Power Grids With Renewable Energy

IET ENERGY ENGINEERING SERIES 67

Power Grids with

Renewable Energy
Other volumes in this series
Volume 1 Power Circuit Breaker Theory and Design C.H. Flurscheim (Editor)
Volume 4 Industrial Microwave Heating A.C. Metaxas and R.J. Meredith
Volume 7 Insulators for High Voltages J.S.T. Looms
Volume 8 Variable Frequency AC Motor Drive Systems D. Finney
Volume 10 SF6 Switchgear H.M. Ryan and G.R. Jones
Volume 11 Conduction and Induction Heating E.J. Davies
Volume 13 Statistical Techniques for High Voltage Engineering W. Hauschild and
W. Mosch
Volume 14 Uninterruptible Power Supplies J. Platts and J.D. St Aubyn (Editors)
Volume 15 Digital Protection for Power Systems A.T. Johns and S.K. Salman
Volume 16 Electricity Economics and Planning T.W. Berrie
Volume 18 Vacuum Switchgear A. Greenwood
Volume 19 Electrical Safety: A Guide to Causes and Prevention of Hazards
J. Maxwell Adams
Volume 21 Electricity Distribution Network Design, 2nd Edition E. Lakervi and
E.J. Holmes
Volume 22 Artificial Intelligence Techniques in Power Systems K. Warwick, A.O. Ekwue
and R. Aggarwal (Editors)
Volume 24 Power System Commissioning and Maintenance Practice K. Harker
Volume 25 Engineers’ Handbook of Industrial Microwave Heating R.J. Meredith
Volume 26 Small Electric Motors H. Moczala et al.
Volume 27 AC–DC Power System Analysis J. Arrillaga and B.C. Smith
Volume 29 High Voltage Direct Current Transmission, 2nd Edition J. Arrillaga
Volume 30 Flexible AC Transmission Systems (FACTS) Y.-H. Song (Editor)
Volume 31 Embedded Generation N. Jenkins et al.
Volume 32 High Voltage Engineering and Testing, 2nd Edition H.M. Ryan (Editor)
Volume 33 Overvoltage Protection of Low-Voltage Systems, Revised Edition P. Hasse
Volume 36 Voltage Quality in Electrical Power Systems J. Schlabbach et al.
Volume 37 Electrical Steels for Rotating Machines P. Beckley
Volume 38 The Electric Car: Development and Future of Battery, Hybrid and Fuel-cell
Cars M. Westbrook
Volume 39 Power Systems Electromagnetic Transients Simulation J. Arrillaga and
N. Watson
Volume 40 Advances in High Voltage Engineering M. Haddad and D. Warne
Volume 41 Electrical Operation of Electrostatic Precipitators K. Parker
Volume 43 Thermal Power Plant Simulation and Control D. Flynn
Volume 44 Economic Evaluation of Projects in the Electricity Supply Industry
H. Khatib
Volume 45 Propulsion Systems for Hybrid Vehicles J. Miller
Volume 46 Distribution Switchgear S. Stewart
Volume 47 Protection of Electricity Distribution Networks, 2nd Edition J. Gers and
E. Holmes
Volume 48 Wood Pole Overhead Lines B. Wareing
Volume 49 Electric Fuses, 3rd Edition A. Wright and G. Newbery
Volume 50 Wind Power Integration: Connection and System Operational Aspects
B. Fox et al.
Volume 51 Short Circuit Currents J. Schlabbach
Volume 52 Nuclear Power J. Wood
Volume 53 Condition Assessment of High Voltage Insulation in Power System
Equipment R.E. James and Q. Su
Volume 55 Local Energy: Distributed generation of heat and power J. Wood
Volume 56 Condition Monitoring of Rotating Electrical Machines P. Tavner, L. Ran,
J. Penman and H. Sedding
Volume 57 The Control Techniques Drives and Controls Handbook, 2nd Edition
B. Drury
Volume 58 Lightning Protection V. Cooray (Editor)
Volume 59 Ultracapacitor Applications J.M. Miller
Volume 62 Lightning Electromagnetics V. Cooray
Volume 63 Energy Storage for Power Systems, 2nd Edition A. Ter-Gazarian
Volume 65 Protection of Electricity Distribution Networks, 3rd Edition J. Gers
Volume 66 High Voltage Engineering Testing, 3rd Edition H. Ryan (Editor)
Volume 67 Multicore Simulation of Power System Transients F.M. Uriate
Volume 68 Distribution System Analysis and Automation J. Gers
Volume 69 The Lightening Flash, 2nd Edition V. Cooray (Editor)
Volume 70 Economic Evaluation of Projects in the Electricity Supply Industry,
3rd Edition H. Khatib
Volume 72 Control Circuits in Power Electronics: Practical issues in design and
implementation M. Castilla (Editor)
Volume 73 Wide Area Monitoring, Protection and Control Systems: The Enabler for
Smarter Grids A. Vaccaro and A. Zobaa (Editors)
Volume 74 Power Electronic Converters and Systems: Frontiers and Applications
A.M. Trzynadlowski (Editor)
Volume 75 Power Distribution Automation B. Das (Editor)
Volume 76 Power System Stability: Modelling, Analysis and Control A.A. Sallam and
Om P. Malik
Volume 78 Numerical Analysis of Power System Transients and Dynamics A. Ametani
(Editor)
Volume 79 Vehicle-to-Grid: Linking electric vehicles to the smart grid J. Lu and
J. Hossain (Editors)
Volume 81 Cyber-Physical-Social Systems and Constructs in Electric Power
Engineering S. Suryanarayanan, R. Roche and T.M. Hansen (Editors)
Volume 82 Periodic Control of Power Electronic Converters F. Blaabjerg, K. Zhou,
D. Wang and Y. Yang
Volume 86 Advances in Power System Modelling, Control and Stability Analysis
