(b)

10

20

30

40

50

Time (weeks) 35 30

Hydro production (GW)

25 20 15 10 100% 5

85% 50%

0

15% 0%

–5

1000

2000

3000

4000

5000

6000

7000

8000

Time (hours)

Figure 16.10 Reservoir handling/hydropower production in Norway 2010. (a) Reservoir handling and (b) hydro production. Reproduced from: Farahmand, H., Jaehnert, S., Aigner, T., Huertas-Hernando, D., 2013. TWENTIES Task 16.3. Nordic Hydro Power Generation Flexibility and Transmission Capacity Expansion to Support North European Wind Power: 2020 and 2030 Case Studies. SINTEF Energy Research, D 16.3.

(a) × 104 9 8 7 Reservoir level (Mm3)

6 5 4 3

100% 85%

2

50% 15%

1

0% 0

10

20

30

40

50

Time (weeks)

(b)

35 30

Hydro production (GW)

25 20 15 10 100% 5

85% 50%

0

15% 0%

–5

1000

2000

3000

4000

5000

6000

7000

8000

Time (hours)

Figure 16.11 Reservoir handling/hydropower production in Norway 2030. (a) Reservoir handling and (b) hydro production. Reproduced from: Farahmand, H., Jaehnert, S., Aigner, T., Huertas-Hernando, D., 2013. TWENTIES Task 16.3. Nordic Hydro Power Generation Flexibility and Transmission Capacity Expansion to Support North European Wind Power: 2020 and 2030 Case Studies. SINTEF Energy Research, D 16.3.


(a)

515

(b)

70 €/MWh

130 €/kW

55 €/MWh

65 €/kW

40 €/MWh

0 €/kW

Figure 16.12 Geographic overview 2010. (a) Area prices and (b) marginal transmission profit. Reproduced from: Farahmand, H., Jaehnert, S., Aigner, T., Huertas-Hernando, D., 2013. TWENTIES Task 16.3. Nordic Hydro Power Generation Flexibility and Transmission Capacity Expansion to Support North European Wind Power: 2020 and 2030 Case Studies. SINTEF Energy Research, D 16.3.

(b)

(a)

70 €/MWh

130 €/kW

55 €/MWh

65 €/kW

40 €/MWh

0 €/kW

Figure 16.13 Geographic overview 2030. (a) Area prices and (b) marginal transmission profit. Reproduced from: Farahmand, H., Jaehnert, S., Aigner, T., Huertas-Hernando, D., 2013. TWENTIES Task 16.3. Nordic Hydro Power Generation Flexibility and Transmission Capacity Expansion to Support North European Wind Power: 2020 and 2030 Case Studies. SINTEF Energy Research, D 16.3.

516

Offshore Wind Farms

The connections between the Nordic area and Continental Europe are still most congested, whereas the marginal profit is more than doubled compared to 2010. In addition, increasing congestions can be observed on the corridor between England and Scotland. These large marginal profits on the transmission corridors show the need as well as the potential for a further expansion of those transmission corridors. The development of the average electricity prices in the future scenarios is mainly influenced by two factors, the additional power production from RES and the expected increase of production costs of thermal power plants.

16.5.3

PSST results

The case study for this analysis covers the whole European transmission grid model consisting of five synchronous regions: the Nordic region, RG Continental Europe6, Great Britain, Ireland, and the Baltic region. The grid model of the Nordic power system is taken from the Norwegian Water Resources and Energy Directorate (NVE) transmission dataset. The model determines the topology of the grid, and the distribution of loads and generation units. For RG Continental Europe ENTSO-E UCTE study model “winter in 2008”, has been used. The grid model represents a snapshot of the power system for that period of year while the demand and generation portfolio are provided by other sources (Woyte et al., 2011; ADAPT, 2008). Network data for Great Britain describing the transmission system, the demand and the generation by fuel type, are available from the 2009 Seven Year Statement on the National Grid Website (National Grid, 2011). For Ireland, no data describing the network have been made available to the project. However, as this is a fairly small network, a two bus equivalent, one for the Republic of Ireland and the other for Northern Ireland, was assumed to be adequate for the purpose of this study. For the Baltic countries, Estonia, Latvia and Lithuania, a reduced equivalent model as in Korpås et al. (2007) is adapted.

16.5.3.1 Offshore grid alternatives to export Nordic hydro flexibility to Continental Europe In this analysis, our main focus is to perform a detailed techno-economic analysis to find out the impact of the grid, both onshore and offshore, to exploit the RES generation and flexibility of the Nordic hydro production. Hence, several grid case studies are investigated for different offshore grid topologies and onshore grid constraints for the 2030 scenario. In order to study the optimal offshore grid topology, the optimal offshore grid topology introduced in IEE-EU OffshoreGrid project (Woyte et al., 2011) is considered as a basis. The increased capacity of hydro power in southern Norway is considered in our simulation for 2030 (Solvang et al., 2012). In addition to the assumptions in the IEE-EU OffshoreGrid project (Woyte et al., 2011), it is important to consider further alternatives for offshore grid topologies in order to exploit the potential additional 6

Regional groups within ENTSO-E (the Regional Group Continental Europe comprises the TSOs of the former UCTE synchronous area.


517

hydropower production in an optimal way. In this respect, three different offshore grid structures are considered, including: • • •

“Case A”: Original offshore grid according to the IEE-EU OffshoreGrid project without any connection between the Ægir offshore wind farm in Norway and the other parts of the grid (Fig. 16.14a). “Case B”: Offshore grid with connection between the Ægir offshore wind farm and Eemshaven, a seaport in northern Netherlands (Fig. 16.14b). “Case C”: Offshore grid with connection between the Ægir wind farm and the Gaia offshore wind farm in Germany (Fig. 16.14c).

Fig. 16.14 shows different offshore grid alternatives in this study. “Case A” is the design proposed in IEE-EU OffshoreGrid project. A rather limited hydro capacity expansion was considered in Norway by 2030 in Woyte et al. (2011). However, the scenarios for hydro capacity expansion considered in Solvang et al. (2012), motivate

(a)

(b)

NO

NO

NO

NO NO

NO

NO

NO

DK

DK UK

GB

UK

DE

GB UK

UK DE NL

DE

DE NL

DE

DE NL

NL

(c)

NO

NO NO NO

DK UK

DE

GB UK

DE NL

DE NL

Figure 16.14 Proposed offshore grid alternatives. (a) Case A, (b) Case B and (c) Case C. Reproduced from: Farahmand, H., Jaehnert, S., Aigner, T., Huertas-Hernando, D., 2013. TWENTIES Task 16.3. Nordic Hydro Power Generation Flexibility and Transmission Capacity Expansion to Support North European Wind Power: 2020 and 2030 Case Studies. SINTEF Energy Research, D 16.3.

518

Offshore Wind Farms

the introduction of new offshore grid alternatives “Case B” and “Case C”. In “Case B”, there is a direct connection between the Norwegian offshore node and the onshore grid in the Netherlands, which allows transfer of hydropower flexibility and offshore wind power directly to the onshore grid. In “Case C” the offshore grid is closed in a loop between Norway, Great Britain and Germany offshore nodes. In this offshore grid configuration, power flows circulate among the countries around the North Sea. This configuration facilitates circulation of power, use of hydropower flexibility from the Norwegian system and penetration of installed offshore wind power to the onshore grid. In order to study the effect of onshore constraints, we superimpose scenarios for onshore grid constraints on the proposed offshore grid alternatives. The scenarios for onshore grid constraints are clustered in three categories including: • • •

No Constraint (NC) Internal Constraint (IC) Internal Constraint with Expansion (ICE)

The “NC” scenario represents the case without internal constraints in the EU grid. In this case, the internal grid in each country is considered as a copper plate, whereas cross-border grid constraints are considered by limiting the transmission capacity for each individual cross-border transmission line and NTC values for the corridors linking countries and areas inside Germany. The NTCs used between areas inside Germany are taken from the dena Grid study-II (dena, 2010). The “IC” scenario considers the onshore grid with today’s internal grid limitations in the German, the Dutch, the British and the Scandinavian systems. The “ICE” scenario represents a situation, where the grid in “IC” scenario is expanded. In Norway, the expansion includes the grid reinforcements according to Statnett’s Network Development Plan (Statnett, 2013), which includes a transmission corridor connecting the proposed hydropower stations located in southern Norway from the CEDREN study (Solvang et al., 2012). In central Europe the grid is expanded according to the Ten-Year Network Development Plan (ENTSO-E, 2012). Table 16.6 presents the annual operating cost taking into account the proposed onshore constraints (“NC”, “IC” and “ICE”), and proposed offshore grid cases. The onshore grid constraints in the German and the Dutch systems limit the transmission of wind and hydro production to the load centres inland, hence increasing the operating cost. The comparison indicates that “NC” is always the solution with the lowest operating cost since the highest penetration of offshore wind is possible into the system. In this scenario, “Case B” of the offshore grid (a direct connection between the Norwegian offshore node and the onshore grid in the Netherlands) results in the lowest operating cost. For “IC” and “ICE” scenarios, the results show a different picture with respect to offshore grid topologies. The optimal offshore grid topology appears to be “Case C”, where the offshore grid is closed in a loop between Norway, Great Britain and Germany offshore nodes. According to investment estimation in L’Abbate and Migliavacca (2011), the investment cost for offshore HVDC grid is expected to be 1700 V/(MW km). The distance between the Norwegian and the German offshore wind farms is roughly


519

Table 16.6 Annual operating costs for 2030 case studies (baseline wind scenario) Onshore grid constraints in the continental system and the UK

Offshore grid cases

Cost [bnEUR/a]

NC

Case A

92.846

Case B

92.749

Case C

92.766

Case A

95.577

Case B

95.527

Case C

95.517

Case A

92.992

Case B

92.928

Case C

92.927

IC

ICE

assumed to be 350 km. The distance between the Norwegian offshore wind farm and Eemshaven in the Netherlands is assumed to be approximately 500 km. The connection capacity is expected to be 1000 MW. In this respect, the investment costs for the extra tie-line in “Case B” and “Case C” with respect to “Case A” are approximated to be 850 MEUR and 600 MEUR, respectively. Taking into account 100 MEUR for the costs of DC/AC converters, switchgears and transformers (L’Abbate and Migliavacca, 2011), the overall cost of the interconnections are roughly expected to be 950 MEUR and 700 MEUR for “Case B” and “Case C”, respectively. The life time factor (LFT) P 1 for the lifetime of 30 years and 5% discount rate is 30 n¼1 ð1þ5%Þn ¼ 15:3725. This factor allows a comparison of the operational savings accumulated throughout the life time of grid project, which is annual savings 15.3725. The accumulated savings for “NC-Case B”, “IC-Case C” and “ICE-Case C” are w1225 MEUR, w936 MEUR and w1005 MEUR, respectively. The reference offshore case for this saving is “Case A” (original offshore grid according to the IEE-EU OffshoreGrid project). Comparing these numbers with investment for all expanded corridors indicates that all accumulated savings are greater than investment cost for each corridor. Therefore, it turns out that making investments in those corridors is profitable in comparison with the accumulated savings throughout the lifetime of the corridors.

16.5.3.2 Case study in Tonstad to assess the correlation between wind and pumping profile The internal grid bottlenecks in the continental system limit the transmission of wind energy from offshore wind facilities in the North Sea to the load centres. Thus, surplus

520

Offshore Wind Farms

power production is redirected towards the Nordic system largely affecting the production behaviour and the reservoir levels where the pump storage facilities are expected to be installed in southern Norway. This impact has been observed in our simulation in the simulated reservoir trajectory at Tonstad hydropower station. The reservoir is expected to serve as a host for a new pump storage unit according to Solvang et al. (2012). Furthermore, Tonstad will be directly connected to the planned HVDC cable to Germany (NordLink). The NordLink HVDC link will link southern Norway to northern Germany. Moreover, two offshore wind production facilities in Germany, the offshore wind farms DanTysk and NordseeOst, are both planned to be connected to the grid point at the German terminal of NordLink cable. Thus, the NordLink cable is connected at one end to Tonstad pump hydro station and at the other end to two offshore wind farms. Fig. 16.15 illustrates the correlation between the Tonstad pumping pattern and the German offshore wind production at wind facilities connected to NordLink HVDC cable. The figure shows that pumping at the Tonstad hydropower station is highly correlated with the variation in wind production at DanTysk and NordseeOst. These results illustrate that in the future power system with a large penetration of wind energy and high level of interconnections, the pumping strategy in the Nordic region will not only be influenced by seasonal inflows but also by the variability of wind production around the North Sea in case of offshore interconnectors in the form of point-to-point HVDC links between countries and a meshed offshore grid.

Tonstad pumping pattern Offshore wind production DanTysk Offshore wind production NordseeOst

1

Energy (p.u.)

0.8 0.6 0.4 0.2 0 4000

4050

h

4100

4150

Figure 16.15 Wind power production of the German offshore facilities connected to the NorGer HVDC cable versus pumping pattern in Tonstad. Reproduced from: Farahmand, H., Jaehnert, S., Aigner, T., Huertas-Hernando, D., 2014. Nordic hydropower flexibility and transmission expansion to support integration of North European wind power. Wind Energy (article in press).


16.6

521

Conclusions

Hydropower systems are a very good option for balancing the natural variability of wind power production, especially when installed offshore. The flexibility of hydropower systems allows power systems with a high share of RES to maintain stability. However, the flexibility of hydropower systems is not unlimited and is determined by the concurrent demand level in the Nordic area. In addition, the operation of the hydropower system is limited by discharge constraints as well as maximum and minimum reservoir levels due to environmental regulations. Furthermore, transmission bottlenecks can also limit the flexibility of hydropower systems. The analysis presented indicates that the value of hydropower flexibility to the European power system is significant, consequently justifying the investment costs for transmission expansion. The results illustrate that in the future power system with a large penetration of wind energy and high level of interconnections, the pumping strategy in the Nordic region will not only be influenced by seasonal inflows but also by the variability of wind production around the North Sea in case of offshore interconnectors in the form of point-to-point HVDC links between countries and a meshed offshore grid.

References Ackerman, T., Orths, A., Rudion, K., 2012. Transmission system for offshore wind power plants and operation planning strategies for offshore power systems. In: Ackermann, T. (Ed.), Wind Power in Power Systems. Wiley, pp. 293e327. ADAPT, 2008. KMB Balance Management in Multinational Power Markets. ADAPT Consulting AS. Barberis Negra, N., Todorovic, J., Ackermann, J., 2006. Electric Power Systems Research 76, 916e927. BTM Wind Report, 2014. World Market Update 2013, Navigant Research. Belsnes, M.M., Feilberg, N., Bakken, B.H., 2009. Stochastic modelling of electricity market prices in Europe with large shares of renewable generation. In: Presented at the 10th European Conference, Vienna, Austria. Callavik, M., Ahstr€om, J., Yuen, C., 2011. HVDC grids for continental-wide power balancing. In: Proceedings 10th International Workshop on Large-Scale Integration of Wind Power as Well as Transmission Networks for Offshore Wind Farms, Energynautics, October, pp. 332e337. Cutululis, N.A., Litong-Palima, M., Sørensen, P., 2012a. North sea offshore wind power variability in 2020 and 2030. In: Proceedings 11th International Workshop on Large-Scale Integration of Wind Power as Well as Transmission Networks for Offshore Wind Farms, Energynautics, November. Cutululis, N.A., Litong-Palima, M., Zeni, L., Gøttig, A., Detlefsen, N., Sørensen, P., 2012b. Offshore Wind Power Data. TWENTIES project, D16.1. ENTSO-E, 2012. 10-Year Network Development Plan 2012 [Online]. Available: https://www. entsoe.eu/fileadmin/user_upload/_library/SDC/TYNDP/2012/TYNDP_2012_report.pdf. European Commission, 2009. EU Energy Trends to 2030-Update. Available: http://ec.europa. eu/energy/observatory/trends_2030/doc/trends_to_2030_update_2009.pdf.

522

Offshore Wind Farms

Farahmand, H., 2012. Integrated Power System Balancing in Northern Europe e Models and Case Studies (Doctoral thesis). Department of Electric Power Engineering, Norwegian University of Science and Technology (NTNU), Trondheim, Norway. Farahmand, H., Jaehnert, S., Aigner, T., Huertas-Hernando, D., 2015. Nordic hydropower flexibility and transmission expansion to support integration of North European wind power. Wind Energy 18 (6), 1075e1103. Farahmand, H., Jaehnert, S., Aigner, T., Huertas-Hernando, D., 2013. TWENTIES Task 16.3. Nordic Hydro Power Generation Flexibility and Transmission Capacity Expansion to Support North European Wind Power: 2020 and 2030 Case Studies. SINTEF Energy Research. D 16.3. Flatabø, N., Haugstad, A., Mo, B., Fosso, O.B., 1998. Short-term and medium-term generation scheduling in the Norwegian hydro system under a competitive power market structure. In: Presented at the EPSOM 98 Zurich, Switzerland. dena German Energy Agency, 2010. dena Grid Study II e Integration of Renewable Energy Sources in the German Power Supply System from 2015e2020 with an Outlook to 2025 [Online]. Available: http://www.dena.de/fileadmin/user_upload/Projekte/Erneuerbare/ Dokumente/dena_Grid_Study_II_-_final_report.pdf. Hansen, A., 2012. Generators and power electronics for wind turbines. In: Ackermann, T. (Ed.), Wind Power in Power Systems. Wiley, pp. 293e327. Jaehnert, S., Doorman, G., 2014. The north European power system dispatch in 2010 and 2020. Energy Systems 5, 123e143. Korpås, M., et al., 2007. Grid Modelling and Power System Data. IEE-EU TradeWind project, SINTEF Energy Research D3.2. L’Abbate, A., Migliavacca, G., 2011. Review of Costs of Transmission Infrastructures, Including Cross Border Connections. Technical Report D3.3.2, EU FP7 REALISEGRID project [Online]. Available: http://realisegrid.rse-web.it/Publications-and-results.asp. Max, L., 2007. Energy Evaluation for DC/DC Converters in DC-Based Wind Farms. Chalmers University of Technology. Moccia, J., Arapogianni, A., 2011. Pure Power. Wind Energy Targets for 2020 and 2030. European Wind Energy Association, [Online]. Available: http://www.ewea.org/fileadmin/ ewea_documents/documents/publications/reports/Pure_Power_III.pdf. Mc Dermott, R., 2009. Investigation of use of higher AC voltages on offshore wind farms. In: Proceedings EWEA. National Grid, 2011. 2011 NETS Seven Year Statement: Chapter 8 e Transmission System Capability. NVE, 2011. Økt Installasjon I Eksisterende Vannkraftverk, Potensial Og Kostnader [Increased Installation at Existing Hydroelectric Power Stations]. NVE Report No. 10 e 2011, Oslo, (in Norwegian). Roshanfekr, P., 2013. Energy-Efficient Generating System for HVDC Off-Shore Wind Turbines. Chalmers University of technology. Saez, D., Iglesias, J., Giménez, E., Romero, I., Rez, M., 2012. Evaluation of 72 kV collection grid on offshore wind farms. In: Proceedings EWEA. Solvang, E., Harby, A., Killingtveit, Å., 2012. Increasing Balance Power Capacity in Norwegian Hydroelectric Power Stations (A Preliminary Study of Specific Cases in Southern Norway). SINTEF Energy Research. CEDREN Project, Project No. 12X757. Statnett, A., 2013. Grid Development Plan [Online]. Available at: http://www.statnett.no/ Global/Dokumenter/Prosjekter/Nettutviklingsplan%202013/Statnett-Nettutviklingsplan 2013-engelsk_03korr.pdf.