F. Milano (Editor)
Volume 87 Cogeneration: Technologies, Optimisation and Implementation
C.A. Frangopoulos (Editor)
Volume 88 Smarter Energy: From Smart Metering to the Smart Grid H. Sun,
N. Hatziargyriou, H.V. Poor, L. Carpanini and M.A. Sánchez Fornié (Editors)
Volume 89 Hydrogen Production, Separation and Purification for Energy A. Basile,
F. Dalena, J. Tong and T.N. Veziroğlu (Editors)
Volume 90 Clean Energy Microgrids S. Obara and J. Morel (Editors)
Volume 91 Fuzzy Logic Control in Energy Systems with Design Applications in
MATLAB‡/Simulink‡ İ.H. Altaş
Volume 92 Power Quality in Future Electrical Power Systems A.F. Zobaa and
S.H.E.A. Aleem (Editors)
Volume 93 Cogeneration and District Energy Systems: Modelling, analysis and
optimization M.A. Rosen and S. Koohi-Fayegh
Volume 94 Introduction to the Smart Grid: Concepts, Technologies and Evolution
S.K. Salman
Volume 95 Communication, Control and Security Challenges for the Smart Grid
S.M. Muyeen and S. Rahman (Editors)
Volume 96 Industrial Power Systems with Distributed and Embedded Generation
R. Belu
Volume 97 Synchronized Phasor Measurements for Smart Grids M.J.B. Reddy and
D.K. Mohanta (Editors)
Volume 98 Large Scale Grid Integration of Renewable Energy Sources
A. Moreno-Munoz (Editor)
Volume 100 Modeling and Dynamic Behaviour of Hydropower Plants N. Kishor and
J. Fraile-Ardanuy (Editors)
Volume 101 Methane and Hydrogen for Energy Storage R. Carriveau and D.S.-K. Ting
Volume 104 Power Transformer Condition Monitoring and Diagnosis A. Abu-Siada
(Editor)
Volume 106 Surface Passivation of Industrial Crystalline Silicon Solar Cells J. John
(Editor)
Volume 107 Bifacial Photovoltaics: Technology, Applications and Economics J. Libal
and R. Kopecek (Editors)
Volume 108 Fault Diagnosis of Induction Motors J. Faiz, V. Ghorbanian and G. Joksimović
Volume 110 High Voltage Power Network Construction K. Harker
Volume 111 Energy Storage at Different Voltage Levels: Technology, Integration, and
Market Aspects A.F. Zobaa, P.F. Ribeiro, S.H.A. Aleem and S.N. Afifi (Editors)
Volume 112 Wireless Power Transfer: Theory, Technology and Application N.
Shinohara
Volume 115 DC Distribution Systems and Microgrids T. Dragičević, F. Blaabjerg and
P. Wheeler
Volume 117 Structural Control and Fault Detection of Wind Turbine Systems
H.R. Karimi
Volume 119 Thermal Power Plant Control and Instrumentation: The Control of Boilers
and HRSGs, 2nd Edition D. Lindsley, J. Grist and D. Parker
Volume 120 Fault Diagnosis for Robust Inverter Power Drives A. Ginart (Editor)
Volume 121 Monitoring and Control using Synchrophasors in Power Systems with
Renewables I. Kamwa and C. Lu (Editors)
Volume 123 Power Systems Electromagnetic Transients Simulation, 2nd Edition
N. Watson and J. Arrillaga
Volume 124 Power Market Transformation B. Murray
Volume 125 Wind Energy Modeling and Simulation Volume 1: Atmosphere and Plant
P. Veers (Editor)
Volume 126 Diagnosis and Fault Tolerance of Electrical Machines, Power Electronics
and Drives A.J.M. Cardoso
Volume 128 Characterization of Wide Bandgap Power Semiconductor Devices F. Wang,
Z. Zhang and E.A. Jones
Volume 129 Renewable Energy from the Oceans: From Wave, Tidal and Gradient
Systems to Offshore Wind and Solar D. Coiro and T. Sant (Editors)
Volume 130 Wind and Solar Based Energy Systems for Communities R. Carriveau and
D.S.-K. Ting (Editors)
Volume 131 Metaheuristic Optimization in Power Engineering J. Radosavljević
Volume 132 Power Line Communication Systems for Smart Grids I.R.S. Casella and
A. Anpalagan
Volume 139 Variability, Scalability and Stability of Microgrids S.M. Muyeen, S.M. Islam
and F. Blaabjerg (Editors)
Volume 145 Condition Monitoring of Rotating Electrical Machines P. Tavner, L. Ran,
C. Crabtree
Volume 146 Energy Storage for Power Systems, 3rd Edition A.G. Ter-Gazarian
Volume 147 Distribution Systems Analysis and Automation, 2nd Edition J. Gers
Volume 152 Power Electronic Devices: Applications, Failure Mechanisms and
Reliability F Iannuzzo (Editor)
Volume 155 Energy Generation and Efficiency Technologies for Green Residential
Buildings D. Ting and R. Carriveau (Editors)
Volume 157 Electrical Steels, 2 Volumes A. Moses, K. Jenkins, Philip Anderson and
H. Stanbury
Volume 158 Advanced Dielectric Materials for Electrostatic Capacitors Q Li (Editor)
Volume 159 Transforming the Grid Towards Fully Renewable Energy O. Probst,
S. Castellanos and R. Palacios (Editors)
Volume 160 Microgrids for Rural Areas: Research and Case Studies R.K. Chauhan,
K. Chauhan and S.N. Singh (Editors)
Volume 166 Advanced Characterization of Thin Film Solar Cells N. Haegel and
M. Al-Jassim (Editors)
Volume 172 Lighting Interaction with Power Systems, 2 Volumes A. Piantini (Editor)
Volume 905 Power System Protection, 4 Volumes
Power Grids with
Renewable Energy
Storage, integration and digitalization