523

The Federal German Government, 2011. Government statement on the energy strategyd switching to the electricityof the future. http://www.bundesregierung.de/ContentArchiv/EN/ Archiv17/Artikel/_2011/06/2011-06-09-regierungserklaerung_en.html?nn=709674. Wood, A.J., Wollenberg, B.F., 1996. Power Generation & Control, second ed. WielyInterscience, USA. Wolfgang, O., et al., 2009. Hydro reservoir handling in Norway before and after deregulation. Energy 34, 1642e1651. Woyte, A., et al., 2011. Offshore Electricity Grid Infrastructure in Europe.


Part Four Installation and operation of offshore wind farms


Assembly, transportation, installation and commissioning of offshore wind farms

17

M. Asgarpour Energy Research Centre of the Netherlands (ECN), Petten, The Netherlands; Aalborg University, Aalborg, Denmark

17.1

Introduction

Installation of offshore wind farms is the last step before commissioning of an offshore wind farm, which contributes to approximately 20e30% of development costs or 15e20% of the price of energy. It is projected that the trend of offshore wind installation will grow rapidly in coming years. In Europe alone, by the end of 2014, about 2500 offshore wind turbines were installed, making a cumulative total capacity of 8 GW in 74 offshore wind farms (Corbetta and Mbistrova, 2015). Moreover, there are governmental plans to install another 32 GW of offshore wind energy in Europe by 2020 (European Commission, 2013). This means that the required effort for future distant and large offshore wind farms will be enormous in coming years. During the development of an offshore wind farm the installation step is typically overlooked, resulting in project delays and noticeable risks and financial consequences. Therefore, it is essential to have a better look into the installation steps and optimise them when possible to reduce the installation costs, risks and delays. Before the actual offshore installation takes place, the components should be designed and manufactured, be delivered to the onshore assembly site at the harbour, be assembled based on the installation strategy, and then, be transported to the location of the offshore wind farm. The design of the turbine components, such as the tower, the nacelle and blades, is done by the wind turbine manufacturer. In the following sections of this chapter, these steps are briefly described for a typical three-bladed horizontal axis wind turbine.

17.2

Delivery of components

The first step towards installation of offshore wind farms is delivering components to the onshore assembly site at the harbour. These components are: • • • •

Foundation Tower sections Nacelle Rotor


528

Offshore Wind Farms

The delivery of onshore and offshore substations is directly to their installation locations and no assembly at the harbour is required. Additionally, array and export cable-laying vessels are already loaded with cables and no harbour assembly is necessary. Foundations of turbines are typically delivered directly to the location of the offshore wind farm and, therefore, no harbour delivery is required. Typically, a civil engineering office designs the foundation of the turbines and another civil and electrical engineering office designs the onshore and offshore substations. In order to achieve the optimal reliability and minimise the costs, the foundation and the turbine structure should be designed using the same structural tool to make sure all aerodynamic loads are considered in the foundation design and the natural frequency of the complete structure is calculated correctly. Unfortunately, this is not always the case, and turbine manufacturers and foundation designers have strict confidentiality agreements. When all components are designed and manufactured, they should be transferred to the onshore assembly site at the harbour (Fig. 17.1). Depending of the location of the manufacturer and the component size, the components can be transported on land using oversized trucks or, through the sea, using offshore vessels. It should be noted that there is no need for delivery of all wind farm components at the same time. In fact, there is limited space available at the onshore assembly site and, based on the planned installation strategy, components should be delivered when the previous pack of components is loaded on the installation vessel. Furthermore, typically components providers manufacture the components just before the scheduled delivery date to avoid storage problems.

17.3

Onshore assembly

The onshore assembly site at the harbour is where, based on the installation strategy, all component assemblies are completed and, then, components are loaded onto the installation vessel to be transported to the location of the offshore wind farm (Fig. 17.2).

Figure 17.1 Onshore transport of components using oversized trucks on the left or railways on the right.


529

Figure 17.2 Example of onshore assembly site for offshore wind farms at the harbour.

As explained before, there is no assembly required for onshore and offshore substations and typically they are transported to their installation locations. Additionally, there is no assembly necessary for foundations and, depending on their manufacturing location, they can be directly transferred to the location of the offshore wind farm. Therefore, assembly at the harbour only applies to wind turbine components. Based on the installation strategy, the following assembly concepts for wind turbine components are possible: 1. No onshore assembly: all components should be transported to the location of the offshore wind farm and then be installed one by one. 2. Tower assembly: the tower sections (typically three or four sections) are assembled at the onshore assembly site. Then, the whole tower structure is bolted on the deck of the installation vessel to maximise the vessel’s loading capacity. 3. Assembly of two blades and the nacelle: the nacelle, hub and two blades are connected together. This concept is also known as the “bunny ear” concept. When the assembly is done, the nacelle with two blades attached is placed on the deck of the installation vessel. 4. Assembly of three blades and the nacelle: this concept is similar to the bunny ear concept, but with the whole rotor attached to the nacelle. The problem of this concept is that the required deck area for each rotorenacelle assembly is huge and assuming existing offshore vessel designs, only one rotorenacelle assembly can be loaded on the deck. A workaround is to place the rotorenacelle assemblies on top of each other, which requires the correct structure on the deck for load handling and damage prevention.

The first concept has been used in installation of several offshore wind farms in the past and has proved to be inefficient for large (more than 15 wind turbines) and far offshore (more than 15 km distance to the shore) wind farms with unsuitable weather conditions. The second concept has proved to be a very efficient choice and nowadays

530

Offshore Wind Farms

most offshore wind farms are installed using preassembled towers. Based on the deck configuration of the installation vessel, the third concept can be an efficient choice. As stated, the last concept requires specific installation vessels for multiple rotorenacelle assembly loading and is not an optimal option. As an optimal assembly concept, depending of the wind farm location and availability of installation vessels, one of the following two concepts can be chosen: • •

Tower assembly only Tower assembly and the bunny ear rotorenacelle concept

However, the most optimal installation and assembly concept should be chosen based on the project size, wind farm location and availability of installation vessels.

17.4

Offshore transport

The last step before the installation of an offshore wind farm is transportation of all components to the location of the offshore wind farm. As stated before, depending on the location of the harbour, wind farm and manufacturing facilities, the foundations and offshore substations can be directly transported to the location of the farm. However, wind turbine components are typically transported to the onshore assembly site at the harbour and then are loaded on installation vessels. Currently, there are several installation vessels customised for offshore wind industry and more optimised vessels are in the design phase. Depending on the project specification, one of the following installation vessels can be selected for foundation, substation and turbine installation: • • •

Floating vessel stabilised with mooring lines Floating vessel equipped with motion-compensated crane Jack-up barge

Currently, jack-up barges (Fig. 17.3) are used most often for close to shore wind farm locations and floating vessels with motion-compensated cranes are used for deep waters. For array and export cable installation, custom-made vessels are used. These vessels are customised for cable laying, trenching and rock dumping. For each specific project and installation strategy an installation vessel is reconfigured for equipment placement and deck preparation. This step is normally called the mobilisation and takes place before loading the components from the manufacturing facilities or the onshore assembly site at the harbour to the deck of the vessel. When the installation is finished, the deck area is reconfigured for the next offshore wind installation. This step is normally called demobilisation. Mobilisation and demobilisation of large installation vessels are costly and time-consuming (each operation can take up to one month). After the mobilisation of the installation vessel and loading the components to the deck of the vessel, the vessel can sail to the location of the wind farm. It should be noted that sailing out to the location of the wind farm can only take place when the weather conditions at the location of the wind farm are suitable for the next installation


531

Figure 17.3 A jack-up barge during loading before sailing out to the location of the offshore wind farm.

step. Otherwise, the vessel will wait at the harbour for suitable weather conditions, but the vessel daily rate should still be paid. This delay is normally known as weather delay and for far offshore wind farms can be a significant project risk. Therefore, it is advisable that, based on the historical weather data, the weather delay per installation step be calculated. If this calculation is done, the optimal starting date for the installation can be found to minimise the total weather delay.

17.5

Offshore installation

The installation step of offshore wind farms is when the years of planning come to reality. This step starts when the installation vessel with foundations arrives at the location of the wind farm to install the first foundation, and finishes when the cable installation vessels connect the offshore substation to the onshore substation through export cables. Installation of offshore wind farms can be categorised in four steps: • •

• •

Foundation installation Turbine installation • Tower • Nacelle • Rotor Substation installation • Offshore substation • Onshore substation Cable installation • Array cables • Export cables

In the following sections these steps are briefly described.

532

17.5.1

Offshore Wind Farms

Foundation installation

Depending on the foundation type, the installation vessel and strategy may differ. Currently, approximately 90% of offshore wind turbines are installed on monopiles and the remainder are installed on jackets, tripods or gravity-based support structures. There are also a few demonstration floating turbines, which have no bottom-fixed foundations.

17.5.1.1 Monopiles Monopiles are large hollow steel or concrete tubes, whose thickness and diameter vary based on the turbine size, soil condition and water depth. Before installation of a monopile, a layer of scour protection should be applied to avoid seabed erosion around the monopile. This first scour protection layer is made by rock dumping around the monopile position. When the first layer of scour protection is made, monopiles are lifted from the installation vessel and then positioned on the seabed. Common installation methods of monopiles are pile driving using a hydraulic hammer or pile drilling (Fig. 17.4). On average it takes about one or two days to install a monopile using these methods. If pile driving using a hydraulic hammer is chosen, depending on the seabed condition and water depth, it takes about 2000e3000 hammer hits to drive the monopile into the ground. During the pile driving or drilling, the piling depth is continuously monitored to make sure the monopile is placed into the correct depth. Since for hammering or drilling a stable platform is required, normally jack-up barges are used for monopile installation. If a monopile is used as a foundation, in order to connect and level the turbine tower to the monopile, an extra component, called the transition piece, is necessary.

Figure 17.4 Pile driving by a hydraulic hammer equipped on a jack-up barge.


533

The transition piece is lifted and placed on the top of the monopile and then the space between the monopile and the transition piece (about 10e20 cm thick) is grouted. The top of the transition piece is used as the work platform and the sides are used for boat landing and ladder placement. Moreover, J-shaped tubes are placed on the side of the transition piece to guide the array cables from the tower to the seabed.

17.5.1.2 Jackets and tripods The installations of jackets and tripods to some levels are similar to each other. Similar to the monopile installation, a first layer of scour protection by rock dumping is required (Fig. 17.5). The jackets or tripods are transported to the location of the wind farm using jack-up barges or floating vessels with mooring line stabilisation. When the installation vessel is positioned, the jacket or tripod is lifted and placed on the seabed. Alternatively, the jacket or tripod can be floated and then, using a crane, be positioned. In that case, a heavy lift crane is no longer required. When the structure is positioned into the location, for jackets, four piles and for tripods, three piles are driven into the seabed to fix the foundation. The pile-driving methods for jackets and tripods are similar to monopiles. When the foundation installation is finished, the turbine tower can be installed directly on the topside of the jacket or the tripod.

17.5.1.3 Gravity-based foundation Gravity-based foundations are normally self-buoyant and can be floated or towed out to the location of the offshore wind farm. Since the placement of the gravity-based

Figure 17.5 Scour protection after installation of the monopile (shown in gray) and the transition piece (shown in yellow).

534

Offshore Wind Farms

foundation on the seabed requires a flat area, seabed preparation and scour protection steps are needed. When the seabed is prepared and the foundation is positioned in the right location, the foundation is sunk by influx of water, and then the base of the foundation is filled with ballast to anchor the foundation. When ballasting is finished, the turbine tower can be directly installed on the topside of the gravity-based foundation.

17.5.2

Turbine installation

The turbine components to be installed are the tower, nacelle, hub and blades. The first installation step starts with the tower. As discussed before, the tower sections are typically assembled at the onshore assembly site at the harbour and the complete tower is transported to the location of the wind farm by a jack-up barge (Fig. 17.6). When the installation vessel is in position and stabilised, the tower is lifted and placed on top of the foundation and then bolted. If tower sections are not assembled at the onshore assembly site, the assembly takes place offshore, which logistically takes more time and effort due to the harsh weather conditions offshore. The second turbine component to be installed is the nacelle. Similar to the tower, the nacelle is lifted by the crane off the installation vessel and placed on the top of the tower. If the blades are not already attached to the nacelle, each blade should be lifted separately and connected to the hub. Then, in order to not change the position of vessel or crane, the rotor is rotated to make space for installation of a new blade. This operation is iterated up to the moment that all three blades are installed.

17.5.3

Substation installation

In order to connect the wind turbine generators to a grid, proper electrical infrastructure is required. If an offshore wind farm is located near to shore, an onshore substation is

Figure 17.6 Offshore wind turbine blade installation using a jack-up barge.


535

sufficient; but if the wind farm is located distant from the shore, both onshore and offshore substations are required. In this section, only the installation of offshore substations is discussed (Fig. 17.7), since onshore substation installation follows typical onshore civil works. Prior to the installation of an offshore substation, its foundation should be installed. Typical choices for the foundation of an offshore substation are jackets or gravity-based foundations. When the foundation is installed, the complete substation should be lifted from the installation vessel and be placed on top of the foundation.

17.5.4 Cable installation The last step of offshore wind farm installation is cable installation. Depending on the size and location of the wind farm, array cables connecting the output power of turbines are connected to one or two offshore substation busbars. Then, using export cables, the high-voltage electricity produced by the offshore wind farm is transferred to the onshore substation and from there, to the local electrical grid. The array and export cable routes are planned in such a way as to minimise the total cable length and follow all environmental laws and marine restrictions. In the following, the installations of array and export cables are discussed separately.

17.5.4.1 Array cable installation The array or infield cables are lines of cables connecting several turbines to an offshore substation. If a monopile foundation is used, the array cables are pulled through J-tubes

Figure 17.7 Offshore wind substation installation.

536

Offshore Wind Farms

Figure 17.8 Cable burial into the seabed using a remotely operated cable trencher.

and then are connected to the wind turbine cables in the tower bottom. After cable pulling, a second layer of scour protection by rock dumping should be applied around the foundation. The array cables should be placed 1 or 2 m under the seabed in the space between wind turbines. This is done using trenching remotely operated vehicles (ROV) departed from an offshore vessel and monitored by an experienced pilot so as to not damage the cables (Fig. 17.8). The trenching ROV buries the array cables 1 or 2 m below the seabed, depending on the environmental requirements and IEC and DNV standards (eg, DNV-RP-J301 guideline). The last turbine in a row is connected to an offshore substation. This operation should be done for each row of connected turbines.

17.5.4.2 Export cable installation After connecting array cables to offshore substations using transformers, the voltage is stepped up for onwards transmission over a longer distance. The export high-voltage AC or DC cables connect the offshore substations to an onshore substation. The installation of export cables is similar to array cables, but larger cable-laying vessels and trenching ROVs are used. Typically, the cables near shore should be buried deeper than those far from the shore. After export cable installation, precommissioning tests can be carried out and then, the offshore wind farm can be commissioned.

17.6

Tests and commissioning

Before and after commissioning of an offshore wind farm several tests are carried out to make sure that the wind farm components have proper functionality and the wind farm can be connected to an electrical grid as a stable power plant. As a generic guideline, DNV-OSS-901 can be used for project certification of offshore wind farms. Depending on the project location and regulations of the wind farm operator, the


537

wind turbine manufacturer and the grid operator, several tests may be mandatory for commissioning of an offshore wind farm. According to NoordzeeWind (2008) and Larsen et al. (2009), the following tests should be carried out on wind turbines, Supervisory Control And Data Acquisition (SCADA) system, foundations and electrical system of offshore wind farms: • • • • •

Factory acceptance tests Site acceptance tests Commissioning tests Completion test Performance tests

17.6.1 Factory acceptance tests Factory acceptance tests are several dimensional, material and functionality tests that should be performed during and after manufacturing of foundations and wind turbine components. During factory acceptance tests all component certifications should be reviewed and approved. After the factory acceptance tests, components can be delivered to the onshore assembly site or their installation location.

17.6.2 Site acceptance tests Site acceptance tests are typically done on the SCADA system to ensure proper communication between wind turbines and the electrical infrastructure of the wind farm. In addition to the communication tests based on IEC 104, aviation lights, uninterruptible power supply (UPS) and automation modules should also be tested.

17.6.3 Commissioning tests Commissioning tests are performed on wind turbines, foundations and electrical system components to demonstrate their safe and proper operation. Typical commissioning tests of wind turbines are: • • • • •

Test of wind turbine generator while connected to the grid (a few hours) Test of wind turbine generator while grid loss occurs Wind turbine vibration test Test of yaw system Test of pitch systems

Typical commissioning tests of the electrical infrastructure and foundations are: • • • • • • •

Test of power measurement system Test of array and export cable voltages Test of transformer cooling equipment Test of control equipment Test of the diesel generator, if available Conductivity tests of the concrete reinforcement Test of proper grounding of cable connections

538

Offshore Wind Farms

17.6.4

Completion tests

Completion tests can be performed when all commissioning tests are carried out successfully. Completion tests are done for wind turbines individually and for the wind farm as a whole to demonstrate their promised functionality. Completion tests of each individual wind turbine are typically planned as a continuous operation of the grid-connected turbine for several days where the turbine faults do not exceed a maximum number. The completion test of the wind farm is also typically planned as a continuous operation of all wind turbines, while all of them are producing power and total availability is above a minimum number. The terms and thresholds of completion tests are normally set by wind farm operator and wind turbine manufacturer.

17.6.5

Performance tests

During the warranty period of the wind farm (typically a 5-year warranty), the wind farm operator has the right to perform several performance tests and check whether the wind farm is functioning and producing power as stated in contract terms. According to Larsen et al. (2009) performance tests on wind turbines are: •

• • •

Availability test: time availability of wind turbines, which is defined as the relationship between the time that the wind turbine has been available to produce power to the gird and the time where the grid has been available to receive the produced power from the wind turbine. Power cure test: the measured power curve of wind turbines is validated against their reported power curve by the wind turbine manufacturer. The power curve measurement should be based on IEC 61400-12 standard and other regulations set by local authorities. Electrical system test: in this test the power losses occurring in the main transformer, high- and medium-voltage cables are measured and their compliance according to defined thresholds is demonstrated. Acoustic noise test: the emitted noise of the wind turbines according to IEC 61400-11 is measured and compared to defined environmental thresholds.

If the wind farm passes all commissioning and completion tests, then it can be officially commissioned. It should be noted that for large wind farms, sometimes half of the wind farm is commissioned earlier and the official commissioning takes place when all the turbines and electrical infrastructure are ready for commissioning.

17.7

Conclusions and future trends

The installation of offshore wind farms can only be done successfully if several consecutive and parallel steps are executed according to the project plan and defined budget. Due to the harsh offshore environmental conditions, the installation of offshore wind farms is associated with high risks and costs. It is estimated that future offshore wind farms will be located even further offshore and consequently in deeper waters and with harsher weather conditions.


539

17.7.1 Optimal installation planning There is no generally optimal installation strategy for all offshore wind farms. Depending on the wind farm size, the wind farm distance to the shore, water depth and climate conditions, an optimal installation strategy can be defined to minimise the installation costs and risks. During recent decades, inadequate attention has been paid to installation planning and optimisation of offshore wind farms and only a limited number of planning and optimisation tools, such as OWECOP (Herman, 2002) and ECN Install (Asgarpour et al., 2014) are available in the market for public use. As an example, using the ECN Install tool, the whole process of foundation, turbine and cable installation can be modelled and available resources can be allocated to each installation step (Fig. 17.9). With the chosen installation strategy, the planning of the whole process can be derived and costs and resources associated with each installation step can be calculated. Furthermore, based on the defined installation strategy, total operation time, weather delays and costs are calculated and the commissioning date is estimated. A generic installation planning tool can be used by all parties involved in the installation of offshore wind farms and facilitate their collaboration. By using such a tool, the owner of the wind farm can monitor the work carried out by different parties and ensure a smooth project execution according to the budget and planned schedule. The insurance and financial institutes can identify risks associated with each individual installation step and update their policies accordingly. The vessel and equipment owners can have an estimation of all possible delays within the project and reschedule their future projects consequently. Moreover, engineering offices can use the

Figure 17.9 ECN Install tool for optimal planning of offshore wind installation projects.

540

Offshore Wind Farms

installation planning tool to demonstrate the added value of their innovative solutions compared to traditional installation strategies. In the following, two innovative installation solutions are described briefly.

17.7.2

Offshore harbour

Offshore installation vessels are the main contributor to high installation costs of offshore wind farms. One way to reduce the installation time and costs is to reduce the operation hours of expensive jack-up barges. In order to reduce the operation hours of the jack-up barges, an offshore harbour can be used (Fig. 17.10). The offshore harbour can be designed as a floating pontoon structure moored to the seabed. If an offshore harbour is available, a cheaper feeder can feed new components to the offshore harbour continuously. Then, the expensive jack-up barge does not need to travel back and forth to the onshore assembly site to load new components. In Asgarpour et al. (2015) it is shown that for a 300 MW offshore wind farm with 85 km distance to the shore, an offshore harbour can significantly reduce the installation time and costs.

17.7.3

Breakwaters

Vessels and equipment used for offshore wind installation can only operate when the significant wave height is below a certain limit, typically when it is less than 1.5 m. Due to the harsh weather conditions offshore and vessel operational restrictions,

Figure 17.10 Example of a pontoon structure loaded at the harbour to be used as a floating offshore harbour.


541

significant weather delays, which are associated with high costs, occur during the project. Therefore, it is beneficial to make use of breakwaters around the installation location to attenuate waves to an acceptable level. Moreover, if breakwaters are placed in the surroundings of the wind farm, less expensive installation vessels with lower significant wave height restrictions can be used. In Asgarpour et al. (2015) it is shown that by using one or two rows of breakwaters moored to the seabed around the location of the installation vessel, the wave height and consequently, the weather delays, can be reduced significantly. As discussed, offshore wind share in the produced electricity market is growing rapidly, but installation of offshore wind farms is still not mature enough. Therefore, new innovative installation solutions are required to reduce the costs and risks of future large and far-offshore wind farms and to make offshore wind farms competitive with conventional power plants.

References Asgarpour, M., Dewan, A., Savenije, L.B., 2015. Commercial proof of innovative installation concepts using ECN install. In: European Wind Energy Association (EWEA). EWEA, Copenhagen. Asgarpour, M., Dewan, A., Savenije, L.B., 2014. Robust installation planning of offshore wind farms. In: International Wind Engineering Conference (IWEC). Fraunhofer, Hannover. Corbetta, G., Mbistrova, A., 2015. The European Offshore Wind Industry e Key Trends and Statistics 2014. Brussels. European Commission, 2013. Report from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions e Renewable Energy Progress Report. Brussels. Herman, S.A., 2002. Offshore Wind Farms e Analysis of Transport and Installation Costs. Petten. Larsen, P.E., Larsson, Å., Jeppsson, J., 2009. Testing and Commissioning of Lillgrund Wind Farm. Sweden. NoordzeeWind, 2008. Off Shore Windfarm Egmond Aan Zee General Report. The Netherlands.


Condition monitoring of offshore wind turbines

18

W. Yang Newcastle University, Newcastle upon Tyne, United Kingdom

18.1

Reliability of offshore wind turbines

Up to now, various wind turbine (WT) concepts have been developed. However, the technology of modern wind power has not yet been fully matured and standardized. Based on the configuration of the drive-train subassemblies, these WTs can roughly be classified as [1]: • • •

Geared WTs, with a gearbox, a high-speed asynchronous generator, and a partially rated converter; Geared WTs, with a gearbox, a medium-speed synchronous generator, and a fully rated converter; Direct-drive WTs, with no gearbox but a low-speed synchronous generator and a fully rated converter.

Recently, a few innovative concepts of WTs have also been developed, such as the semidirect-drive WT developed by Goldwind [2]; the WT adopting digital displacement transmission [3], etc. In principle, these innovative designs are superior to conventional concepts in system reliability and efficiency. However, they are still currently in research. Further verification of their actual performance under various operation conditions is still required. Therefore, nowadays the mainstream products in the commercial wind power market are geared and conventional direct-drive WTs. To understand the basic reliability and failure modes of both types of WTs, a survey of different concepts of onshore WTs was conducted by the UpWind project [4], and the results are shown in Fig. 18.1. Fig. 18.1 shows that direct-drive WTs are superior to the conventional gear-driven WTs in the following aspects: • • •

free of gearbox failure; improved reliability for the hydraulic system; less problems in mechanical brakes.

However, as concluded in Ref. [6], direct-drive WTs suffer from more problems in electric subassemblies (eg pitch control and power electronic converter), rotor blades, and generators. Similar conclusion also drawn in the survey of the reliability of Enercon E32/33 direct-drive WTs, see Fig. 18.2, where the failure rate and downtime data were taken from Scientific Measurement and Evaluation Programme (WMEP) [7] managed by Fraunhofer IWES under the German publicly funded programme


544

Offshore Wind Farms

1.6 Direct-drive WTs

Annual failure rate

1.4

Standard gear-driven WTs

1.2 1.0 0.8 0.6 0.4 0.2

Support & housing

Drive train

Gearbox

Generator

Mechanical brake Rotor blades

Rotor hub

Yaw system

Sensors

Hydraulic system

Electronic control

Electrical system

0.0

Figure 18.1 Reliability of geared and direct-drive wind turbines. Reproduced from W. Yang, Condition monitoring the drive train of a direct drive permanent magnet wind turbine using generator electrical signals, J. Sol. Energy Eng. 136 (2014) 021008.

Electrical system Electronic control Sensor Hydraulic system Yaw system Rotor hub Mechanical brake Rotor blades Gearbox Generator Structhousing Drive train

Electrical system Electronic control Sensor Hydraulic system Yaw system Rotor hub Mechanical brake Rotor blades Gearbox Generator Structhousing Drive train 1

0.75

0.5

0.25

Annual failure rate

0

2

4

6

Downtime per failure (d)

8

1

0.75

0.5

0.25

Annual failure rate

0

2

4

6

8

Downtime per failure (d)

Figure 18.2 Reliability characteristics of WMEP survey. (a) E32/E33 e 1357 turbine-years (b) WMEP survey e 7800 turbine-years. Reproduced from P. Tavner, D.M. Greenwood, M.W.G. Whittle, R. Gindele, S. Faulstich, B. Hahn, Study of weather and location effects on wind turbine failure rates, Wind Energy 16 (2013) 175e187.

‘250 MW Wind’ during a 17-year period. The WMEP programme collected 193,000 monthly operation reports and 64,000 maintenance reports from 1500 WTs, which cover approximately 15,000 turbine years. These data are considered to be the most thorough collection of publicly available onshore WT reliability information to date [8]. From Fig. 18.2, it is seen that in spite of turbine types, electrical, electronic control, hydraulic and yaw systems have shown much higher failure rates than rotor blades, gearboxes and generators do. However, they lead to shorter downtimes as they are easier to replace and repair. In contrast, rotor blades, gearboxes and generators show relatively low failure rates, but result in much longer downtimes due to the difficulties in logistics, lifting, replacing and repairing. However, it should be known that the information illustrated in Fig. 18.2 was for onshore WTs. A recent study also discloses that in onshore cases, 75% of the faults


545

cause 5% of the downtime, whereas 25% of the faults cause 95% of the downtime [9]. In other words, downtime onshore is dominated mainly by a few large faults, many of which were associated with gearboxes, generators and blades that are difficult to replace or repair. This 5% of the downtime was mostly associated with the electrical and power electronic control systems, whose defects are relatively easy to fix in an onshore environment. As failure modes of an offshore WT are similar to its onshore counterpart, the failure rates of an offshore WT would be similar to those shown in Fig. 18.2, although the wet and corrosive offshore environment could affect more or less the failure rates of WT components. However, it is well known that downtime offshore is significantly affected by the accessibility of the offshore wind farm. Therefore, it is deemed that the downtime figure of an offshore WT would be quite different from those shown in Fig. 18.2. For example, the failure of a power electronic control system results in short downtime onshore attributed to the potential ease of repair. However, it could lead to a very long downtime offshore, as the downtime offshore is often dictated by unfavourable weather and sea conditions rather than being solely dependent on how quick the fault can be fixed [10]. Condition monitoring (CM) of offshore WTs should extend beyond monitoring only drive-train subassemblies (eg blades, gearbox, and generator), to monitoring non-mechanical subassemblies (eg electrical and power electronic control systems). Early practice has shown that offshore operation and maintenance (O&M) is much more costly and sophisticated than onshore O&M [1]. Particularly in unfavourable seasons, the wind farm may be inaccessible for long periods. As a consequence, any breakdown that needs manual repair or reset could lead to a long downtime and significant revenue loss. However, if the breakdown can be predicted in advance with the aid of a CM system, it will be beneficial to the operator in finance. Even in favourable seasons, visiting an offshore site is still expensive due to the high cost of hiring suitable vessels. Nowadays, the operator still prefers to conduct routine offshore WT maintenance via visual inspection at the site. Frequent site visits could lead to many unnecessary costs that increase the cost of energy of offshore wind. However, the kind of costs resulting from unnecessary site visits can be reduced if the WT is equipped with a remote CM system. Therefore, an effective CM system is essential to an offshore WT in order to achieve high availability and good economic return.

18.2

Challenges in offshore wind turbine operation and maintenance

From the point of view of WT CM, the challenges in the O&M of an offshore WT can be summarized as follows: 1. Offshore WTs show few unexpected failure modes associated with the offshore environment, except those due to the offshore AC connector cable arrays and AC export cables. For this reason, the CM techniques used for onshore WTs are equally applicable to offshore WT CM. However, in offshore circumstances more functions need to be considered to avoid the long downtime associated with offshore logistics.

546

Offshore Wind Farms

2. Offshore weather has a significant influence on the O&M of offshore WTs. Early experience has shown that the higher the wind speed, the lower the availability. This is due to both the increased outages and the limited access to defective WTs in windy weather. Such a view is supported by the data presented in Ref. [9], which shows that the availability of offshore WTs can be improved if the maintenance and repair activities can be well scheduled. In addition to the impact of offshore weather, the availability of an offshore WT is also influenced by the availability of suitable vessels, spare parts and maintenance crews. 3. The accessibility of the offshore wind farm is pivotal to keeping the desired availability of offshore WTs. Take one of the UK Round 3 projects, Dogger Bank as an example, its minimum distance to Blyth is 118 km, while the maximum distance is 200.6 km [1]. This means that an oilfield support vessel (with excursion speed 12 knots) from Blyth would take over 10 h sailing time to the nearest edge of Dogger Bank and 17 h to the furthest edge of the site. Obviously, the long sailing time and narrow window of favourable offshore weather bring challenges to the maintenance of offshore WTs. 4. At present, the average cost of a transfer boat is about £73 per hour [1]. The daily rate of jack-up vessels that are used for offshore WT installation and big maintenance varies from £67,800 to £259,900, depending on the vessel types and the contract with the ship owner [11]. However, it should be known that many of these vessels are not able to cover those offshore wind farms located far from shore, eg Dogger Bank, because they can only travel up to 111 km from a safe harbour. Moreover, they are not allowed to travel on sea when the significant wave height is more than 1.5 m, which makes the >98% accessibility an impossible target to achieve. In addition, the carrying and lifting capability of such vessels is often limited. For these reasons, installation vessels with access systems and helicopters are still options when carrying out the maintenance of those WTs deployed far from shore. The extremely high cost so caused is an issue. 5. Offshore WT maintenance is a very young industry field. To date, there has not been a purpose-designed vessel available specifically to fit the needs of this market. Currently, the offshore WT maintenance work is undertaken either by offshore wind farm installation vessels or by oil and gas jack-up vessels. Wind farm construction vessels are often bounded on busy wind farm construction activities, their availability to WT maintenance work is limited and, moreover, they are usually oversized compared to the actual requirement by offshore WT maintenance. Meanwhile, oil and gas jack-up vessels are extremely expensive to use and, moreover, they cannot always meet the purposes of offshore WT maintenance. For example, their crane is small, their legs are too short, poor jacking performance, much downtime due to poor workability, etc. Therefore, a purpose-designed offshore WT maintenance vessel is still wanted today. 6. Compared to vessels, helicopters enable quicker access to far offshore sites. Thus, they seem an ideal tool for carrying out maintenance at distant offshore sites. However, the practice of Horns Rev 1 indicates that helicopters are unable to cover all site access issues. In other words, helicopters do overcomes the issue of distance, however, their safe operation is still limited by poor weather conditions (eg low visibility, strong wind shear and turbulence, etc.). Moreover, it should be noted that the oil and gas industry has reduced the use of helicopters in operations because, historically and statistically, helicopters have been identified as one of the most dangerous aspects of offshore activities [1].

In summary, accessing difficulties, low availability, and high cost characterize the O&M of offshore WTs. However, in principle all these challenges can be mitigated through efficient CM of WTs. Therefore, remote CM is essential for offshore WT from the point of view of O&M.


18.3

547

Offshore wind turbine condition monitoring techniques

Basically, the CM techniques used for monitoring onshore WTs are seen to be applicable to offshore WT. However, there are still a number of fundamental challenges, as listed below, that need to be addressed specifically when developing offshore WT CM systems (CMSs) [12].

18.3.1 Far offshore distance of newly developed offshore wind farms Statistics [1] show that the average offshore distance of the UK Round 3 offshore wind farms is 93.7e189.4 km, compared to the offshore distance of the Round 1 and 2 offshore wind farms of 8.6e16 km. Moving wind farms to farther offshore brings more challenges to WT maintenance because the far offshore distance makes the site more difficult to access, particularly in winter. For this reason, the time window suitable for WT maintenance becomes narrower as the maintenance activity cannot be performed under unfavourable sea and weather conditions. As a consequence, components that even have minor problems would have to be repaired or replaced in advance to avoid potential long downtime that could result in their failure during unfavourable seasons. This will inevitably waste much of the remaining life of the WT components thus leading to significant financial loss. For this reason, different from those CMSs installed in onshore WTs, the CMSs designed for offshore WTs will play a vital role in remaining life prediction in addition to conducting the conventional operation and health CM tasks. However, it is well known that accurate remaining life prediction is not easy to achieve in practice, particularly in harsh offshore environments, as the ultimate loads in extreme weather and wet, salty and corrosive sea air will accelerate the development of faults.

18.3.2 Large diversity of the offshore WT concepts Since 2008, the market share of the gearless or direct-drive turbines has increased from 12% to 20% [13]. This means that there are more and more WTs of different concepts appearing in onshore wind farms. At present, the offshore wind market is still dominated by gear-driven turbines. However, the diversity of offshore WTs will increase sooner or later. Since different concepts of WTs have different working mechanisms and hardware configurations, different CMSs and the associated CM techniques are needed to meet the specific CM purposes. Currently, the commercially available WT CMSs are mostly vibration-analysis-based systems, which are good at detecting gear and bearing faults occurring in gear-driven WTs. However, whether these systems can be equally effective in detecting the fault occurring in other concepts of WTs is questionable. For example, Fig. 18.2 shows that a power electronic system leads to a significant number of failures in a WT. However, these failures cannot be successfully detected by the existing vibration-analysis-based systems. Therefore, how to

548

Offshore Wind Farms

enhance the CM capability of a WT CMS and enable it to be applicable to monitoring more types of faults and more concepts of WTs is a challenging issue that needs to be solved when developing the future offshore WT CMSs.

18.3.3

Increased offshore WT size

Due to less visual, noise, land usage and social constraints, offshore WTs are, in general, larger than their onshore counterpart to maximise the economic benefit. In theory, the reliability of a WT is independent of its size. However, a survey of more than 6000 onshore WT years for turbines ranging from 300 to1800 kW in Denmark and Germany over 11 years has shown that larger WTs did experience more failures than smaller ones [14]. The immature design technology, variable-speed operation, sophisticated control, and lack of experience of operating large WTs can account for such of the survey conclusions. Therefore, an efficient CMS is demanded by large WTs to mitigate this issue. The added value of a CMS is also highlighted by the incurring significant revenue losses due to the breakdown of large offshore WTs. However, how to detect incipient mechanical and/or electrical faults from large offshore WTs operating at constantly variable speeds is a challenging issue that remains to be solved.