Edited by
Abdelhay A. Sallam and Om P. Malik
The Institution of Engineering and Technology

Published by The Institution of Engineering and Technology, London, United Kingdom
The Institution of Engineering and Technology is registered as a Charity in England &
Wales (no. 211014) and Scotland (no. SC038698).
† The Institution of Engineering and Technology 2021
First published 2020
This publication is copyright under the Berne Convention and the Universal Copyright
Convention. All rights reserved. Apart from any fair dealing for the purposes of research
or private study, or criticism or review, as permitted under the Copyright, Designs and
Patents Act 1988, this publication may be reproduced, stored or transmitted, in any
form or by any means, only with the prior permission in writing of the publishers, or in
the case of reprographic reproduction in accordance with the terms of licences issued
by the Copyright Licensing Agency. Enquiries concerning reproduction outside those
terms should be sent to the publisher at the undermentioned address:
The Institution of Engineering and Technology

Michael Faraday House
Six Hills Way, Stevenage
Herts, SG1 2AY, United Kingdom
www.theiet.org
While the authors and publisher believe that the information and guidance given in this
work are correct, all parties must rely upon their own skill and judgement when making
use of them. Neither the authors nor publisher assumes any liability to anyone for any
loss or damage caused by any error or omission in the work, whether such an error or
omission is the result of negligence or any other cause. Any and all such liability is
disclaimed.
The moral rights of the authors to be identified as authors of this work have been
asserted by them in accordance with the Copyright, Designs and Patents Act 1988.
British Library Cataloguing in Publication Data

A catalogue record for this product is available from the British Library
ISBN 978-1-83953-027-2 (hardback)

ISBN 978-1-83953-028-9 (PDF)
Typeset in India by MPS Limited

Printed in the UK by CPI Group (UK) Ltd, Croydon
No significant work can be accomplished without the moral
support of the families and we wish to acknowledge that with
our appreciation and dedication
To our wives,
Hanzada Sallam and Margareta Malik
Contents
About the authors xix

Preface xxi
Acknowledgements xxv
1 Towards the new trend of power grids 1

1.1 Conventional power systems 2
1.2 Microgrids 4
1.3 Energy storage systems 5
1.4 Smarter grid 6
1.4.1 Consumers’ role 8
1.4.2 DSOs’ role 9
1.4.3 Smart grid definition 9
1.5 Power grid examination 12
References 12
Part I: Renewable energy sources 15

2 Wind energy 17
2.1 Wind turbines 17
2.1.1 Horizontal axis wind turbines 18
2.1.2 Vertical axis wind turbines 21
2.2 Basic relations 22
2.2.1 Power from wind 22
2.2.2 Factors affecting wind power 25
2.2.3 Average wind power 32
2.3 Wind characteristics 35
2.3.1 Airfoils 35
2.3.2 Capacity factor 39
2.4 Electrical generators in wind systems 41
2.4.1 Asynchronous induction generators 41
2.4.2 Permanent magnet synchronous machines 60
References 63
3 Solar energy 65
3.1 Introduction 65
x Power Grids with Renewable Energy
3.2 Solar PV technology 66

3.2.1 Basics 66
3.2.2 Energy loss in a solar cell 73
3.2.3 Electricity generation 75
3.2.4 PV cell equivalent circuit 76
3.2.5 Solar modules 81
3.2.6 Solar arrays 83
3.2.7 PV cell power curve 84
3.3 Concentrating solar power 91
3.3.1 Linear concentrator systems 92
3.3.2 Dish-engine systems 94
3.3.3 Power tower systems 95
References 99
4 Ocean energy: tidal energy 101

4.1 Introduction 101
4.2 Physics of tidal phenomena 101
4.2.1 Types of tides 104
4.3 Tidal energy estimation 105
4.3.1 Tide generating forces 106
4.3.2 Tidal wave 110
4.3.3 Tidal resonance 112
4.4 Tidal energy extraction 112
4.4.1 Tidal barrage 113
4.4.2 Tidal stream 119
4.5 Tidal current energy conversion devices 122
4.5.1 Tidal current turbines 122
4.5.2 Tidal current generators 123
4.5.3 Drawbacks of tidal stream generators 124
4.6 Tidal power plants connected to utility grid 124
4.6.1 Capacity limits 124
4.6.2 Power quality 124
4.6.3 System analysis 125
4.7 Environmental impacts 125
4.8 Tidal power pros and cons 125
4.8.1 Pros of tidal power 126
4.8.2 Cons of tidal power 126
References 127
5 Ocean energy: wave and thermal energy 129

5.1 Wave energy concept 129
5.2 Ocean waves creation 129
5.2.1 Ripples 130
5.2.2 Shallow waves or shore waves 133
Contents xi
5.3 Wave propagation 133

5.4 Fundamentals of deepwater waves 137
5.5 Estimation of wave energy 139
5.5.1 Derivation of power formulae 142
5.6 Wave energy converters 143
5.6.1 Location of WECs 144
5.6.2 Types of WECs 144
5.6.3 Modes of operation 147
5.7 Mechanical interfaces 149
5.7.1 Air turbines 149
5.7.2 Hydraulic turbines 151
5.7.3 High-pressure oil hydraulic motors 152
5.8 Electricity generation 153
5.8.1 Direct drive wave energy 154
5.9 PTO systems 156
5.10 Pros and cons 157
5.11 Principle of OTEC operation 158
5.11.1 Open-cycle OTEC 159
5.11.2 Closed-cycle OTEC 160
5.12 OTEC and environment 163
5.13 Technical limitations and challenges 164
5.14 Site selection 164
5.15 Advantages and disadvantages 165
References 166
6 Biomass energy 169

6.1 What is biomass? 169
6.2 Types of biomass 169
6.3 Biomass energy 171
6.4 Biomass properties 172
6.5 Biomass energy conversion technologies 173
6.6 Pros and cons of biomass energy 175
6.6.1 Burning wood 175
6.6.2 Burning municipal solid waste or wood waste 176
6.6.3 Disposing ash from waste-to-energy plants 176
6.6.4 Collecting landfill gas or biogas 176
6.6.5 Liquid biofuels: ethanol and biodiesel 177
References 177
Part II: Energy storage systems 179