18.3.4

Increased use of electrical and power electronic components

Modern WTs use sophisticated control systems for pitch, generator and converter controls. However, practice shows that electrical and power electronic components are in general less reliable than mechanical components [15,16]. Their failure may not a matter on onshore occasions attributed to the potential ease to repair. However, their failure would lead to a long downtime with significant economic losses farther offshore due to reduced accessibility. Current WT CMSs are not designed for detecting the faults that occur in these electrical and electronic systems. Therefore, how to enhance the CM function of the CMSs and enable them to look after more WT components is an issue worthy to consider when developing offshore WT CMSs. Many preliminary researches have been conducted in recent years [17], the achievements of which will be very helpful in obtaining a final solution.

18.3.5

High cost of offshore WT CM

With the continual boom of the offshore wind industry, more and more offshore WTs will be deployed farther offshore in coming years [18]. In the future, hundreds of WTs would be installed in a wind farm. Assume the unit prices of a vibration-analysis-based CMS is approximately £10,000, then to equip each WT with such a CMS will cost the operator at least £10 m. As a CMS is in essence a power electronic system, it is not reliable in a wet, salty and corrosive offshore environment. Once the WT CMSs breakdown due to component failure, not only is the individual WT left unprotected but a large capital loss will result. For this reason, the future offshore WT CMS must be not only efficient and cost-effective but also reliable enough when operating offshore.


549

To date, there are a number of non-destructive testing (NDT) techniques that can be used in WT CM. They are either proven in the laboratory or have already been used in practice. A brief review of them can be found from Refs. [19e23] and their features are reviewed in Table 18.1 [12]. In the table, the costs of different NDT techniques are classified as follows: • • • •

Low: £10,000.

However, it should be known that the costs vary depending on measurement accuracy, resolution, functionality and applicable environment. From Table 18.1, the features and potential applications of various NDT techniques can be summarized as follows: •

•

• • •

•

•

• •

Vibration analysis and oil particle counters, being low-cost and well-proven, are feasible monitoring techniques. Moreover, their combined use could be a key to WT drive-train monitoring. Currently vibration analysis is more widely used for tracing the growth of WT gearbox and bearing faults than oil particle counters. Oil quality analysis is valuable for gearbox gear and bearing monitoring, in particular fault diagnosis through analysing the composition and shapes of lubrication oil metal particles. In the meantime, it is also an effective approach to monitoring the aging and contamination of the lubrication oil itself. However, it is most likely to be used offline due to high cost. Shaft torque and torsional vibration measurement have been investigated but torque transducer installation will be costly and may be limited by the compact structures of new-generation WTs. Ultrasonic testing is a potentially effective tool for detecting the early WT blade or tower defects, although its application requires methods for scanning the individual components. Thermocouples are cheap and reliable. They are extensively used for monitoring the nacelle, gearbox and generator bearings, lubrication and hydraulic oil, and power electronic temperatures. By contrast, thermography is rarely used because of the high cost of the thermographic camera and difficulties in practical application in operating WTs, although its potential application in WT CMS has been investigated [21]. Fibreoptic strain measurements are proving a valuable technique for measuring bladeeroot bending moments as an input to advanced pitch controllers and can be used to monitor WT blades. They have been demonstrated in operation and improvements in costs and reliability are expected. By contrast, mechanical strain gauges are used only in lab tests as they are prone to failure under impact and fatigue loads. Acoustic emissions could be helpful for detecting drive-train, blade or tower defects during type tests but is has wide bandwidth and is costly, both to measure and analyse. Vibroacoustic techniques have had success, for example, in the aerospace industry, but their costs would be prohibitive for the wind industry and the WT nacelle is not ideal for collecting microphone data. Measurement of electric signals has been widely used in the practice of monitoring rotating electric machines [24,25]. But it has not been adopted by the commercially available WT CMSs. Shock pulse method (SPM) [22] could be an alternative online approach to detecting WT bearing faults, although further experience is still needed in the wind industry.

550

Table 18.1

Non-destructive techniques applicable to wind turbine condition monitoring

CM techniques

Cost

Online CM

Fault diagnosis

Deployment

WT components

1

Thermocouple

Low

Y

N

Already used

Bearings Generator Converter Nacelle Transformer

2

Oil particle counter

Low

Y

N

Already used

Gearbox Bearing

3

Vibration analysis

Low

Y

Y

Already used

Main shaft Main bearing Gearbox Generator Nacelle Tower Foundation

4

Ultrasonic testing

Low to medium

Y

N

Being tested

Tower Blades

5

Electric effects (eg discharge measurement)

Low

Y

N

Already used

Generator

6

Vibroacoustic measurement

Medium

Y

N

N

Blade Main bearing Gearbox Generator

7

Oil quality analysis

Medium to high

N

Y

N

Gearbox Bearing

Offshore Wind Farms

No

Acoustic emission transducers

High

Y

N

N

Blade Main bearing Gearbox Generator Tower

9

Torsional vibration (encoder-based)

Low

Y

N

Being tested

Main shaft Gearbox

10

Fibre optic strain gauges

Very high

Y

N

Already used

Blade

11

Thermography

Very high

Y

N

N

Blade Main shaft Main bearing Gearbox Generator Converter Nacelle Transformer

12

Shaft torque measurement

Very high

Y

N

Being tested

Blades Main shaft Main bearing

13

Shock pulse method (SPM) [22]

Low

Y

N

N

Bearing Gearbox


8

Reproduced from W. Yang, P. Tavner, C. Crabtree, Y. Feng, Y. Qiu, Wind turbine condition monitoring: technical and commercial challenges, Wind Energy 17 (5) (2014) 673e693.

551

552

Offshore Wind Farms

It is worth noting that the aforementioned CM techniques were designed mainly for monitoring WT drive-train components. However, the high failure rate of power electronic components and their potentially long downtime offshore suggest that it is essential to perform the in situ monitoring of those power electronic components. This will be very helpful in achieving high availability of an offshore WT. In recent years, much effort has been spent in this field. Unfortunately, a fully successful technique dedicated to monitoring WT power electronic systems has not yet been achieved. For example, although investigations have been done to understand the failure modes of power electronic components (eg [16,26e28]) and based on which some interesting CM techniques have been developed, such as temperature measurement [29,30], eddy current pulsed thermography [31], resistance and capacitance measurements [32], etc. However, the majority of these techniques were designed for monitoring individual components rather than the whole power electronic control system. In reality, it is not realistic to install a large number of transducers to monitor all individual components within the constraints of both space and cost. Therefore, to develop an innovative technique applicable to the CM of the whole WT power electronic system is still an open issue that remains to be solved.

18.4

Offshore wind turbine condition monitoring systems

At present, WT CM is accomplished mainly through the following two types of CMSs [1,12]: • •

The Supervisory Control and Data Acquisition (SCADA) system. This has already been installed in wind farms in order to provide low-resolution monitoring and supervise WT operation; and The purpose-designed WT CMSs. These provide high-resolution monitoring of WT subassemblies for diagnosis and prognosis purposes.

Both types of CMSs have been recently surveyed [12,33]. Their features and issues are discussed below.

18.4.1

Wind farm SCADA system

Fig. 18.3 shows how WTs and the associated equipment in a wind farm are connected to a SCADA system. The SCADA system monitors signals and alarms, usually at 10-min intervals to reduce the transmitted data bandwidth from the wind farm, and will include the following parameters [34]: • • • •

Active power, and its standard deviation Reactive power Power factor Generator currents and voltages


Substation

553

#1

#2

#n

SCADA PC

Telecommunication network

Wind farm externals

Central control

Service centre

Figure 18.3 Schematic diagram of a typical WT SCADA system. Reproduced from W. Yang, P. Tavner, C. Crabtree, Y. Feng, Y. Qiu, Wind turbine condition monitoring: technical and commercial challenges, Wind Energy 17 (5) (2014) 673e693.

• • • • • • •

Anemometer-measured wind speed, and its standard deviation Turbine and generator shaft speeds Gearbox bearing temperatures (for geared-drive turbines) Gearbox lubrication oil temperature (for geared-drive turbines) Generator winding temperatures Generator bearing temperatures Average ambient temperature within nacelle.

Alarm status will also be monitored by the SCADA system for operational purposes. Potentially, these alarms can help the turbine operator to basically understand the operation condition of the WT key components. However, in a large wind farm, these alarms are often too frequent for rational analysis. In comparison with those purpose-designed CMSs, the wind farm SCADA system collects operational information from some WT subassemblies, which could be used to accomplish some basic WT CM tasks. However, the concerns are: • •

SCADA data are usually 10-min average data. The conventional spectral analysis-based CM techniques that are being popularly adopted in WT CM cannot be applied to interpret the data at so low a rate; SCADA system was not initially designed for CM purposes. It does not collect all the information needed for conducting full CM of a WT;

554

•

Offshore Wind Farms

Values of SCADA data (eg bearing vibration and temperature) vary over wide ranges under varying operational conditions. It is hard to detect an incipient fault from raw SCADA data without an appropriate data analysis tool.

Nevertheless, applying SCADA data to WT CM also possesses a number of advantages. For example, SCADA systems have already been installed in wind farms. No more hardware investment is needed when developing SCADA-based CMSs, hence they are subject to low cost [33]. However, to date few operators are aware of such an added value of SCADA data because its low sampling rate has been considered too low for implementing accurate fault diagnosis. To overcome this issue, SCADA data at higher sampling frequency have been investigated in the EU FP6 CONMOW project [35,36]. However, a successful SCADA-based WT CM has not yet been achieved for the practical reasons mentioned in Ref. [35]. To further explore the added value of wind farm SCADA data on WT CM, more efforts are ongoing. For example, the EU FP7 ReliaWind project [37] is using SCADA data to provide CM for WT generator, gearbox and pitch faults and progress has been made in developing simple signal algorithms to prevent false alarms [38,39]; the CM of WT rotor blade and generator failures were studied in Ref. [40], and so on. In summary, wind farm SCADA data are potentially applicable to WT CM to reinforce a WT CMS at no additional cost, however further improvement is still needed for the following reasons: •

• • •

WT SCADA data are collected under varying operational conditions. In addition to the nonlinear control effects that could dampen the features of faults, it is difficult to detect an incipient fault from SCADA data unless the fault has been fully developed to degenerate the turbine performance significantly. In other words, a serious fault can lead to a change in SCADA data, however the change in SCADA data does not necessarily mean the occurrence of a fault. As mentioned above, a number of advanced techniques have been developed to process WT SCADA data. However, further verification of these techniques is still required before commercialization. The constantly varying operational condition of a WT requests the CM technique to be intelligent and adaptive. However, such a technique has not yet been fully developed. Some SCADA-based CM techniques that are available to date, eg conducting CM through comparing the performances of neighbouring WTs, are simple to implement. However, the reliability of the CM result obtained by using such a method is often affected by the local terrain and wake effect.

Despite low cost, the SCADA-based CMS still cannot replace a purpose-designed WT CMS for the following reasons: • • • •

The SCADA system is not designed to collect all required signals necessary for conducting full WT CM. SCADA data are collected at a low sampling rate, which misses detailed information that is essential for conducting full WT CM and fault diagnosis. To date, a successful CM-purposed SCADA data analysis tool has not been fully developed. The fault-related change on SCADA data, for example an increase in bearing temperature, is usually a late-stage indication of a fault, not giving necessary prognostic lead time for useful WT CM.


555

18.4.2 Purpose-designed WT CMSs The application of a purpose-designed WT CMS has been strongly recommended by certification bodies, for example Germanischer LIoyd [41e43], following a series of catastrophic WT gearbox failures in the late 1990s. Today, a number of WT CMSs are commercially available to the wind industry [12]. Many of these were developed by experienced CM practitioners, such as SKF, Br€uel & Kjær, GE Bently Nevada, Pr€ ueftechnik and Gram & Juhl, based on the long-term experience of monitoring conventional rotating machines. Some of these commercial CMSs are listed in Table 18.2. Table 18.2 shows that the majority of commercially available WT CMSs are vibration-analysis-based systems, although some are used in combination with oil particle counters and fibreoptic strain gauges to enhance their WT CM capabilities [44,45]. There are a few systems based on shaft torque or torsional vibration measurement and some use structural health monitoring [19,20]. It can also be seen that the available WT CMSs mainly focus on monitoring the WT drive train, i.e. the main bearing, shaft, gearbox and generator, using spectral analysis techniques. A few systems are specifically designed for WT gearbox lubrication oil or blade monitoring. This is because the WT drive train and blade components are expensive and their failure can cause long downtimes [14e16]. A typical vibration-analysis-based WT CMS is outlined in Fig. 18.4, where fibreoptic strain gauges and oil particle counters are optional. In principle, the application of this system will be helpful in reducing WT operational risk. However, to date no published work has demonstrated its effectiveness in improving WT availability and there have been false alarms and ineffective fault reports [12]. The high cost of WT CMSs and their interpretation complexity have discouraged WT operators from making wider use of them, despite the fact that they are fitted to the majority of large WTs (>1.5 MW) to ease certification. The unreliable CM results account for this embarrassed situation. Different from traditional rotating machines, WTs often operate in remote locations, rotate at low and variable speed and work under constantly varying loads. As a consequence, WT CM signals, such as vibrations and temperatures, are dependent not only on component integrity, but also the operating conditions (eg rotational speed, loading and ambient temperature). In other words, WT component vibration and temperature changes do not necessarily indicate a fault occurrence, although the presence of a fault may lead to such changes. To demonstrate this, Fig. 18.5 shows the transverse vibration of a perfect shaft rotating at variable speed and subject to varying load, collected from a specially designed WT CM test rig, introduced in detail in [46,47]. From Fig. 18.5, it is seen that even without damage the shaft transverse vibration varies with WT load torque and rotational speed. Moreover, the larger the torque delivered, the stronger the vibration. Thus, shaft vibration is affected not only by machine dynamic integrity itself, but also by its operating conditions. Likewise, component temperatures, for example generator bearing or gearbox oil temperatures, also correlate with WT load and nacelle temperature. In order to demonstrate this, a practical CM dataset measured from an operational WT is shown in Fig. 18.6, where generator bearing temperature fluctuations clearly correspond to generator power fluctuations.

556

Table 18.2

Some purpose-designed WT CMSs Name

Product information

Product

Company

Major functions

Notes

1

WindCon3.0

SKF (Sweden)

Collect, analyse and compile condition-monitoring data that can be configured to suit management, operators and maintenance engineers

The system focuses on the condition monitoring of wind turbine blades, main bearing, shaft, gearbox, generator and tower by the combined use of vibration transducers and a lube oil debris counter

2

TCM

Gram & Juhl (Denmark)

Advanced signal analyses on vibration, vibroacoustic and strain, combined with automation rules and algorithms for generating references and alarms

The WT blades, main bearing, shaft, gearbox, generator, nacelle and tower are monitored by using spectral analysis methods

3

WP4086

Mita-Teknik (Denmark)

Integrated with WT SCADA, the system provides real-time frequency and time domain analyses of turbine operational signals and gives off alarms based on predefined thresholds

With the aid of eight accelerometers, the WT main bearing, gearbox and generator are monitored by using both time and frequency domain analysis techniques

4

Br€uel & Kjær Vibro

Br€uel & Kjær (Denmark)

Collect and process data at fixed intervals and remotely send results to diagnostic server. The time-waveform of the data at any time is accessible for further analysis

The WT main bearing, gearbox, generator, and nacelle (temperature and noise) are monitored by the approach of vibration analysis in combination with temperature and acoustic analyses

5

CBM

GE Bently Nevada (USA)

The system gives monitoring and diagnosis of drive-train parameters. Correlate CM signals with WT operational information (eg wind and shaft rotating speeds), and give off alarms via SCADA

The vibrations of WT main bearing, gearbox, generator and nacelle as well as bearing and oil temperatures are monitored

Offshore Wind Farms

No

CMS

Nordex (Germany)

Actual vibration values during WT start-up period are compared with the reference values. Some Nordex turbines also use the Moog Insensys fibre optic measurement system

The system focuses on the monitoring of main bearing, gearbox and generator. The WT blades are also monitored if the WTs also install Insensys’ fibreoptic system

7

SMP-8C

Gamesa Eolica (Spain)

Continuous online analysis of WT main shaft, gearbox and generator and compare their spectral trends. Warnings and alarms are given through wind farm management system

WT main shaft, gearbox and generator are online-monitored through the spectral analyses of their vibrations

8

PCM200

Pall Europe Ltd (UK)

This is a real-time system for testing and assessing fluid cleanliness

The cleanliness of gearbox lubrication oil is monitored

9

TechAlert 10/20

MACOM (UK)

TechAlert 10 is an inductive sensor to count and size the ferrous and non-ferrous debris, while TechAlert 20 is a sensor only for counting ferrous particles

Both systems are designed for monitoring the debris contained in lubrication or other circulating oils

10

Rotor Monitoring System (RMS)

Moog Insensys Ltd (UK)

RMS is in fact a blade-monitoring system, conducting the condition monitoring of wind turbine blades and rotor by measuring the strains in bladeeroot sections using fibreoptic technology

The load measurement by RMS is also helpful for the load control of pitch regulated wind turbines

11

MDSWind MDSWind-T

VULKLN SELCOM (Germany)

MDSWind measures the vibrations of main bearing, gearbox, generator, and tower of the wind turbine, calculates and displays their statistic indices (eg RMS and Crest factor) online

MDSWind-T is a four-channel portable system developed based on MDSWind

557

Continued


6

558

Table 18.2

Continued Name

Product information

Product

Company

Major functions

Notes

12

Ascent

Commtest (New Zealand)

Ascent is a vibration analysis system for monitoring the main shaft, gearbox and generator of the turbine by the approach of spectral analyses and time domain statistics

System available in three complexity levels. Level 3 includes frequency band alarms, machine template creation, and statistical alarming

13

Condition Diagnostics System

Winergy (Germany)

The system analyses vibrations, load and oil to give diagnosis, predict and recommend for corrective action. Automatic fault identification is provided. Pitch, yaw and inverter monitoring can also be integrated into the system

It mainly focuses on the health monitoring of wind turbine main shaft, gearbox and generator through vibration analysis and oil debris counter

14

Condition Management System

Moventas (Finland)

This is a compact system initially designed for monitoring wind turbine gearbox by measuring temperature, vibration, load, pressure, speed, oil aging and oil particles

The system can be extended to monitor the generator and rotor as well as the controller of the wind turbine

15

OneProd Wind

Areva (France)

This system monitors the main bearing, gearbox and generator of the wind turbine by measuring oil debris, structure and shaft displacement, and electrical signals

It consists of operating condition channels to trigger data acquisitions, measurement channels for surveillance and diagnosis, optional additional channels for extended monitoring

Offshore Wind Farms

No

Flender Services GmbH (Germany)

This is a vibration-monitoring system for assessing the health condition of wind turbine main bearing, gearbox and generator. Both time and frequency domain analyses are adopted

Vibration measurements are taken when load and speed trigger are realized

17

WiPro

FAG Industrial Services GmbH (Germany)

Temperature and vibration measurements are taken for monitoring the main bearing, main shaft, gearbox and generator of the wind turbine

Time frequency analysis used in the system allows speed-dependent frequency band tracking and speed-variable alarm level

18

HYDACLab

HYDAC Filtertechnik GmbH (Germany)

This is a system for monitoring the particles (including air bubbles) in hydraulic and lubrication systems.