7 Electrical energy storage 181
7.2 EES and renewable energy 182
7.3 EES concept 183
xii Power Grids with Renewable Energy
7.3.1 Elements of EES 183

7.4 The need for EES 183
7.5 Critical challenges 184
7.6 Functions of EES 185
7.7 Classification of EES technologies 189
7.8 Characteristics of EES technologies 190
7.9 Fields of energy storage applications 192
References 193
8 Mechanical energy storage systems 195

8.1 PHES system 195
8.1.1 Physical principles 195
8.1.2 Modelling of PHES dynamics 197
8.1.3 Applications 201
8.1.4 Types of PHES 201
8.1.5 Advantages and disadvantages of PHES 202
8.2 CAES system 203
8.2.1 Principles of CAES operation 204
8.2.2 CAES system modelling 206
8.2.3 Advantages and disadvantages of CAES 207
8.3 FW energy storage 208
8.3.1 FES architecture and operation 209
8.3.2 Mathematical relations 210
8.3.3 FES characteristics 212
References 213
9 Chemical energy storage systems: fuel cells and power-to-gas 217

9.2 Principles of FC operation 218
9.3 Reversible open-circuit voltage of FCs 221
9.4 FC’s efficiency 224
9.5 Factors affecting VOC 226
9.5.1 Activation losses 229
9.5.2 Ohmic losses 230
9.5.3 Concentration losses 230
9.6 FC equivalent circuit 231
9.7 PEMFC characteristics 233
9.7.1 Calculation of resistances 233
9.7.2 PEMFC computer modelling 234
9.8 Types of FCs 235
9.9 Power-to-gas 236
9.9.1 Principle of operation 236
References 241
Contents xiii
10 Electrochemical energy storage systems 243

10.1 Conventional secondary BES technologies 243
10.1.1 PbA batteries 243
10.1.2 Nickel-based batteries 244
10.1.3 Li-ion batteries 245
10.1.4 Me-air batteries 246
10.1.5 NaS batteries 247
10.1.6 NaNiCl batteries 248
10.2 Flow batteries 250
10.2.1 Redox flow batteries 251
10.2.2 Hybrid flow batteries 252
10.3 Battery modelling 252
10.4 Comparison of BES technologies 255
10.4.1 Technical comparison 256
10.4.2 Economical comparison 256
References 258
11 Electrostatic, magnetic and thermal energy storage 261

11.1 Electrostatic energy storage systems 261
11.1.1 Working principle of SCES 261
11.1.2 Electrical performance 262
11.1.3 Lifetime aspect 264
11.1.5 Advantages and disadvantages 265
11.2 Superconducting magnetic energy storage 266
11.3 Thermal energy storage 267
11.3.1 Classification and characteristics of TES 268
11.3.2 Sensible heat storage 269
11.3.3 Latent heat storage 269
11.3.4 TCH energy storage 270
11.3.5 Storage systems 271
11.4 TES performance 274
References 275
12 Energy storage system application: electric vehicles 277

12.2 PEV technologies 278
12.2.1 Operation modes 279
12.2.2 PEV design features 280
12.3 Batteries 282
12.4 Electric motors 282
12.4.1 Brushed DC motors 282
12.4.2 Dynamic and regenerative braking in DC motors 289
xiv Power Grids with Renewable Energy
12.4.3 Brushless DC motors 292

12.4.4 SR motors 296
12.5 Electric vehicles performance 298
12.6 Battery charging infrastructure 301
12.6.1 Charging stations 303
12.6.2 Chargers 303
12.6.3 Charging times 306
References 308
Part III: Renewables integration 311

13 Power conditioning systems 313
13.1 General 313
13.2 AC–DC conversion 313
13.2.1 Uncontrolled rectifiers 314
13.2.2 Phase-controlled rectifiers 315
13.2.3 Diode rectifiers cascaded with a boost converter 317
13.2.4 PWM-controlled rectifiers 319
13.3 DC–DC conversion 320
13.3.1 Buck converter 321
13.3.2 Boost converter 322
13.3.3 Buck-boost converters 324
13.4 DC–AC conversion 324
13.4.1 Sinusoidal PWM 325
13.4.2 Operation of single-phase inverters 326
13.4.3 Operation of three-phase inverters (two-level) 328
13.5 AC–AC conversion 329
13.5.1 ON–OFF control 329
13.5.2 Phase control 330
13.6 Output filters 331
13.6.1 LC filters 331
13.6.2 LCL filters 333
13.7 Case studies 334
13.7.1 Case 1: grid-connected PV 336
13.7.2 Case 2: grid-connected DFIG WT 339
References 342
14 Integration of distributed energy resources 345

14.2 Powering DG systems 347
14.2.1 Non-renewable resources 347
14.2.2 Renewable resources 347
14.3 Benefits of DGs 348
Contents xv
14.4 Operation requirements for DERs integration into power systems 352
14.4.1 Voltage and frequency function 352
14.4.2 Power system restoration 355
14.4.3 Confirming reliability 357
14.4.4 Virtual inertia 359
14.5 Microgrids 362
14.5.1 MG concept 363
14.5.2 Operation functions 363
14.5.3 Challenges facing MG implementation 364
14.5.4 Role of ESSs 365
14.5.5 Protection strategy 366
14.5.6 Control of MGs 367
14.5.7 Islanding detection 369
14.6 Concluding remarks 369
References 369
15 Economic dispatch of hybrid thermal and wind plants 373

15.1 Description of ED problem 373
15.2 Modelling of system comprising thermal units and wind energy
sources 374
15.2.1 Wind power modelling using proper pdf 374
15.3 Formulation of optimization problem 377
15.4 Single-objective ED model 379
15.4.1 Conventional ED with thermal units 379
15.4.2 Deterministic ED incorporating wind power 380
15.4.3 SED incorporating wind power 380
15.5 Economic/emissions dispatch problem 382
15.5.1 Multi-objective stochastic dispatch 383
15.6 Role of ESS 385
References 396
16 Generation scheduling for power grids with renewables 401