It is used mainly for monitoring the gearbox of the wind turbine

19

BLADEcontrol

IGUS ITS GmbH (Germany)

BLADEcontrol is a system specifically designed for monitoring wind turbine blades by comparing spectra with historic spectra obtained from normal blades

Accelerometers are bonded directly to the blades and a hub measurement unit transfers data wirelessly to the nacelle

20

FS2500

FiberSensing (Portugal)

This is also a fibreoptic system designed for monitoring wind turbine blades with the aid of Fibre Bragg grating sensors

This system can be potentially applied to wind turbine blade monitoring, but at the moment it has not been extensively deployed

21

Oil Condition Monitoring System

Rexroth Bosch Group (Germany)

This is a system for the early detection of gearbox damage and the monitoring of oil cleanliness. High-dissolving sensors for the measurement of particles and water content in the lubricating oil are available. Both permit an estimate of the remaining life time of the lubricating oil

This system not only improves the reliability of wind turbines but also their efficient operation by predictable maintenance

559

WinTControl


16

Continued

560

Table 18.2

Continued Name

Product information

No

Product

Company

Major functions

Notes

22

Gearbox Oil Condition Monitoring

Intertek (UK)

Intertek oil condition monitoring services include testing of gearbox oils and lubricants, helping clients extend runtimes for expensive turbines, windmills and other equipment, while minimizing downtime and costly repairs

This is an offline oil analysis system

23

Icount system and IcountPD Particle Detector

Parker (Finland)

Parker’s system is an all-in-one system to determine whether or not system oil is contaminated and the best way to detect particles online or offline

IcountPD is a particle detector; while the Icount system provides early warning of any unwanted changes in hydraulic or lubrication oil quality. Thus increasing the availability of the machinery by reducing the need for unnecessary downtime


Offshore Wind Farms


Blade

561

1 5 3

6

4 Gearbox

9

7

10 Generator

8

2

11

Bearings 12

13

Nacelle 14

Tower

1 --- fibre optic transducers; 2, 8 --- speed transducers; 3, 4, 5, 6, 7, 9, 10, 11 --- accelerometers; 12--- oil debris counter; 13 --- online CMS; 14 --- PC at control center.

Figure 18.4 Outline of a typical vibration-based WT CMS. Reproduced from W. Yang, P. Tavner, C. Crabtree, Y. Feng, Y. Qiu, Wind turbine condition monitoring: technical and commercial challenges, Wind Energy 17 (5) (2014) 673e693.

Vibration (mm)

0.25 0.2 0.15 0.1 30 25

To N.m e( r qu

20

)

15 10

300

320

340

360

380

400

eed (rev/min) Rotational sp

Figure 18.5 The transverse vibration of a perfect shaft under varying loading conditions. Reproduced from W. Yang, P. Tavner, C. Crabtree, Y. Feng, Y. Qiu, Wind turbine condition monitoring: technical and commercial challenges, Wind Energy 17 (5) (2014) 673e693.

Offshore Wind Farms

2.0

90 80

1.5

70 60 50 40 30 20 10 0

1.0 0.5

Power Generator bearing temperature

0 01 Jan

02 Jan

03 Jan

04 Jan

05 Jan Days

06 Jan

07 Jan

08 Jan

Temperature (°C)

Power (MW)

562

09 Jan

Figure 18.6 Correlation between WT output power and generator bearing temperature. Reproduced from W. Yang, P. Tavner, C. Crabtree, Y. Feng, Y. Qiu, Wind turbine condition monitoring: technical and commercial challenges, Wind Energy 17 (5) (2014) 673e693.

Thus, it is inferred that whilst a generator bearing fault leads to an increase in bearing temperature, the change of temperature is also correlated with the changes in ambient temperature and the power generated by the WT generator even if the generator cooling system is in good condition. From the above discussion, one can conclude that vibration, temperature or other WT fault-related parameter responses are not solely dependent on WT integrity. In other words, changes in these fault-related parameters do not necessarily indicate the presence of a fault. In order to reduce false alarms, WT CMSs must carry out more detailed investigations than merely measuring amplitude to discern the true cause of variation, and triggering a fault alarm. Additionally, the harsh operating environment exposes the WT to extreme temperatures, wind gusts and lightning strikes. As a consequence, WT electrical and power electronic systems are also prone to failure, see Fig. 18.7 [48] taken from the ReliaWind project, and their deterioration may be accelerated in worse offshore environments. Such failures could be repaired quickly onshore. But the strong wind speed offshore and access difficulties to offshore sites will exacerbate the effects of these failures and cause long downtimes. Existing WT CMSs have not fully considered the detection, diagnosis and prognosis of these failures. The cost of a WT CMS is also an issue, since a large wind farm may require millions of pounds to equip the entire wind farm with such a WT CMS, excluding the fees for periodic recalibration. Additionally, the operator also faces the challenge of processing, transmitting, storing and interpreting the large amount of data generated by these systems. Apart from this, the development of WT CMS is also limited by present IT technology. It is well known that both low- and high-speed rotating components are included in a WT. In theory, CM data from components with different rotational speeds should be collected by using different sampling frequencies to minimize the CM data size. For example, shaft vibration data are usually sampled at 2 kHz; gearbox and bearing vibration data at 20 kHz and so on. However, this is not normally realized in practice due to hardware limitations. To minimize data transmission, WT CMS processes WT data

30%

25%

20%

15%

10%

Power module

Rotor module Comms & control system

Nacelle module Unknown

Drive train module Auxilliary equipment

Structural module

Whole Condition wind monitoring farm system

Frequency converter LV switch gear Generator assembly Transformer MV switch gear Power feeder cables Unknown Power cabinet Protection cabinet Pitch system Hub Blades Sliprings Unknown Safety chain Sensors Communication system Controller H/W Contyroller /W Unknown Ancillary equipment Yaw system Nacelle cover Nacelle bedplate Unknown Nacelle sensors Unknown Gearbox assembly High speed shaft transmission Mechanical brake Main shaft Unknown Generator silent blocks Hydraulic system Cooling system Top Lift Electrical protetion & safety devices Service crane Lighting & power points WTG meterorological station grounding Electrical cabinets Hub cabinet Beacon Lightning protection system Unknown Tower Foundations Wind farm system Common facilities Unknown Data logger Protocol adapter card for data logger Sensors Condition sensors & cables

5%

0%


Figure 18.7 Failure rate proportions attributable to geared-drive onshore WT subassemblies. Reproduced from M. Wilkinson, F. Spinato, Measuring & understanding wind turbine reliability, Proceeding of European Wind Energy Conference & Exhibition (EWEC) 2010, Warsaw, Poland, 20e23 April, 2010.

Failure rate (%)

563

564

Offshore Wind Farms

and transmits trends to the microprocessor continuously. Spectral analysis is conducted only when an unusual change is detected. Such a strategy mitigates the burden of data processing and transmission from offshore to onshore. However, it increases the risks of losing raw historical data once they are requested for detailed analysis. The last issue concerns WT CMS maintenance. In essence, a CMS consists of a number of electronic components and transducers, which are prone to failure in inclement offshore environments. Therefore, maintenance and recalibration of these electronic components and transducers should be conducted regularly as well to keep their good condition and correct performance. However, it is well known that the offshore WT maintenance window is very narrow. Maintenance activities can hardly cover the maintenance of CMS. Due to a lack of care, how to keep the CMS in good condition during its long service life is a challenging issue.

18.5

Signal processing techniques used for WT CM

The selection of appropriate signal processing and data analysis techniques is critical for the success of WT CM. If the fault-related characteristics can be correctly extracted using these techniques, fault growth can be assessed by observing characteristic variations and these characteristics are also important clues for fault diagnosis. To present a clear review the techniques that are already used by commercial WT CMSs and those that are still in research are discussed separately in the following.

18.5.1

Techniques adopted by commercial CMSs

Both time and frequency domain signal processing techniques have been adopted by commercial CMSs, for example the SKF WindCon3.0 shown in Fig. 18.8.

18.5.1.1 Time domain analysis The SKF’s WindCon system sets time domain warning and alarm levels and plots data trends against time, load or rotational speed, when a trend reaches a predefined threshold, the system triggers an alarm. Time domain trends are usually obtained from well-known parameters, such as overall vibration level, Crest factor, average vibration level and so on and the whole CM process is implemented online. However, CM results are inevitably influenced by varying load and environmental factors. The system also allows the user to review raw signal time waveforms and shaft vibration orbits. However, experience has shown that it is hard to assess WT’s health condition by observing the signal waveforms alone, particularly when the turbine is working under varying load. In practice, some WT performance monitoring systems evaluate a WT’s health status by comparing signals with neighbouring WTs [49], see Fig. 18.9. From Fig. 18.9, it can be seen that an almost constant correlation is maintained when both monitored WTs are normal. Once the health condition of either one is


565

Figure 18.8 Software interface of SKF WindCon3.0 (Observer software). Reproduced from W. Yang, P. Tavner, C. Crabtree, Y. Feng, Y. Qiu, Wind turbine condition monitoring: technical and commercial challenges, Wind Energy 17 (5) (2014) 673e693.

After gearbox failure occurs

90

1.2 The value of measured temperature The value of correlation coefficient

1.1

80

1.0

70

0.9

60

0.8

50

0.7

40 Jun

0.6 Jul

Aug Months

Sep

Oct

Figure 18.9 Performance monitoring by the correlation with neighbouring WTs. Reproduced from W. Yang, P. Tavner, C. Crabtree, Y. Feng, Y. Qiu, Wind turbine condition monitoring: technical and commercial challenges, Wind Energy 17 (5) (2014) 673e693.

Correlation coefficients

Gearbox bearing temperature (°C)

Before gearbox failure occurs 100

566

Offshore Wind Farms

changed, their relationship indicated by the correlation coefficient will fall apart. Such a technique is simple in calculation but its reliability will be affected by local terrain of the wind farm and wake effect.

18.5.1.2 Frequency domain analysis Frequency domain techniques used in the WTCMS, for example envelope analysis [50,51], cepstrum analysis [52] and spectral Kurtosis [53], are based upon the fast Fourier transform (FFT). Considering the FFT was initially designed for processing linear stationary signals with constant amplitudes and fixed-frequency compositions, the maximum variations of the inspected WT CM signals need to be defined in advance, so that signal analysis accuracy can be guaranteed. Historic spectra can be traced with the aid of a waterfall diagram. However, as WTs rotate at variable speed and load, the waterfall diagram obtained is difficult to observe, even after calibration by the rotational speed. The WindCon system also provides a tool aiding the user to calculate the characteristic frequencies of the components at any rotational speed. The cursor function enables the user to select peaks, harmonics and sidebands in the spectrum. If adequate information about the machine or component is known, the type of fault can be readily judged with the aid of this tool. The signal processing techniques used in other WT CMSs are similar, although differences exist between different systems. For example, some WT CMSs use either FFT or cepstrum analyses; some, for example, the WP4086 developed by Mita-Teknik and the CBM by GE Bently Nevada, use acceleration envelope spectra; while some, for example the system developed by Br€ uel&Kjær, use both envelope and cepstrum analysis. Both envelope and cepstrum analyses are based upon the FFT and have been proved to be powerful in extracting faulty features from gearbox and bearing vibration signals. Although the FFT is widely used, it is not an ideal tool for processing WT CM signals which are nonlinear and non-stationary, due to varying speeds and loads and the negative influences of the environment on WT control. Therefore, more advanced signal processing techniques need to be investigated for WT CM.

18.5.2

Techniques in research

Today, a number of advanced signal processing techniques [54e66], including time-frequency analysis and neural network, are being researched to overcome the problems of conventional FFT-based techniques and to find a better solution for WT CM. To give an overview of these newly developed techniques, a brief review has been made and some are listed in Table 18.3 with the consideration of the following four aspects: 1. 2. 3. 4.

Advantages Disadvantages Online CM capability Fault diagnosis capability.

Table 18.3

Technologies being researched for WT CM Online CM

Fault diagnosis

It is still a tool for processing linear signals, not ideal for analysing WT CM signals

N

N

Able to analyse non-stationary signals satisfactorily

It involves intensive calculations and is still a tool for processing linear signals. However, WT CM signals are often nonlinear

N

Y

Discrete wavelet transform

Able to analyse non-stationary signals efficiently

Unable to analyse nonlinear WT CM signals correctly, and unable to locate a desired frequency range flexibly

N

Y

4

Empirical mode decomposition

An ideal tool for processing non-stationary and nonlinear signals, like WT CM signals

Unable to locate a desired frequency range flexibly

N

Y

5

Energy tracking

An efficient tool for analysing WT CM signals

It inherits the disadvantages of wavelet transforms and the accuracy of its results is highly dependent on the correctness of WT speed

Y

Y

6

WignereVille distribution

Able to analyse non-stationary signals satisfactorily

Unable to analyse nonlinear WT CM signals correctly

N

Y

7

Neural network

An ideal tool for developing real-time CMS. It takes all CM information into account, however it is able to process them efficiently

Difficult to train the neural network

Y

Y

8

Data-driven technique

Attributed to ‘natural’ decomposition of original signal and the use of phase information, it is ideal to detect incipient mechanical and electrical defects occurring in WTs

Complex computation and manual selection of interested intrinsic mode functions make it difficult to use online

N

Y

9

Genetic programming

Able to simulate complex problems mathematically

The physical mean of the obtained mathematical model is unknown

Y

Y

No

Technique

Advantages

Disadvantages

1

High-order spectrum

Able to detect the nonlinear relationships between different orders of the harmonics contained in the signal

2

Continuous wavelet transform

3


568

Offshore Wind Farms

From Table 18.3 the following concerns arise: 1. Most of the techniques were only suitable for offline CM and fault diagnosis but were not ideal for online use because of the computational complexity; 2. Some newly developed techniques have not been fully demonstrated, although they may have been tested on one or two types of faults. The difficulties of gathering true WT conditions from overhauls and rebuilds and the fact that not all operational WTs have been monitored properly, means that at the moment it is a challenging task to test or prove these techniques; 3. Neural network [34] and genetic programming [66] can be used for online application but appropriate training is challenging.

Other techniques, such as self-organizing maps [67] and support vector machines [68], are potentially applicable to WT CM attributed to their powerful nonlinear classification capability. However, their application to WT CM has not yet been reported.

18.6

Existing issues and future tendencies of WT CM

The CM of offshore WTs is reviewed in this chapter with the following conclusions: 1. SCADA-based and purpose-designed CMSs are currently available to modern offshore WTs with the following features: a. The SCADA-based CMS requires additional cost. However, the low sampling frequency of SCADA data disables the system from carrying out detailed CM analysis, diagnosis and prognosis functions; b. The purpose-designed CMS is an independent CM system consisting of data acquisition, data conditioning, data transmission modules and a number of transducers. Therefore, its hardware is expensive. But it collects CM data using high sampling frequency, which enables the detailed analysis, diagnosis and prognosis of key WT components; c. More effort should be made to integrate these disparate monitoring methods. 2. Massive deployment of offshore WTs brings a pressing requirement for effective WT CM techniques, which should be: a. Applicable to a wider range of WT concepts than hitherto; b. Able to monitor the key components in the whole WT system rather than solely focussing on drive-train subassemblies; c. Capable of detecting incipient defects and preventing secondary damage; d. Capable of fault detection, diagnosis and prognosis; e. Cost-effective in hardware and reliable in CM result. 3. In contrast to onshore WTs, offshore WTs are larger in size and moreover experience: a. Stronger wind and harsher environment; b. Longer repair and replacement downtimes due to accessing difficulties; and c. Higher repair/replacement costs owing to the additional expenditure on special vessels and lifting facilities.

Therefore, offshore WT CM would play a more vital role in increasing economic return. 4. Vibration and oil analysis are currently effective in monitoring onshore WTs. However, they will have difficulty in meeting the new requirements of future offshore WT CMSs.


569

5. Further researches are required to tackle the following issues: a. Process nonlinear, non-stationary CM signals quickly and accurately; b. Successfully detect incipient WT faults under constantly varying operation conditions and improve the reliability of the CM result by overcoming the negative influences of external loads and other operational conditions; c. Develop techniques dedicatedly to monitoring WT electrical and power electronic systems that contribute a large number of turbine failures. The faults of these systems are not an issue onshore due to the potential ease of repairs. However, they can result in long offshore downtimes due to site-accessing difficulties.

References [1] P. Tavner, Offshore Wind Turbines e Reliability, Availability and Maintenance, The Institution of Engineering and Technology, 2012. [2] Beijing Goldwind Science & Creation Windpower Equipment Co., Ltd. (Online). Available: http://www.goldwind.cn. (accessed December 2014). [3] Artemis Intelligent Power Ltd, Digital Displacement Technology, July 2012. [4] S. Faulstich, B. Hahn, Comparison of Different Wind Turbine Concepts Due to Their Effects on Reliability, 2009. UpWind, EU supported project no. 019945(SES6), deliverable WP7.3.2, public report, Kassel, Germany. [5] W. Yang, Condition monitoring the drive train of a direct drive permanent magnet wind turbine using generator electrical signals, J. Sol. Energy Eng. 136 (2014) 021008. [6] P. Tavner, F. Spinato, G.J.W. van Bussel, E. Koutoulakos, Reliability of different wind turbine concepts with relevance to offshore application, Proceedings of European Wind Energy Conference, Brussels, Belgium, March 31eApril 3, 2008. [7] S. Faulstich, B. Hahn, K. Rohrig, Windenergie Report Deutschland, Institut fur solare Energieversorgungstechnik, Kassel, 2008. [8] P. Tavner, D.M. Greenwood, M.W.G. Whittle, R. Gindele, S. Faulstich, B. Hahn, Study of weather and location effects on wind turbine failure rates, Wind Energy 16 (2013) 175e187. [9] S. Faulstich, B. Hahn, P. Tavner, Wind turbine downtime and its importance for offshore deployment, Wind Energy 14 (3) (2011) 327e337. [10] Y. Feng, P. Tavner, H. Long, Early experience of UK round 1 offshore wind farms, Proc. Inst. Civ. Eng. 163 (4) (2010) 167e181. [11] Y. Dalgic, I. Lazakis, O. Turan, Vessel charter rate estimation for offshore wind O&M activities, 15th International Congress of the International Maritime Association of the Mediterranean (IMAM 2013), October 14e17, 2013. [12] W. Yang, P. Tavner, C. Crabtree, Y. Feng, Y. Qiu, Wind turbine condition monitoring: technical and commercial challenges, Wind Energy 17 (5) (2014) 673e693. [13] International Energy Agency, Technology Roadmap e Wind Energy, 2013 (edition). [14] P. Tavner, F. Spinato, G.J.W. Bussel, E. Koutoulakosb, Reliability of wind turbine subassemblies, IET Renew. Power Gen. 3 (4) (2009) 1e15. [15] J. Ribrant, L. Bertling, Survey of failures in wind power systems with focus on Swedish wind power plants during 1997e2005, IEEE Trans. Energy Convers. 22 (1) (2007) 167e173. [16] P. Tavner, J. Xiang, F. Spinato, Reliability analysis for wind turbines, Wind Energy 10 (2007) 1e18.

570

Offshore Wind Farms

[17] S. Yang, D. Xiang, A.T. Bryant, P. Mawby, L. Ran, P. Tavner, Condition monitoring for device reliability in power electronic converters e a review, IEEE Trans. Power Electron. 25 (11) (2010) 2734e2752. [18] P. Greenacre, R. Gross, P. Heptonstall, Great Expectations: The Cost of Offshore Wind in UK Waters e Understanding the Past and Projecting the Future, Technical report of UK Energy Research Centre, September 2010. [19] C.C. Ciang, J.R. Lee, H.J. Bang, Structural health monitoring for a wind turbine system: a review of damage detection methods, Meas. Sci. Technol. 19 (2008) 1e20. [20] Z. Hameed, Y.S. Hong, Y.M. Cho, S.H. Ahn, C.K. Song, Condition monitoring and fault detection of wind turbines and related algorithms: a review, Renewable Sustainable Energy Rev. 13 (1) (2009) 1e39. [21] E. Jasiuniene, R. Raisutis, R. Sliteris, A. Voleisis, A. Vladisauskas, D. Mitchard, M. Amos, NDT of wind turbine blades using adapted ultrasonic and radiographic techniques, Insight Non-Destr. Test. Cond. Monit. 51 (9) (2009) 477e483. [22] M.A. Rumsey, W. Musial, Application of infrared thermography nondestructive testing during wind turbine blade tests, J. Sol. Energy Eng. 123 (4) (2001) 271. [23] L. Zhen, Z.J. He, Y.Y. Zi, X.F. Chen, Bearing condition monitoring based on shock pulse method and improved redundant lifting scheme, Math. Comput. Simul. 79 (3) (2008) 318e338. [24] P. Tavner, L. Ran, J. Penman, H. Sedding, Condition Monitoring of Rotating Electrical Machines, IET, Stevenage, 2008, ISBN 978-0-86341-741-2. [25] P. Tavner, Review of condition monitoring of rotating electrical machines, IET Electr. Power Appl. 2 (4) (2008) 215e247. [26] V. Smet, F. Forest, J.J. Huselstein, F. Richardeau, Z. Khatir, S. Lefebvre, M. berkani, Ageing and failure modes of IGBT modules in high-temperature power cycling, IEEE Trans. Ind. Electron. 58 (10) (2011) 4931e4941. [27] H. Wang, M. Liserre, F. Blaabjerg, P. de Place Rimmen, J.B. Jacobsen, T. Kvisgaard, J. Landkildehus, Transitioning to physics-of-failure as a reliability driver in power electronics, IEEE J. Emerg. Sel. Top. Power Electron. 2 (1) (2014) 97e114. [28] B. Ji, V. Pickert, B. Zahawi, M. Zhang, In-situ bond wire health monitoring circuit for IGBT power modules, in: The 6th IET International Conference on Power Electronics, Machines and Drives (PEMD), 2012. [29] D. Xiang, L. Ran, P. Tavner, S. Yang, A. Bryant, P. Mawby, Condition monitoring power module solder fatigue using inverter harmonic identification, IEEE Trans. Power Electron. 27 (1) (2012) 235e247. [30] D. Xiang, L. Ran, P. Tavner, A. Bryant, S. Yang, P. Mawby, Monitoring solder fatigue in a power module using case-above-ambient temperature rise, IEEE Trans. Ind. Appl. 47 (6) (2011) 2578e2591. [31] K.J. Li, G. Tian, L. Cheng, A. Yin, W. Cao, S. Crichton, State detection of bond wires in IGBT modules using eddy current pulsed thermography, IEEE Trans. Power Electron. 29 (9) (2014) 5000e5009. [32] A.M. Imam, D.M. Divan, R.G. Harley, T.G. Habetler, Real-time condition monitoring of the electrolytic capacitors for power electronics applications, in: The 22nd IEEE Applied Power Electronics Conference, February 25eMarch 1, CA, USA, 2007, pp. 1057e1061. [33] B.D. Chen, Survey of Commercially Available SCADA Data Analysis Tools for Wind Turbine Health Monitoring, Technical report of Supergen Wind EPSRC Project, November 2010. [34] A. Zaher, S.D.J. McArthur, D.G. Infield, Y. Patel, Online wind turbine fault detection through automated SCADA data analysis, Wind Energy 12 (6) (2009) 574e593.


571

[35] E.J. Wiggelinkhuizen, T.W. Verbruggen, H. Braam, L. Rademakers, J. Xiang, S. Watson, Assessment of condition monitoring techniques for offshore wind farms, J. Sol. Energy Eng. 130 (3) (2008) 031004. [36] S. Watson, B.J. Xiang, W. Yang, P. Tavner, Condition monitoring of the power output of wind turbine generators using wavelets, IEEE Trans. Energy Convers. 25 (3) (2010) 715e721. [37] ReliaWind, Web-link: http://www.reliawind.eu, (last accessed 25.11.10). [38] Y. Qiu, P. Richardson, Y. Feng, P. Tavner, SCADA alarm analysis for improving wind turbine reliability, in: Proceedings of European Wind Energy Conference (EWEA), Brussels, 2011. [39] Y. Feng, Y. Qiu, C. Crabtree, H. Long, P. Tavner, Use of SCADA and CMS signals for failure detection & diagnosis of a wind turbine gearbox, in: Proceedings of European Wind Energy Conference (EWEA), Brussels, 2011. [40] W. Yang, C. Richard, J. Jiang, Wind turbine condition monitoring by the approach of SCADA data analysis, Renewable Energy 53 (2013) 366e376. [41] Germanischer LIoyd, Rules and Guidelines, IV Industry Services, 4 Guideline for the Certification of Condition Monitoring Systems for Wind Turbines. Edition 2007. [42] Germanischer LIoyd, Guideline for the Certification of Wind Turbines. Edition 2003 with Supplement 2004 and Reprint 2007. [43] Germanischer LIoyd, Guideline for the Certification of Offshore Wind Turbines. Edition 2005, Reprint 2007. [44] S. Sheng, Investigation of oil conditioning, real-time monitoring and oil sample analysis for wind turbine gearbox, in: AWEA Project Performance and Reliability Workshop, San Diego, California, USA, 2011. [45] S. Sheng, P. Veers, Wind turbine drive train condition monitoring e an overview, in: Applied Systems Health Management Conference, Virginia Beach, Virginia, USA, 2011. [46] W. Yang, P. Tavner, M. Wilkinson, Condition monitoring and fault diagnosis of a wind turbine synchronous generator drive train, IET Renew. Power Gen. 3 (1) (2009) pp.1e11. [47] W. Yang, P. Tavner, C. Crabtree, M. Wilkinson, Cost-effective condition monitoring for wind turbines, IEEE Trans. Ind. Electron. 57 (1) (2010) 263e271. [48] M. Wilkinson, F. Spinato, Measuring & understanding wind turbine reliability, Proceeding of European Wind Energy Conference & Exhibition (EWEC) 2010, Warsaw, Poland, 20e23 April, 2010. [49] D. McLaughlin, Wind farm performance assessment: experience in the real world, Renewable Energy World Conference & Expo Europe 2009, KoelnMesse, Cologne, Germany, 26e28 May, 2009. [50] C. Hatch, Improved wind turbine condition monitoring using acceleration enveloping, Orbit (2004) 58e61. [51] C. Hatch, A. Weiss, M. Kalb, Cracked bearing race detection in wind turbine gearboxes, Orbit 30 (1) (2010) 40e47. [52] P. Caselitz, J. Giebhardt, M. Mevenkamp, Application of condition monitoring systems in wind energy converters, in: Proceedings of European Wind Energy Conference (EWEC) 1997, Dublin, Ireland, 1997. [53] J. Antoni, R.B. Randall, The spectral kurtosis: application to the vibratory surveillance and diagnostics of rotating machines, Mech. Syst. Signal Process. 20 (2) (2006) 308e331. [54] W.Q. Jeffries, J.A. Chambers, D.G. Infield, Experience with bicoherence of electrical power for condition monitoring of wind turbine blades, IEE Proc. Vis. Image Signal Process. 45 (3) (1998) 141e148.

572

Offshore Wind Farms

[55] C.S. Tsai, C.T. Hsieh, S.J. Huang, Enhancement of damage-detection of wind turbine blades via CWT-based approaches, IEEE Trans. Energy Convers. 21 (3) (2006) 776e781. [56] K. Basset, C. Rupp, D.S.K. Ting, Vibration analysis of 2.3 MW wind turbine operation using the discrete wavelet transform, Wind Engineering 34 (4) (2010) 375e388. [57] W. Yang, P. Tavner, C. Crabtree, An intelligent approach to the condition monitoring of large scale wind turbines, Proceeding of European Wind Energy Conference (EWEC) 2009, Marseille, France, 16e19 March, 2009. [58] W. Yang, J. Jiang, P. Tavner, C. Crabtree, Monitoring wind turbine condition by the approach of empirical mode decomposition, The 11th International Conference on Electrical Machines and Systems (ICEMS) 2008, Wuhan, China, October 17e20, 2008. [59] W. Yang, L. Christian, R. Court, S-transform and its contribution to wind turbine condition monitoring, Renewable Energy 62 (2014) 137e146. [60] W. Yang, L. Christian, P. Tavner, R. Court, Data-driven technique for interpreting wind turbine condition monitoring signals, IET Renew. Power Gen. 8 (2) (2014) 151e159. [61] W. Yang, P. Tavner, R. Court, An online technique for condition monitoring the induction generators used in wind and marine turbines, Mech. Syst. Signal Process. 38 (1) (2013) 103e112. [62] W. Yang, C. Ng, J. Jiang, Wind turbine condition and power quality monitoring by the approach of fast individual harmonic extratcion, J. Sol. Energy Eng. 135 (3) (2013) 034504. [63] W. Yang, S.W. Tian, Research on a power quality monitoring technique for individual wind turbines, Renewable Energy 75 (2015) 187e198. [64] B.P. Tang, W.Y. Liu, T. Song, Wind turbine fault diagnosis based on Morlet wavelet transformation and Wigner-Ville distribution, Renewable Energy 35 (12) (2010) 2862e2866. [65] W. Yang, R. Court, P. Tavner, C. Crabtree, Bivariate empirical mode decomposition and its contribution to wind turbine condition monitoring, J. Sound Vib. 330 (15) (2011) 3766e3782. [66] A. Kusiak, A. Verma, A data-driven approach for monitoring blade pitch faults in wind turbines, IEEE Trans. Sustainable Energy 2 (1) (2011) 87e96. [67] M. Prokopenko, Advances in Applied Self-organizing Systems, Springer, 2008, ISBN 978-1-84628-981-1. [68] B. Sch€olkopf, C.J.C. Burges, A.J. Smola, Advances in Kernel Methods: Support Vector Learning, MIT Press, 1999, ISBN 0-262-19416-3.

Health and safety of offshore wind farms

19

P.O. Lloyd Siemens Centre of Competence EHS Offshore Beim Storhhause, Hamburg, Deutschland The marine environment is inherently dangerous and very unforgiving. What would be a minor inconvenience on shore can become a major incident off shore.

19.1

Limits of this chapter

Health and Safety at Work is a substantial topic and a chapter within a book can only briefly touch on the subject. To put this into perspective; a general managers’ guide to health and safety at work can stretch to over 300 pages, the RenewableUK guidance on Offshore Wind and Marine Energy Health and Safety Guidance is 268 pages and the Trade Industry Organization, G9, guidance on one specific aspect, Working at Height in the Offshore Wind Industry extends to 143 pages. Accordingly, this chapter does not attempt to be a reference document on Offshore Health and Safety, but will explain the process and procedures that have to be followed and provide guidance on where up-to-date and more detailed information may be found. It is written for non-health and safety professionals and is not a substitute for seeking specialist advice on this topic; however, it should help identify the questions that should be asked of such individuals. Documents are not referenced, as change is so rapid in the industry that versions become obsolete very rapidly. All documents referred to are open source and the most recent version may be found by searching the Organization’s Name and the Document Title, which are highlighted within the chapter in Italic.

19.2

Introduction

A wind turbine generator (WTG) only becomes a reality once it is constructed offshore. The construction process encompasses: survey work, cable laying, subsea structure construction in the form of mono-piles or jacket construction, tower build, nacelle lifting, hub and blade attachment and finally commissioning. Hazards within these activities include: heavy lifting, cables under tension, electricity and chemical handling. Wind turbine components have been dropped during vessel loading and lifting mechanisms have failed, dropping blades. Injuries have occurred. Although original equipment manufacturers (OEM) plan to minimise visits to wind turbines, regular servicing and rectification visits are needed. Servicing of wind turbines is a planned and coordinated process usually taking a week to complete. Similarly to automobile Offshore Wind Farms. http://dx.doi.org/10.1016/B978-0-08-100779-2.00019-2 Copyright © 2016 Elsevier Ltd. All rights reserved.

574

Offshore Wind Farms

servicing, oil and fluid levels are checked, as are other key components. Outside the servicing regime additional visits can be undertaken to carry out retrofit or warranty work as well as fault-finding and component changes. Fires have occurred whilst turbines have been manned and on very rare occasions have resulted in fatalities. Rectification can involve major component exchange, necessitating the use of jack-up vessels. The requirement to reach and transfer to a turbine offshore introduces logistic issues, be it marine or aviation risk. Vessels range from small agile crew transfer vessels (CTV) carrying 12e24 technicians, service/construction operation vessels with around 60 on board, to hotel ships with hundreds on board. Marine incidents have included vessel fires leading to abandonment and striking an object under power leading to lifeboat and search and rescue helicopter call out. Although offshore helicopter accidents are extremely rare, the consequences tend to be catastrophic. Inherent in turbine design is a requirement to work at height. This includes climbing the ladders on a turbine foundation following a marine transfer from a CTV, to climbing the full height of a turbine using the internal ladder system when the lift is unavailable. This physical exertion has led to heart attacks. Working at height also introduces risk from dropped objects and falling from height. In addition, there are risks to offshore structures from collision of passing vessels or planes which could lead to structural collapse and finally the offshore environment can throw extremes in temperature, wind, wave and tidal forces. Such complexity requires a methodical approach to risk management. Fifteen pages can only skim the subject and only attempt to describe the task that has to be completed.

19.3

Legal framework

All employers have a duty of care for those in their employment. The Health and Safety Executive (HSE), the Maritime and Coastguard Agency (MCA) and the Marine Accident Investigation Branch (MAIB) are responsible for different aspects of the law applied to offshore activities. For example, WTG must meet statutory requirements for machines. All marine activity will come under the authority of the MCA. It can help clarify the requirement by thinking of what would happen if an accident were to occur. The HSE would investigate if the accident was to occur on an Offshore Renewable Energy Installation (OREI) and either the MCA or the MAIB would investigate a marine accident. There is memorandum of understanding between the organizations to determine who would have precedence if the accident occurred at the boundary of responsibility, for example whilst transferring from a vessel to an OREI. The health, safety and welfare of personnel are protected by law. The UK Health and Safety Laws follow a goal-setting approach. Instead of a prescribed checklist of things to do, which may not be right in all circumstances, goal-setting law sets out the objectives to be achieved. Those that are responsible for creating the risks in the workplace are responsible for controlling them and must as a minimum: • • •

systematically identify hazards; assess the risks and the consequences of those hazards being realized; and put in place suitable procedures and measures to control the risks.


575

The offshore wind industry is still in its infancy and the ‘pushing of boundaries’ in terms of turbine size and distances from shore continue to produce additional health and safety challenges. The Health and Safety Executive document Reducing Risk and Protecting People (2001) advises if the level of risk is unclear due to new technology or processes then a ‘precautionary principle’ should be adopted. If there is a plausible and credible risk that serious harm can occur then that risk should be mitigated regardless of lack of frequency information or data. In simple terms, employees should have the expectation of going to work and at the end of the work day returning home without any deterioration in their physical or mental health.

19.4

Safety management system

How can an employer meet such a requirement? Experience has shown that an organization must have a safety management system (SMS) in place to control risk. There are various models of SMS available; however, British Standard Occupational Health and Safety Management (BS OHSAS) 18001 and BS EN ISO 9001 on safety and quality respectively, are seen as international standards adopted by the major players in the offshore renewable energy industry. Employers use compliance with these standards as a recognized short hand that they have safety and quality under control. Some companies refuse to work with other businesses that do not carry such recognition. There is therefore a need for businesses to obtain independent certification that they meet these standards; this service is provided by international certification bodies such as Det Norske Veritas e Germanischer Lloyd (DNV-GL). The process of certification can in itself be a challenging, time-consuming though very beneficial process. However, there is a danger that certification becomes the companies’ aim and not protecting the safety of its employees. Regulators do not only police health and safety law but also provide guidance on what is expected from employers and employees. The UK HSE portal is an excellent place to start e http://www.hse.gov.uk with HSG65 Managing Health and Safety providing a foundation of what is expected of an organization’s management system. The key to effectively managing for health and safety are: • • •

Leadership and management (including appropriate business processes); A trained/skilled workforce; An environment where people are trusted and involved.

HSG65 does not prescribe what exactly has to be put in place, but uses the Plan, Do, Check, Act cycle of goal-setting.

19.5

Plan, do, check, act

The Plan, Do, Check, Act approach, Fig. 19.1, should be seen as a circular process and may need to be completed more than once as a project is developed and implemented.

576

Offshore Wind Farms

Plan

Act

Do

Check Figure 19.1 Plan, Do, Check, Act. http://www.hse.gov.uk/managing/plan-do-check-act.htm.

19.6

The offshore renewable energy industry

The first ever commercial offshore windfarm was installed in Vindeby, Denmark in 1991. The wind farm consisted of 11 450-kW turbines located approximately 2 km from the coastline. The first offshore windfarm in the UK was a near-shore installation in Blyth harbour, north-east England in 2000. For comparison, the total Vindeby output was around 5 MW, a single turbine in 2015 is capable of producing 7 MW and larger turbines are in development. As turbine size has increased they are also being located further offshore, The Bard offshore wind farm in Germany became the first to be located 100 km offshore, connecting to the power grid in September 2013. This mixture of size, complexity and distance to shore, combines to produce a complex working environment with many and varied hazards that could cause injury and loss of life. Nevertheless, the renewable industry is not assessed as a Hazardous Industry and as such is not currently subject to the UK’s Control of Major Accident Hazards Regulations (COMAH), as is the case with oil and gas for example. The Plan, Do, Check, Act cycle will be used as a framework to look at the Offshore Renewable Energy Industry.