16.2 Thermal generating units 402
16.2.1 Characteristic of thermal units 402
16.3 Formulation of the thermal scheduling problem 403
16.3.1 Unit commitment 404
16.3.2 Problem objective function 404
16.3.3 Fuel costs 404
16.3.4 Transition costs 405
16.3.5 System constraints 405
16.3.6 Unit constraints 407
16.4 Economic dispatch 408
xvi Power Grids with Renewable Energy
16.5 Review of thermal scheduling techniques 408

16.5.1 Priority list 409
16.5.2 Dynamic programming 409
16.5.3 Lagrangian relaxation 410
16.5.4 Branch and bound 410
16.5.5 Linear programming 410
16.5.6 Artificial intelligence 411
16.6 Problem representation and solution strategy 413
16.6.1 GA implementation 413
16.6.2 Decomposition methods 415
16.6.3 PL implementation 416
16.6.4 ANN implementation 416
16.6.5 ED implementation 417
16.7 SCGS for smart grid incorporating wind and solar thermal energy 419
References 427
17 Load flow analysis for power systems with renewables 431

17.2 Steady-state load flow calculations 432
17.2.1 Load flow algorithm 433
17.2.2 Role of the ESS 443
17.3 Load flow calculations for transient stability 449
References 454
18 Integration of electric vehicles with renewables into power grids 457

18.1 Electrification of transportation 457
18.2 EVs integration with RESs 459
18.3 EVs for ancillary services in power systems 460
18.3.1 V2G concept 460
18.3.2 EVs for power grid frequency regulation 461
18.4 V2G challenges 480
References 480
19 Microgrids: modelling and control 485

19.2 MG architecture 486
19.3 Modelling and simulation of MG components 488
19.3.1 PV system 488
19.3.2 Wind energy system 491
19.3.3 Proton-exchange membrane fuel cell 494
19.3.4 Battery energy storage system 496
19.3.5 Load profiles 496
19.4 Application of D-STATCOM 500
19.5 Application of SFC 505
References 508
Contents xvii
Part IV: Associated systems 511

20 Communications in power systems 513
20.1 Perspective of future power grid 513
20.1.1 Classification of communication networks 515
20.2 Wired communications 517
20.2.1 Power line communications 517
20.2.2 Optical communications 522
20.2.3 Digital subscriber lines 522
20.3 Wireless communications 523
20.3.1 Microwave radio systems 523
20.3.2 Cellular communication networks 523
20.3.3 Satellite communication systems 524
20.4 Internet access 524
20.5 Concluding remarks 525
References 526
21 Cybersecurity in power systems 529

21.2 Communication network protocols in power grids 529
21.3 Cybersecurity objectives and requirements 531
21.4 Network attacks in power grids 532
21.4.1 DoS attacks 533
21.4.2 Attacks targeting integrity and confidentiality 534
21.5 Grid privacy 535
21.6 Security of grid components 536
21.6.1 Process control systems 536
21.6.2 Power system state estimation 537
21.6.3 Smart meters 537
21.6.4 Communication protocol 538
21.6.5 False data injection attacks 538
21.7 Cloud computing 538
21.7.1 Advantages, disadvantages and threats of
cloud computing 540
References 541
Appendix A: Probability density and Weibull distribution functions 543
Appendix B: Test system 549
Index 553
About the authors
Abdelhay A. Sallam is Professor Emeritus in Port Said University, Egypt. He has

published over 100 papers in international journals and conferences and 2 books,
has supervised thesis of over 60 students and has had 17 years of experience in
industry.
Om P. Malik is Professor Emeritus in the University of Calgary, Canada. His pub-

lication record comprises over 375 papers in international journals such as IEEE
Transactions and IET (formerly IEE) Proceedings, and he has presented over 360
papers in international conferences. He is the author of two books. Under his
supervision, 52 PhD and 58 Masters students have successfully completed their
degrees.
Preface
Traditionally, growth and development of the electric power system has been based
on the idea of establishing large size centralized generating power plants to provide
electricity and meet the continuous growing demand in an optimal manner, eco-
nomical and technical. The common types of these plants are that fuelled by fossil
fuels, hydro-turbine power plants, or nuclear power plants. A power grid is dedi-
cated to serve both large and small consumers with electrical energy. In developing
the power grid, attention of power system planners and operators is primarily aimed
at providing the consumers with electrical energy as economically as possible and
with a high degree of reliability and quality of supply.
With the rising environmental concerns about global warming trends, there has
been a big change in thinking about the source of energy for the generation of
electricity. This has resulted in the generation of electricity from renewable sour-
ces, having attracted a tremendous amount of attention worldwide. Renewable
energy, derived from imperishable resources of nature being inexhaustible, has the
major advantage of being replenished or replaced naturally. Examples of potential
sources of renewable energy include biomass, solar, wind, hydro, geothermal and
ocean waves, among others. Geothermal and hydro sources have been exploited for
a long time.
Over the past few decades, there has been a tremendous amount of progress in
the development of electricity generation from renewable energy sources such as
wind and solar. Solar and wind-based plants ranging in capacity from hundreds of
megawatts to small distributed plants serving individual residences are in existence
now. Now, these may be considered as having reached the stage of mature tech-
nologies. At the same time, additional research is going on in parallel on the
development of electricity generation from other sources such as oceans and
biomass.
Considered for long-term use, renewable energy is sustainable and less
expensive as it does not require external feeding of fuels that cannot be replenished.
There is, however, the concern of the non-availability of renewable sources under
certain conditions and lack of control on the resource (fuel) input can significantly
affect the production of sustainable power. Despite this, the benefits of sustainable
energy outweigh its drawbacks.
Nature of the various renewable sources, such as variability and inability to
accurately predict and control, unlike the fossil fuel sources of energy used in
conventional generation, creates its own problems when integrating with the con-
ventional power systems at both the grid and the distribution levels. A tremendous
xxii Power Grids with Renewable Energy
amount of research and development is taking place in ensuring utilization and