19.7

Plan

Policy e Leadership and management need to provide strong guidance on what is expected within the organization. This is more than just a glib statement; it has to be followed by actions and resources. A clear and succinct policy will be felt and followed throughout the organization. Many of the companies working in the offshore renewable sector have a background in the petrochemical industry and bring their zero tolerance of harm culture to the renewable industry. Think about where you are now and where you need to be e An analysis of the organization’s current capability at managing health and safety is a prerequisite to


577

assessing the risks as projects proceed offshore. New competences may be required, which may have to be sourced from outside the organization and from other industries. As a relatively new industry; health and safety rules, regulations and procedures are constantly evolving and the option of controlling risk through the perspective of compliance in isolation, will not work. Say what you want to achieve and who will be responsible e Objectives and targets must be achievable and there has to be an understanding of who is responsible for meeting such targets. Organizations that have a long track record of managing complex and high hazard activity, for example aviation, realise that risk is managed by the owners of the risk and not the prerogative of health and safety professionals. Clarity in accountability and responsibility will ensure correct resources are applied and that those that have to implement policy have the training and competence to complete their actions. Accountability can best be described as the owner of the risk, who is ultimately responsible should anything go wrong and has the budget to implement change. The responsible individual implements the process and procedures put in place to manage risk and should have received the appropriate training and equipment to undertake his duties. Decide how to measure performance e Performance can be measured in many ways. The easiest is to look at historical data e how often an event occurs per 100,000 h worked. Such historic data showed that there was a correlation between the number of major injury (accidents), minor injury (incidents) and non-injury accidents (near-misses). Fig. 19.2 shows the ratios found by Heinrich in 1931. This led to the belief that in pursuing the non-injury events, this would also eliminate major injury accidents. This activity tended to concentrate on the causes of the minor events and not the precursors of the major events. Recent work is looking at concentrating on the precursors of major injury events. A popular statistic is the number of days worked without a lost time incident (time-off following an incident). Where an organization rewards such milestones, the pressure on an individual not to report an accident on the 499th day out of a 500-day target is immense.

1 major injury

29 minor injuries

300 non-injury accidents Figure 19.2 Derived from Heinrich HW (1931). Industrial accident prevention: a scientific approach. McGraw-Hill.

578

Offshore Wind Farms

Such historic information is known as lagging indicators. They are the bottom-line numbers of safety performance within a project. They tell you how many incidents occurred and how many people were hurt. They do not say how well an organization is at preventing accidents. Conversely, low reporting may give a false impression that safety is under control. A leading indicator is a measure preceding or indicating a future event used to drive and measure activities carried out to prevent and control injury. Examples include: safety training, employee perception surveys and safety audits. An example that has worked well in the renewable industry is the recording of the number of management safety walks, these demonstrate leadership’s commitment to safety and provide an interaction with the work force. Consider hazards and cooperate with those that share those risks. Wind farm projects are a complex interaction of risk owners. A WTG is a complex piece of equipment that introduces risk from rotating equipment, combustible materials, electricity and stored energy. Design risk assessments (DRA) are completed by the OEM and where possible these risks are eliminated in initial design. Not all risks can be eliminated and process and procedures are introduced to control and mitigate these risks. Such processes and procedures need to be clearly transferred between the OEM and the customer. Most projects involve numerous subcontractors, from marine heavy lift down to a specialist inspection team of one. In identifying risks, those that are exposed to, or could be affected by it, should be involved in the deliberation. This also applies to adjoining risk producers, for example colocated oil and gas installations. Plan for change and identify any legal requirements that have to be met e Changes in circumstances can alter the risk profile. If we consider the different phases of a wind farm development • •

• • •

Preconsenting survey Construction phase • Under water • Cable laying • Foundations • Turbine construction Operation and Maintenance (O&M) • Planned maintenance • Rectification Repowering Decommissioning

The number of personnel exposed to hazards and the type of hazards can vary significantly. A construction site can have 10 times the number of vessels and personnel offshore compared to the O&M phase. Emergency procedures and first aid provision may be well found during construction, but may reduce dramatically as the project moves to Operation and Maintenance. Weather changes can alter marine risk and night operations may complicate working procedures and delay rescue procedures. Change has to be monitored and its impact has to be anticipated to ensure risk remains under control.


19.8

579

Do

Identify your risk profile e With such a new industry, identifying the risk is the foundation of any safety management system relating to offshore renewable energy and will be covered in depth. As with similar industries, such as oil and gas, it is the hazard that has failed to be identified and treated that causes the accident. It is worth starting with some clarification: • •

A hazard is anything that may cause harm, such as electricity, working at height or rotating equipment. The risk is the chance, high or low, that someone could be harmed by these and other hazards, together with an indication of how serious the harm could be, for example whether an event could cause a single or multiple fatalities.

In simple terms the risk assessment process consists of: • • •

Identifying the hazards; Deciding on who might be harmed and how; Evaluating the risk and deciding on precautions.

There is also a requirement to record significant findings and to put in a process to review the assessment on a regular basis and following an incident or a major change in what has being assessed. Turning to the systematic identification of hazards. There are currently no guidelines of what constitutes a renewable energy hazard list. However, there are industries with similar profiles. For example, ISO 17776 has been developed to aid the oil and gas industry. Although offshore renewables occupy a similar geographical footprint it does not have the petrochemical risk and personnel tend to be dispersed and work in smaller groupings. The following gives an outline of the likely hazard areas: The risk associated with the product: • •

Wind turbine Transformer platforms. • In both cases, these can be further divided into control of hazardous energy: - Fire - Electrical - Mechanical - Hydraulics and pneumatics.

The geographical and environmental impact of placing the product offshore: • •

Vessel and aircraft collision Structural failure.

The human interaction with the product: • • • •

Working at height Lifting Marine and aviation access Occupational activity.

580

Offshore Wind Farms

Activities to mitigate risk or to respond to incidents: • • • •

Offshore medical fitness Offshore first aid training Offshore emergency response Onshore integrated emergency response planning.

Trade bodies provide industry guidance and customer groupings also work together to establish best working practices. The RenewableUK portal www.renewableuk.com provides access to health and safety documents. Their document, Offshore Wind and Marine Energy Health and Safety Guidelines, gives a structured approach to identifying the possible risks that could be expected and are split into 24 major categories from Access and Egress to Work at Height. The offshore renewable industry is primarily a danger to its own employees; interaction with the public is limited. Nevertheless, consideration has to be given to the unplanned visitor to a WTG. Here there is a conflict between the requirement to provide refuge for a mariner in distress and the physical control of access to a WTG. Turning to risk assessments. The HSE Information Sheet e Guidance on Risk Assessments for Offshore Installations, outlines the procedures that may be followed. Although specifically developed with the oil and gas industry in mind, the key principles apply equally well to the renewable energy industry. In evaluating risk, it states, the risk assessment methodology applied should be efficient (cost-effective) and of sufficient detail to enable the ranking of risks in order, for subsequent consideration of risk reduction. The rigour of assessment should be proportionate to the complexity of the problem and the magnitude of risk. Risk assessments can progress through the following different tiers of assessment: • • •

Qualitative (Q), in which frequency and severity are determined purely qualitatively. Semiquantitative (SQ), in which frequency and severity are approximately quantified within ranges. Quantified risk assessment (QRA), in which quantification occurs.

This approach ensures that time and effort are expended in appropriate areas. In the renewable industry, Q assessment is the norm as most hazards are understood and there is considerable experience in how to control these risks from other industries. Examples include heavy lifting, working at height, electricity and rotating machinery. Such an assessment would be conducted in-house and would form the vast majority of assessments undertaken. The risk can be plotted on a risk assessment matrix, which can be superimposed by the risk tolerance of the organization. For example, in the Risk Assessment Matrix in Fig. 19.3, any risk within the red area would be unacceptable and the activity would not be contemplated until mitigation measures were applied that either reduced the severity or probability of the event occurring. Areas in green would be deemed as under control and no further actions would be required. The yellow area would be the subjective area, sometimes referred to as the As Low as Reasonably Practical (ALARP) region. Here the residual risk is acceptable to the organization as any further activity to reduce risk would be deemed as requiring excessive effort measured in time and money. This recognizes that there


581

Severity

Risk assessment matrix Extensive Major Medium Minor

Categories Not acceptable

No impact

Very

Likely

Possible

Unlikely

Highly

ALARP Acceptable

Probability

Figure 19.3 Risk Assessment Matrix.

will always be pressure on resources and by comparing residual risk against other risks, attention is placed on where best value for money can be achieved. Nevertheless, this heightened risk remains and business takes the responsibility that the level of risk is within the business’ tolerability. As this is a very subjective decision it is essential that this decision is made by the risk owner. In some organizations the limit of authority to accept such risk varies within the company, with the highest risk being approved at the highest level, for example the Chief Executive Officer. There are some areas where the consequence of a hazard may be more severe and with historic information from such a new industry lacking, an SQ assessment may be appropriate. The risk from fire being an example, where the consequences to an individual are understood but the probability of these occurring is not clear. In such a situation detailed computer modelling may be considered to populate an SQ assessment. These may be conducted in-house, but may also require specialist support. At the top end where consequences may be very severe, possibly involving multiple fatalities and the hazard could affect the general public, then a QRA conducted by a specialist organization will be undertaken. An example would be the risk associated with the loss of a blade from a WTG. Once the risk is identified, quantified and prioritized it should be treated to bring the residual risk down to an acceptable level. The following is a recommended hierarchical approach to risk reduction: • • • • • •

Elimination and minimization of hazards by design (also known as safer by design); Prevention (reduction of likelihood occurring); Detection (raising the warning); Control (limitation of scale, intensity and duration); Mitigation of consequence (protection from effects); and Escape, evacuation and rescue (EER).

582

Offshore Wind Farms

Advice on risk mitigation may be found from many organizations. The following are examples of organizations that provide specific advice on key hazards; the list if far from exhaustive.

19.8.1 • •

HSE e Safe Use of Work Equipment e Provision and Use of Work Equipment 1998 (commonly known as PUWER 1998). Approved Code of Practice and Guidance HSE e Memorandum of guidance on the Electricity at Work Regulations 1989 e Guidance on regulations Guidance on regulations

19.8.2 • • •

• •

•

Occupational hazards

HSE e Manual handling assessment charts (the MAC Tool) Guidance on regulations HSE e Safe Use of Lifting Equipment e Lifting Operations and Lifting Equipment Regulations 1998 (commonly known as LOLER 1998) Approved Code of Practice and Guidance G9 e Good Practice Guideline e Working at Height in the Offshore Wind Industry G9 e Good Practice Guideline e The Safe Management of Small Service Vessels used in the Offshore Wind Industry

19.8.4 • •

Marine hazards

IMCA e Guidance on the Transfer to and from Offshore Vessels and Structures IMCA e Guidelines for Lifting Operations MCA and NWA e The Workboat Code e International Working Group Technical Standard

19.8.3 • •

WTG hazards

Emergency preparation and response

RenewableUK-Incident Response: Offshore Wind and Marine Projects* MCA e MGN371 e Offshore Renewable Energy Installations (OREIs) e Guidance on UK Navigational Practice, Safety and Emergency Response Issue* MCA e Offshore Renewable Energy Installations. Emergency Response Co-operation Plans (ERCoP) for Construction and Operations Phase, and Requirements for Emergency Response and SAR Helicopter Operations* • *These documents are to be replaced by Integrated Emergency Response e Renewables (IOER-R) in late 2015.

Nevertheless, the above is only advice and the final decision on how risk is to be treated rests with the management responsible for putting individuals in harm’s way. Organise activities to deliver the plan. In the process of determining how the risk is to be controlled it is important to involve workers and communicate, so that everyone is clear on what is needed and can openly discuss issues. This assists in developing positive attitudes and behaviours. Management must provide adequate resources to deliver the plan, including specialist advice where needed.


583

Implement the plan. In implementing the plan management must: • • •

•

Decide the preventative and protective measures needed and put them in place. Such information should be clearly published to enable a safe system of work to be established. The RenewableUK e Wind Turbine Safety Rules are a good example of such procedures. Provide the right tools and equipment to do the job and keep them maintained. Train and instruct, to ensure that everyone is competent to carry out their work. As the renewable energy workforce tends to move from project to project, the main industry leaders have standardized the minimum safety training required to work within the industry. Information can be found through the Global Wind Organization portal at http://www.globalwindsafety.org/. Supervise to make sure that arrangements are followed.

Within the offshore renewable industry, managing contractors is a special cause for concern. Contractors have a higher accident rate than core employees. The reason for this can be many-faceted and each situation must be assessed on its own merit and special procedures will need to be developed to monitor and control the situation. This may require external specialist support where in-house experience is lacking. A key component of any plan is to have contingency measures in place should any incident occur. Within the renewable wind industry events can occur organically from within the wind farm’s activity or a wind farm can also collect events from passing or adjoining general marine activity. The United Kingdom’s emergency services operate a three-level escalatory response to any incident, see Fig. 19.4. The activity related to dealing with the incident at its source is deemed to be at the operational level. Where the incident requires the coordination of a number of responding assets, this level of response is defined as the tactical level. Where an incident is likely to exceed predefined contingencies, then it can be escalated to the strategic level (sometimes also known as a crisis response)

Gold crisis or strategic level Silver tactical level ERCoP - MRCC Bronze operational ERP – windfarm

Figure 19.4 Strategic, Tactical and Operational levels of Command and Control e https:// www.app.college.police.uk/app-content/operations/command-and-control/definitions-andprocedures/.

584

Offshore Wind Farms

so that all the organization’s assets may respond. These three levels are also referred to as Bronze, Silver and Gold. The regulatory authorities expect windfarms to have the ability to immediately respond to foreseeable events. Historic evidence shows that individuals suffer life-threatening medical emergencies, eg heart attack, industrial accidents affecting limbs and ergonomic events leading to strains and dislocations. Multiple casualties are rare, although marine-related events have seen personnel forced to enter life rafts, with a risk of an unintended water entry, exposure and drowning. All offshore incidents eventually reach shore and recovery to the mainland should be factored into all response plans. This may require the creation of reception centres, next of kin notification and survivor support. Liaison officers and communication specialists are most likely to be also required. How a windfarm principal duty holder prepares to respond to an incident will be determined by the likely magnitude of the incident and assets that are available to respond. This will change as the windfarm evolves through the lifecycle of the project. Accordingly, any assessment and response plans will require updating as risk and responding assets change. Activity at the operational level is detailed within the principle duty holders’ Emergency Response Plan. This should cover all activity and all personnel, including subcontractors, within a defined windfarm. This clarity of command cannot be overemphasized and is one of the key lessons identified from the oil and gas industry. Should the incident require a complex response or external assistance, it should be escalated to the tactical level. This would be in the first instance to the MCA Maritime Rescue Coordination Centre (MRCC) and would be in accordance to the predetermined Emergency Response Cooperation Plan (ERCoP). It is mandatory for ERCoPs to be established before wind farm construction commences and for them to be updated as the wind farm progresses into Operation and Maintenance. The full requirement for ERCoPs can be found at MCA e Offshore Renewable Energy Installations. Emergency Response Co-operation Plans (ERCoP). ERCoPs are documents jointly prepared by the principal duty holder and the MCA, they cannot be created in isolation and therefore ensure face-to-face dialogue. Where there is a risk to life or where there is a risk to the quality of life, the incident should be immediately escalated to the MCA in accordance with the ERCoP. The MCA will then take responsibility for the coordination of the incident and determine whether national search and rescue assets should be deployed. In a medical incident the MCA will seek telemedicine advice on the urgency and efficacy of an immediate SAR helicopter transfer. On occasions the windfarm assets may still be the preferred means of transportation; however, coordination with other emergency services, eg the ambulance authority, will remain the responsibility of the MCA. Where an incident is greater than the capability of the windfarm and normal standby national emergency response, then the incident may be escalated to a strategic/crisis incident. It is difficult to imagine a windfarm originating an incident that would require such an escalation. It is more likely to arise from a marine incident occurring within or near the windfarm. All mariners are obliged to respond to personnel in distress at sea. As windfarms progress further offshore and into oil and gas territory, geographical


585

integrated response procedures will be established, detailing how the different stakeholders will provide mutual support. Any incident that causes loss of life will also be investigated by the police. This can require the interviewing of witnesses and fact-finding visits to OREI.

19.9

Check

Measure your performance. A safety management system is only as good as the feedback it generates. Monitoring can be broken into ‘active’ or ‘reactive’. Active include routine inspections, health surveillance and safety equipment checks. Reactive methods include investigating accidents and incidents and monitoring causes of ill health and sickness absence records. Active monitoring is difficult in the offshore renewable industry. Work teams are small and work independently from senior supervision. In addition, no-notice visits are difficult to conduct as marine logistics require prior notification. Accordingly, self-supervision with remote monitoring of activity is becoming commonplace. Investigate the causes of accidents, incidents or near misses. The key to reactive monitoring is to choose the correct level of investigation. There is little to be gained in completing a full-blown investigation when the injured party and supervisor are fully aware of the cause and the corrective actions required. Investigating a near miss that could have had severe consequences may be more beneficial. Accident investigation requires a book on its own. However, care has to be taken to identify what went wrong and not who did wrong. A hunt for a culprit is a sure way of ensuring that incidents and near misses are never reported. Such a culture is known as a hanging culture and can be very corrosive. The opposite approach is the non-blameworthy culture, where any event is forgiven, as long as it is reported. This also has its issues, with wilful breaches of procedures seen as having no repercussions, thereby encouraging unsafe actions. A middle ground has been christened a Just Culture, where honest mistakes are recognized as more of a systems failure than a personal failure. However, wilful breach of safety procedures will not be tolerated and will be subject to disciplinary action. On occasions management may perceive wilful breach, but the individual may claim a mistake. On these occasions it might be useful to look at the incident from the eye of work colleagues, who may have also experienced a similar situation, albeit without the unfortunate consequences. For a Just Culture to succeed it must be seen to be fair by the work force. There are a number of tools and methodologies available to assist in investigating an accident, incident or near-miss. They help in identifying the root cause of any incident. Management oversight is often credited as a root cause or at least a contributory factor. After all, not many individuals come to work to harm themselves. An in-depth investigation can identify many shortcomings within an organization; however, they can be time-consuming and costly. It is therefore imperative that any lessons are acted upon and that the information gained is shared with the workforce in an open and honest manner, however uncomfortable the findings may be to management.