integration of the electricity generated from renewable sources.
Lack of control on the basic input source of energy, unlike that of the con-
ventional fossil or even hydro-based generation, and intermittence of generated
electricity requires special consideration while integrating renewable energy sour-
ces in conventional power grids. Energy storage and proper interfacing play an
indispensable role in the optimal application and utilization of natural resources,
and efficient overall system operation of the conventional and renewable energy
sources integrated in power systems.
With the increasing generation of electricity from renewable sources at all
levels of the power system, operation of the power grid is becoming more complex
as significant power will flow in both directions, i.e. not only from the power
system to the consumer but also from the consumer to the power system. Problems
like excessive voltage variations and bottlenecks may occur more frequently, and
then, additional investments in the network reinforcement, in more sophisticated
protective relaying and control systems, are required. Thus, grid design modifica-
tion is needed to benefit from the potential of greater flexibility offered by dis-
tributed generation in the form of microgrids, flexible loads, storage systems,
computer monitoring and control technologies, and communication networks. Such
flexibility will require that the power grid become smarter.
The technology of renewable power generation is well covered in many
existing books. Material relating to both the mature, wind and solar, and other
emerging technologies for the generation of electricity from renewable sources is
limited in this book to that necessary. The primary thrust of this book is to cover the
requirements for proper interfacing and integration of the renewable generation
technologies in the existing power grids considering the special operational char-
acteristics of these variable generation sources with no control on the basic input
source of energy unlike that of the conventional fossil or even hydro-based
generation.
An outline of the contents of this book is given as follows:
● A general background to the topic of this book, new trends developing in
power grids, i.e. generation of electricity from renewable sources, is provided
in the introductory (Chapter 1).
● The state of development of various renewable energy sources being pursued
at present is presented in brief in Part I ‘Renewable Energy Sources’ consisting
of Chapters 2–6.
● Energy storage is one of the important aspects in the integration of renewable
generation for efficient operation of the conventional and renewable genera-
tion integrated power grid. The essential role played by energy storage sys-
tems, various types of energy storage systems, such as electrical, mechanical,
chemical, thermal, electromechanical, electrostatic and magnetic including
superconducting, and their applications, especially, the electric vehicles is
included in Chapters 7–12, Part II ‘Energy Storage Systems’.
Preface xxiii
● Despite the generation and use of energy from sustainable sources having
increased significantly recently, there is still a need for optimizing the appli-
cation of natural resources. The renewable energy sources need to be inte-
grated properly with the existing power network for efficient system operation.
The most significant part of the book, Part III ‘Renewables Integration’,
Chapters 13–19, covers interfacing, integration and operational aspects of
renewable energy and includes energy storage, power electronics interfacing
and other operational considerations.
● Communications and cybersecurity, essential for the safe and secure operation
of the integrated system, are covered in Chapters 20 and 21, in the concluding
Part IV ‘Associated Systems’.
● All chapters are supported by a limited number of relevant references so that
the interested reader can pursue more in-depth coverage of any specific topic.
The authors have drawn on their own research and many other tools such as
MATLAB/Simulink and toolboxes.
Abdelhay A. Sallam
Om P. Malik
Acknowledgements
In the present day, no individual is unto self. It is the cooperative effort of many,
and this book is no exception in that regard. The authors, therefore, wish to record
their appreciation of the help received from various sources in preparing this book.
A part of the material included is based on their research and the work of a number
of graduate students, though not individually named here, who have worked under
their supervision over the years and are making their own contribution to the pro-
fession. Their work and contribution are well appreciated.
The authors, however, feel it appropriate to express their gratitude to certain
individuals by name for their specific contributions and help in putting this book
together. These are
● Dr Azza A. El-Desouky,
● Dr Mai M. Eladany,
● Ahmed H. Elmetwaly,
● Mohamed A. Kapeel,
● Mohamed G. Anany, and
● Mahmoud M. Mashaal.
Chapter 1
Towards the new trend of power grids
Historic growth and development of the electric power system has been based on
the idea of economies of scale and building bigger and bigger electricity generating
plants primarily fuelled by fossil fuels with some hydro generating plants, where
such possibility exists, and a few nuclear generating plants. Global warming trends
and the environmental concerns have brought in a big change in thinking with the
result that over the past couple of decades a tremendous amount of attention has
been devoted to the generation of electricity from renewable sources such as wind,
solar, oceans and biomass.
Electricity generation from wind and solar as primary sources of energy has
advanced tremendously with the result of solar and wind-based plants in all sizes,
ranging in capacity from hundreds of megawatts to small distributed plants serving
individual residences, that are in operation today. Very significant cost reductions
have also resulted in parallel with these developments.
While the state of technology in the development of the wind-based and solar
plants may be considered to have reached the state of maturity, development in the
generation of electricity from other renewable sources has been progressing in
parallel. These are the energy in oceans, utilization of biomass and so on.
The nature of the various renewable sources such as its variability and our inability
to accurately predict and control, unlike the fossil fuel-based sources of energy used in
the conventional generation, creates its own problems when integrating them with the
conventional sources at both the grid and the distribution levels. A tremendous amount
of research and development is taking place in ensuring proper utilization and inte-
gration of the electricity generated from renewable sources. The aspects of both gen-
eration and integration are covered in the subsequent chapters in the book.
A power grid is dedicated to serve both large and small consumers with elec-
trical energy. In developing the power grid, the focus of power system planners and
operators is primarily aimed at providing electrical energy to the customers as
economically as possible and with a high degree of reliability and supply quality.
The term ‘grid’ denotes the entire electric system infrastructure, which is also
commonly known as ‘electric power system’. Thus, the two terms ‘grid’ and
‘power system’ are often used interchangeably [1].
A brief description of conventional power systems, recent developments and
future perspectives based on new enabling technologies, advanced controllers and
communication facilities is presented in this chapter.
2 Power grids with renewable energy
1.1 Conventional power systems

A complete power system is a chain of three consecutive sectors. These sectors are
denoted by their functions as generation, transmission and distribution, respec-
tively. The conventional power system infrastructure comprises centralized bulk
generating stations (e.g. fossil-fuel thermal, nuclear and hydro stations) supplying
electricity, through extra-high-voltage (EHV) and high-voltage (HV) transmission
networks (also called transmission and sub-transmission networks, respectively) to
local distribution systems. Local distribution systems serve the consumers’ loads
are classified into three major types, industrial, commercial and residential, as each
type has its own characteristics and requirements. In the conventional power sys-
tems, electricity flows in one direction starting at generating stations and
terminating at the individual load points in distribution network as shown in
Figure 1.1.
With the progress in the level of technology in power systems, many pro-
grammes using available new technology (e.g. flexible AC transmission system
(FACTS), fast controllers and communication systems) have been applied to
improve power system characteristics, especially, reliability, security and sta-
bility [2].
Power system reliability: In general, system reliability is specified by the
overall ability of the system to perform its function. In a power system, the basic
aspect of identifying its reliability is the power system adequacy. It relates to the
presence of enough facilities within the system to satisfy the consumer load
demand, i.e. it is necessary to generate sufficient energy and to use both trans-
mission and distribution networks with sufficient power transfer capacity to enable
the system to transport the energy to the load points. Therefore, adequacy is asso-
ciated with static conditions excluding contingencies and is evaluated by identify-
ing suitable indices such as the following:
● Expected frequency and duration of failure: It identifies the expected
frequency of encountering a deficiency and the expected duration of the
deficiencies.
● Loss of energy expectation: It identifies the expected energy that will not be
supplied by the generating station when the load demand exceeds the available
generating capacity.
● Outage time of consumer load points: It is the annual time of unavailability of
feeding load points with energy.
Power system security: The term ‘security’ refers to the degree of risk in the
ability of the electric power systems to withstand without serious consequences
to any of a preselected set of sudden disturbances such as short circuits or
unanticipated loss of system elements. From a control perspective, security is the
degree of risk in the ability of a power system to reach a steady-state operating
point without violating system operating constraints, mainly, thermal limits of
the transmission lines and bus voltage limits. These constraints are referred to as
Towards the new trend of power grids 3
Tr
an
Thermal
sm
iss
io
n
ne
Step-up substation
tw
Nuclear
or
MV/EHV
k
U
Hydro
ni
di
re
ct
io
na
lf
Large-size
lo
Step-down substation
w
generating stations
of
EHV/HV
el
ec
tri
ci
ty
rk
wo
net
ion
iss
sm
Pri ran
ma
b-t
ry
ork
fee
Su
etw
der
nn
tio
Industrial load
ibu
str
HV/MV
Di
Commercial
load
Secondary
feeder
Residential
load
Figure 1.1 A typical conventional power system
security constraints when referred to the post contingency scenarios and can be
written as
Ng
X
PGi ¼ PD þ Ploss and Pmin max
Gi PGi PGi (1.1)
i¼1

min max
Vj Vj Vj and Sjk Sjkmax 8 branch j k (1.2)
where
D
PGi ¼ generated real power at bus # i;
D
PD ¼ total system demand;
D
P D¼ total real power loss in the transmission network;
loss
V j ¼ voltage magnitude at bus # j;
D D
Sjk ¼ apparent power (MVA) flow in branch j k and Ng ¼ number of
generators.
Power system stability: It commonly pertains to rotor angle stability. Stability
phenomenon is characterized by generator rotor speed oscillations under a severe
perturbation and can be analysed by solving non-linear dynamic equations describing
the transient behaviour of the system under a set of credible contingencies.
Nowadays, electricity has become all pervasive in the world resulting in tre-
mendous growth in load demand. Consequently, the conventional power system
infrastructure has become
● large scale;
● multi-input/multi-output;
● distributed over large geographical areas requiring the energy needs to be
routed through long distances, resulting in increased transmission losses;
● complex with many interconnections;
● costly due to the investments needed to expand generation, transmission and
distribution systems while still maintaining a desired level of reliability,
security and stability.
In addition, a large part of generation is based on fossil fuel (coal, oil and
natural gas) as source. Inflation and astronomical increase in fuel prices has resulted
in rapid increase in consumer tariffs as well as adverse effects on the environment.
Their combined effects introduce considerable uncertainty in predicting the future.
To satisfy this continued growth in load demand necessitates further development of
the power system that is compatible with environmental considerations.
In a new energy economy, it will be beneficial to empower every consumer to
conserve energy and become an energy producer by generating electricity from
renewable energy sources (RESs). This will become a major issue in satisfying
system adequacy as it creates a reduction in forecast demand, improves system
security and stability as well as reduces greenhouse gas emissions.
Distributed generation (DG) including RESs (e.g. wind, solar, photovoltaic (PV),
tidal and micro-hydro) and non-RESs (e.g. fuel cells, diesel generators and micro-gas
turbines) can be interconnected together to form what is called a ‘microgrid’ (MG)
that can operate in isolation, islanded or, in turn, can be connected to a utility
distribution system. Therefore, power system planning and configuration must be
modified to recognize this new scenario [3,4].
1.2 Microgrids
MGs are small distribution systems encompassing a group of electricity consumers