586

19.10

Offshore Wind Farms

Act

Review performance. All employees should have the means to report accidents, incidents and near-misses. These can be as simple as verbally reporting to line management, to IT solutions that record, distribute and process all information into user-friendly categories that support trend analysis. The larger the organization the stronger the requirement for such IT systems. In addition, trade organizations, such as IMCA, G9 and RenewableUK, run industry-specific reporting systems, that attempt to collate information across the whole industry so that collective improvement can be instigated. Regulatory authorities also require the mandatory reporting of accidents and incidents above a certain threshold. Where they believe that action is required they can raise an improvement notice and where the authorities believe there could be a risk of serious personal injury then an enforcement notice requiring the cessation of the activity can be raised. Take action on lessons learned, including from audit and inspection reports. Lessons identified must be converted into lessons learned. Following the systematic identification of the root cause, a corrective action (an attempt to prevent reoccurrence) or a preventative action (to stop the event occurring) should be applied. Collectively these actions are known as Corrective Actions Preventative Actions (CAPA) and organizations have systems in place to ensure that CAPA owners are identified and those actions are monitored and closed.

19.11

For the future

With wind farms becoming larger and further offshore, personnel are being employed permanently offshore for long periods, either within vessels (hotel ships and service operation vessels) or fixed offshore transformer platforms. The oil and gas industry has been in a similar situation for decades. However, the current renewable industry approach to risk management may not stand up to the demands of planned activity. History has shown that incidents occur at the interaction and boundaries between risks and risk owners. This can be well-illustrated by the tragic Piper Alpha Platform accident in 1988, which led to the death of 167 workers. Lord Cullen led a review into causes of the disaster and to recommendations for the changes to the safety regime. The cause was identified as a hydrocarbon leak following the removal, blanking of a pump and then recharging of the pipe. It is believed that the consequence of pump removal was not fully understood by those responsible for the restart. Lord Cullen’s report resulted in the oil and gas industry adopting a Safety Case approach to managing risk. The safety case is required to provide full details of the arrangements for managing health and safety and controlling major accidents on an installation. This umbrella approach is intended to ensure that risks are not missed, especially where activities and risk owners meet. Safety Cases were mandated by law for the oil and gas industry; however, they are currently not a requirement for the offshore renewable industry. However, it may become very difficult to justify that a principal duty holder has risk controlled without adopting such a process.


19.12

587

Conclusion

If you want to be 100% risk-free; then avoid offshore activity. However, large WTG work well in the stable and strong wind environment that exists offshore. Those that manage the risk well will have a commercial advantage. Risk management is built on detailed hazard identification. This activity must commence at the beginning of a project and requires a systematic approach; it is not an afterthought. Only with an early understanding of the risk exposure can those accountable for the activity put control measures in place at the design phase; the only time where elimination of a risk is a possibility. Although a young industry, it has made tremendous progress through adopting and adapting process and procedures from similar industries. With strong leadership, management and a well-trained and dedicated workforce, it is possible to go to work offshore and return home safely, but it requires everyone involved, from designers, supply chain managers, lifting supervisors, marine coordinators, operational managers and engineers to all play their part. Safety is a team game, even more so when played offshore.

Abbreviations ALARP

As low as reasonably practical

CAPA

Corrective actions preventative actions

CTV

Crew transfer vessels

DRA

Design risk assessments

EER

Escape, evacuation and rescue

ERCoP

Emergency response cooperation plan

HSE

Health and safety executive

MAIB

The marine accident investigation branch

MCA

The maritime and coastguard agency

MRCC

Maritime rescue coordination centre

O&M

Operation and maintenance

OEM

Original equipment manufacturers

OREI

Offshore renewable energy installation

Q

Qualitative

QRA

Quantified risk assessment

SMS

Safety management system

SQ

Semiquantitative

WTG

Wind turbine generator


Offshore wind turbine foundations e analysis and design

20

B.C. O’Kelly Trinity College Dublin, Dublin, Ireland M. Arshad University of Engineering & Technology, Lahore, Pakistan

20.1

Foundation options for offshore wind-turbine structures

The economically viable development of wind-farms depends on efficient solutions being available for a number of technical issues, one aspect being the foundations. The foundation choice largely depends on water depth, seabed characteristics, applied loading, available construction technologies and, importantly, economic costs (Malhotra, 2010). Offshore wind turbine (OWT) structures may be found on gravity base, suction caisson, monopile, tripod or jacket/lattice foundations (collectively categorized as bottom-mounted structures) or using floating platforms tethered to the seabed (Fig. 20.1). The most widely adopted foundation choice in terms of its ease of installation, economy and logistics is the monopile, a single large-diameter hollow (pipe) pile; with an estimated 75% of all installed OWTs employing this solution (Blanco, 2009; Fischer, 2011; Malhotra, 2010). Hence, the focus of this chapter is on the geotechnical aspects of monopile foundations, with some discussion also provided on the other main foundation options. Monopiles are typically used in shallow water depths (ie, 20e40 m), but may become too flexible for water depths of between w30 and 40 m, in which case monopiles fitted with guy wires, or alternatively tripod and jacket/lattice structures, are considered as economical alternatives. For greater depths, time-consuming installation and the effect of soil degradation, which occurs around the pile shaft at seabed level, make monopile foundation solutions prohibitive (Irvine et al., 2003). Tripods consist of a large-diameter central, steel tubular section that is supported over its lower length by three braces (Fig. 20.1(d)), which are connected to the seabed using different foundation types, including gravity base, suction bucket or piles. In this manner, the loads applied to the OWT and its support structure are mostly transferred axially (via the braces) to the seabed foundation. Complete installation of a tripod foundation system having, for example, a water surface to seabed length of up to 50 m typically takes two to three working days, often


590

Offshore Wind Farms

(a)

(b)

(c)

(d)

(e)

(f)

(g)

Sea level 0

Water depth (m)

10

Seabed

20 30 40 50 60 70 80

a) Gravity b) Monopile c) Monopile with guy wires d) Tripod e) Jacket/lattice structure f) Tension leg with suction buckets (ballast stabilized) g) Buoy with suction anchor

Figure 20.1 Support structure options showing typical ranges of water-depth application (adapted from Malhotra, 2010).

requiring special equipment for driving/drilling and working underwater (Esteban et al., 2011; Fischer, 2011). A jacket structure (see Fig. 20.1(e)) is a lattice frame comprising small-diameter steel struts that, similar to the tripod, is anchored to the seabed using different foundation types. Complete installation of the jacket structure generally takes up to 3 days. These structures are particularly suited for severe maritime weather conditions because of the additional structural stiffness and larger moment arm (across the seabed) for reacting against the bending loads, compared with monopile foundations. Jacket/lattice structures are also more adaptable to the conditions encountered onsite, increasing their application range, with geometrical variations of the substructure achieved relatively simply, but without altering the stiffness of the overall structure (De Vries, 2007). It is estimated that by 2020, 50e60% of OWTs will be supported by monopiles, with a further 35e40% supported by jacket and tripod systems (Babcock and Brown Company, 2012). The main reason for this shift is the attraction of jacket and tripod systems for deeper sea locations that provide consistently higher wind speeds and hence greater wind-energy production, which is a cube function of the wind speed (Tempel and Molenaar, 2002).


591

In the future, it is anticipated that floating structures, which are currently at the research and development stage, will be commercially used, particularly for water depths greater than 50 m (Saleem, 2011). Such floating platforms for wind turbines will impose many new design challenges. Among these, tension-leg platform concepts (Fig. 20.1(f)) are currently considered as more economical (Fischer, 2011) since the rigid body modes of the floater are limited to horizontal translation (surge and sway) and rotation around the vertical axis (yaw). For spare floater systems (Fig. 20.1(g)), buoyancy is provided to the wind turbine structure by a long slender cylinder/capsule that protrudes well below the water line (De Vries, 2007; Esteban et al., 2011; Fischer, 2011). For barrage floater systems, the wind turbine structure is placed on a barrage and attached, via anchor lines, to the seabed. The design of floating offshore wind turbines is discussed in chapter ‘Energy storage for offshore wind farms’.

20.2

System of loading on offshore foundations

Offshore foundations are subjected to a combination of loading, namely: axial (self weight) forces of the structure/machinery; repeating horizontal/lateral loads; bending and torsional moments. Apart from the self-weight forces, this system of loading is generated by environmental conditions and the operation of the machinery (Fig. 20.2). The foundations must be designed to resist large numbers of wind and

Ambient turbulence Wake turbulence Operational and accidental loads

Icing

Marine growth

Waves

Currents Seabed

Figure 20.2 Environmental impacts on offshore wind turbines.

Scour

592

Offshore Wind Farms

hydrodynamic (ocean waves, current flow and tidal/swell action) load cycles of varying direction, amplitude and frequency occurring at the proposed site over the project’s lifetime of typically 25 years or more (Sahin, 2004). Another variable/cyclic load acting on OWT structures, depending on the geographic setting, is from ice sheets. Seismic loading is considered as a special type of dynamic loading. However, the focus of this chapter is on wind and wave loading since these are considered more important in the assessment of the OWT structure’s fatigue life. The wave loading (forces) acting on OWT structures is often greater than the wind loading. However, in terms of the overturning (bending) moments generated, the wave loading generally has only a minor role, compared with the rotorethrust reaction to the wind loads. This occurs due to the smaller lever arm for the bending moment generated by the wave loading, as compared with the overall tower length, and longer lever arm of the rotor thrust in considering the overall overturning moment acting on the foundation system (LeBlance, 2009). For instance, for OWT monopiles installed in the North Sea, Byrne and Houlsby (2003) reported that the rotor thrust reaction contributed to approximately 25% of the total horizontal load, but generated approximately 75% of the total overturning moment. The density of the medium must also be considered when comparing wind and wave loading, with the density of sea water significantly greater than that of air. Hydrodynamic loads generally only become significant for greater water depths and (or) wave heights, which cause the lever arm of the wave load to increase, along with the intensity of the lateral load generated by the water (Fischer, 2011). Fluctuations in wind speed about a mean value impose repeated aerodynamic loads, although when considering the dynamic behaviour of offshore structures, this cyclic nature is generally insignificant compared with the repeated wave loads (Journée and Massie, 2001). Dynamic analysis of offshore structures that takes into account the fluctuating wind load is necessary in cases where the wind field contains energy at frequencies near the offshore structure’s natural frequency, although for monopile foundations, the difference between these frequencies is usually high. When the loading frequency gets closer to the structure’s natural frequency of vibration (API, 2010), the repeating load can be termed dynamic load. This tends to excite the structure dynamically, leading to resonance and the development of higher stresses in the support structure and foundation, but more significantly to a higher range of stresses; an unfavourable situation in considering fatigue life. A correct evaluation of the total hydrodynamic load acting on an offshore structure must consider the combined current flow and wave particle velocities. Morison et al. (1950) formulated an equation to predict the wave loads acting on a vertical pile exposed to horizontal sinusoidal oscillatory flow. In their equation, the linear inertial force and adapted quadratic drag force (from real flows and constant currents) are superimposed to obtain the resultant force acting on the projected area of the pile. Morison’s formula is strictly limited for use with slender structural elements, characterized by D/l < 0.2, where D is the diameter of the structural element between seabed level and the transition piece, and l is the impinging wavelength of the ocean wave. For larger offshore structures (eg, gravity foundations and OWT monopiles), the wave field is significantly influenced and a diffraction regime emerges. Potential


593

flow theory is more suitable for the calculation of wave loads on such structures (Batchelor, 1967). However, a significant number of existing offshore structure designs have employed Morison’s equation even though the criterion of D/l < 0.2 may not have been fully satisfied (Haritos, 2007). For geotechnical design, the relative proportions and importance of the different types of loads applied essentially depend upon the kind of foundation system being considered. For gravity base foundations, potential failure modes are in bearing capacity or excessive settlement; hence the vertical (self-weight) loads are generally the major design consideration (Malhotra, 2010). For monopile foundations, the lateral deflection (rotation) response of the monopile largely controls the serviceability limit state of the whole structure. Hence, lateral loads and resulting moments are more critical compared with the vertical loads for monopile foundations. In other words, the monopile’s response under repeated lateral loading is the major design consideration, with the monopile design dominated by considerations of its dynamic and fatigue responses under working loads, rather than its ultimate load-carrying capacity. For instance, existing OWTs have rotor diameters ranging between w90 and 120 m (power-generation capacities of 3e6 MW (Tong, 2010)) and produce gravitational loading in the range of w2e8 MN. For instance, Byrne and Houlsby (2003) reported vertical loading of 6 MN acting on the monopile foundation for an anticipated 3.5 MW OWT located in the North Sea, with lateral loading from wind and wave factors accounting for up to 66% of the vertical loading. The precise magnitude of these loads will vary with the size of the installation, the detailed design, and local environmental conditions. This scenario is more onerous when the repeating lateral loading occurs at varying frequency, load amplitude and direction (Arshad and O’Kelly, 2013). At some critical level of load amplitude and (or) frequency, the repeating lateral loads can cause significant reductions in the lateral soil resistance for a monopile foundation (Ramakrishna and Rao, 1999).

20.3

General aspects of OWT monopile foundation system

Monopiles with outer diameters of 4e6 m have been successfully installed, with embedment (penetration) depths of between 20 and 40 m, depending upon the wind-power generation capacity of the OWT (Peng et al., 2006; Achmus et al., 2009; LeBlanc et al., 2010; Peng et al., 2010; Cuéllar et al., 2012; Pappusetty and Pando, 2013). These monopiles are manufactured from steel tubular sections with wall thicknesses (depending on installation and loading conditions) ranging from 55 to 150 mm (Haiderali et al., 2013), but more frequently 60e80 mm, and have an overall mass of up to 1000 tonnes. Depending on soil conditions, monopiles are typically installed in the seabed using largeimpact hammers, by vibratory pile driving (vibropiling), prebored piling, or drill/driving techniques (Malhotra, 2010; Igoe et al., 2013a). Noise-mitigation measures, including bubble curtains, isolation casings (pile sleeve), dewatered cofferdams, and hydro sound dampers, are employed during pile driving in order to limit emissions to environmentally friendly levels (Saleem, 2011; OSPAR

594

Offshore Wind Farms

Commision, 2014). Complete installation of the monopile is usually achieved within 24 h (Junginger et al., 2004; Fischer, 2011; Saleem, 2011). The transition piece of the OWT structure provides a means of correcting vertical imprecisions of the monopile that may have arisen during its installation. The transition piece, which has a slightly larger (or smaller) diameter than the monopile, is grouted to the monopile, with an overlap. Mechanical shear connectors increase the reliability of this connection and alleviate the effect of long-term grout shrinkage on the connection capacity. New design features include grouted conical-shaped connections and shear keys (DNV, 2011). In some instances, the installed monopile can extend above the water surface, connecting directly to the tower. The monopile diameter and embedment length are primarily dependent on the OWT’s power-generation capacity (an indirect measure of the applied loading), seabed characteristics/soil properties and severity of environmental loading. OWT monopiles invariably have slenderness (in terms of embedment length to diameter) ratio values of less than 10 (typically ranging from 5 to 6), and the embedded portion of the monopile is therefore considered to behave as a ‘rigid’ structure, for which rotation is prominent over bending (Tomlinson, 2001; Peng et al., 2011). In other words, the surrounding soil would fail in bearing capacity rather than the monopile failing over its embedded length by plastic hinges; ie, its structural behaviour under lateral loading is defined by its rotation as a rigid body. The embedment length must be sufficient for the monopile to meet design criteria, including vertical stability and limiting horizontal deflection/ rotation over its design life. Limiting the rotation is more important for ‘rigid’ monopiles. As a general rule, under field loading, rotation of the monopile by up to 0.5 degree from its vertical alignment (Achmus et al., 2009; LeBlanc, 2009; LeBlanc et al., 2010; DNV, 2011) or lateral deflections occurring at seabed level of up to 120 mm (based on practical experience) (Malhotra, 2010) are considered as limiting values for the proper operation of the wind turbine. Transport of sediment from beneath the scour protection (typically rock/stone layers) zone around OWT monopiles may cause sinking of the scour protection, particularly for piles founded in sandy soil deposits. This alters the natural frequency of the dynamic response in an unfavourable manner (van der Tempel et al., 2004). For a typical 5-MW OWT installed in the North Sea, a hub height of w95 m above mean sea level and rotor diameter of 125 m would produce the approximate quasistatic loading scenario (acting at the seabed level) given in Table 20.1. This example scenario demonstrates that for the monopile foundation, horizontal loading from wind and wave causes extremely high bending moments (resisted in lateral compression of Table 20.1 Loading at seabed level for monopile supporting 5-MW OWT (Lesny and Wiemann, 2005) Axial load

35 MN

Horizontal load

16 MN

Bending moment

562 MN m

Torsional moment

4 MN m


595

the surrounding soil) and principally controls the foundation design. To ensure the monopile is torsionally stable, sufficient circumferential shear resistance must be mobilized at the pileesoil interface, although the torsional moment to be resisted is usually small (Table 20.1). Further, the connections between the tower and transition piece and between the transition piece and monopile foundation must be capable of transferring these bending and torsional moments, with adequate factors of safety.

20.4

Offshore design codes and methods

At present, current practice for the analysis, design and installation of OWT monopiles usually relies on general geotechnical standards which are complemented by more specific guidelines and semiempirical formulae developed mainly by the offshore oil/gas industries (American Petroleum Institute (API), 2010; DIN, 2005; Det Norske Veritas (DNV), 2011). However, large-diameter monopile foundations for current and future OWT structures are well outside the scope of current experience and analysis/design methods, including the API (2010) and DNV (2011) recommended practices, which are largely based on limited field data obtained for relatively small-diameter (ie, flexible) piles under low numbers of load cycles. For these standards, wave loading is of primary concern when extrapolating to predict extreme events. However, designers of OWT structures must consider wave and wind load spectra simultaneously (IEC, 2005). Hence, careful consideration of these differences in applied loading, as well as other inherent limitations underlying semi-empirical offshore oil/gas industry formulae, is required in extrapolation of these formulations for the design of OWT monopile foundations. Often these formulations cannot be applied with confidence by the offshore wind-power industry to achieve optimum results and economy (Dobry et al., 1982). Current design standards and guidelines on the serviceability of piles under cyclic lateral loading are limited. Over its service life, a typical 2-MW OWT structure is subjected to w100 cycles of 2.0-MN magnitude and 107 cycles of 1.4-MN magnitude in lateral loading, which correspond to the serviceability limit state and fatigue limit state, respectively (GL, 2005). Factors affecting the cyclic response include the diameter and wall thickness of the monopile, its free spanning and embedded lengths, the soil properties, soilepile relative stiffness, the characteristics of the applied loading and the pile installation method (Malhotra, 2010).

20.5

Investigation of monopileesoil behaviour

20.5.1 Soil behaviour and testing Apart from very small strain levels of