(industrial, commercial and residential) connected to several distributed generators
and storage units that can be interfaced by power electronics.
MG technologies help electric system evolve into one that is more efficient, is
less polluting, has lower chances of transmission congestion, has reduced losses
and is more flexible to provide energy consumers’ wants and needs. MGs can help
to make better use of energy generated, stored and used at a local level, e.g. local
generation and islanding operation. In addition, in future each client can become an
autonomous entity with the capability of not only demanding but also injecting
power into the grid when the generated power exceeds the local needs [5,6].
Lately, utility companies have been motivated to encourage connection of
local RESs at the distribution level. This can modernize the conventional power
systems into multiple, interconnected distribution systems, i.e. MGs. However, it is
not a simple proposition to move from the centralized system that has developed
over the past more than 135 years to one that is distributed. Implementing DG
comes with challenges as the present electric power infrastructure is not designed to
meet problems pertaining to operational capability and quality of supply of elec-
tricity generation using RESs such as wind and solar, which are intermittent and
unpredictable.
An example of MG architecture connected to an utility distribution network is
illustrated in Figure 1.2. It typically comprises non-renewable (e.g. AC diesel
generators) and renewable generation units (e.g. wind, solar, micro-hydro, ocean
wave and tidal) in addition to energy storage systems (ESSs). These energy sources
(variable frequency AC, DC and energy storage) are connected to DC bus through
AC/DC converter, DC/DC converter and interface power electronic circuits. The
local DC loads are fed from the DC bus. The excess of DC energy can be converted
into AC energy by DC/AC inverter to be connected to AC bus at the point of
common coupling (PCC). The AC bus is fed from both MG and utility grid to
supply the AC loads [7]. More details are given in Chapter 14, Section 14.5.
1.3 Energy storage systems

RESs (e.g. wind speed and solar radiation) can fluctuate quickly, and at random, it
is difficult to ensure a constant stable supply. So, ESSs can be used to compensate
the fluctuation from RESs and maintain power quality. Installation of energy sto-
rage devices can help alleviate limits on the size of the RESs.
Services provided by storage depend on the type of storage technology and are
also based on their capacity. Two types of energy storage technologies can be used:
(i) bulk storage (such as compressed air energy storage, pumped hydro and large
batteries) is able to provide a wide range of grid services, e.g. day-night arbitrage,
peak shaving regulation and spinning and non-spinning reserves. This type of sto-
rage can be connected to utility transmission network. (ii) Short-term storage
(including batteries, flywheel, superconducting magnetic energy storage (SMES)
[8], plug-in electric vehicles (PEVs), etc.) is generally used for providing shorter-
term regulation services to accommodate load and variation of renewable genera-
tion on a limited time scale. Therefore, the development of DG can be supported by
deploying small-scale, low-voltage (LV) or medium-voltage (MV) grid-connected
DC bus
AC
energy AC/DC
sources converter
DC loads
AC bus
DC DC/DC
energy smoothing AC loads
sources circuit
DC/AC
inverter
Energy Interface PCC
storage power electronic
systems circuit
Microgrid Utility
distribution
network
Figure 1.2 Schematic diagram of an MG connected to utility distribution network
electricity storage systems at different locations in the MGs (decentralized storage).

Such storage systems could affect the management of distribution grid in different
ways: (a) energy management by delivering capacity to satisfy energy arbitrage
(generation and consumption), (b) optimizing the operation of distribution grid and
(c) keeping the quality of service and security of supply. Under certain circum-
stances, non-storage-type solutions, e.g. micro-gas turbine and micro-hydro tur-
bine, may be employed to level load.
Accordingly, a typical overall power grid perspective can be visualized as in
Figure 1.3. It is to be noted that the MG is connected to the utility distribution
system at the PCC. The MG encompasses RESs (wind, solar and PV array), non-
RESs (fuel cells and PEVs) and ESSs such as SMES, flywheel and battery storage.
PEVs demand power from the MG (as a load) in the charging state and inject power
into the MG (as a source) in discharging state. For the transmission system, bulk
power delivered by the compressed-air storage system can be injected to it as and
when needed. More details are given in the forthcoming chapters (Part II).
1.4 Smarter grid
It is to be noted that the traditional concepts and practices of power systems have
changed as different energy sources and storage systems are added. More compli-
cated controllers, information systems and new technologies encompassing power
electronics and two-way communication systems are of paramount importance for
Centralized
generation
Step-up substation
MV/HV
rk
wo
net
ion
iss
nsm
Tra
Compressed air
storage
HV/MV
Utility distribution network

Fuel cell Solar-thermal
Load
PCC
Plug-in electric
vehicle
FLYWHEEL
Flywheel MOTOR/GENERATOR
Microgrid
storage
SMES
PV array
Battery storage
Wind farm
Figure 1.3 Power grid supported by MG and compressed air storage
system components’ integration, interface and better utilization of the existing

power system. In consequence, efforts are continuously being made to restructure
the electric power industry into a more competitive industry focusing on the
wholesale electricity market. In this market, generating companies produce elec-
tricity and sell it to energy service providers, which, in turn, resell the electricity to
residential, industry and commercial consumers in a retail market.
Operation of the power grid shown in Figure 1.3 will become more complex as
significant power will flow in both directions, i.e. not only from the power system
to the consumer but also from the consumer to the power system. Problems such as
excessive voltage variations and bottlenecks may occur more frequently, and then,
additional investments in the network reinforcement in more sophisticated protec-
tive relaying and control systems are required. Thus, modification in the grid design
is needed essentially to benefit from the potential of greater flexibility offered by
DG, flexible loads, storage systems, computer monitoring and control technologies,
and communication networks [9].
Such flexibility would help reinforce the conventional network, make it
smarter and provide enhancement to ensure the following:
● High level of security, quality, reliability and availability of electric power.
● Optimization of the use of the existing network, thereby minimizing distribu-
tion grid extensions.
● Improved economic productivity and quality of life.
● Minimization of environmental impact while simultaneously maximizing
safety and sustainability.
● To deal with this paradigm shift, active involvement of both consumers and
distribution system operators (DSOs) is a must [10].
1.4.1 Consumers’ role

Design of the distribution systems is based on the load analysis performed at the
time. As the load levels increase and load patterns change, some sort of upgrading
and more investment are required to make the design accommodate the new con-
ditions and obviate system overloads. On the other hand, proper management of
distribution system overloads can help reduce the required investment. It may be
like congestion management in the transmission system where it needs distribution
capacity reservation, use of retail pricing, curtailable schedules and distribution
loading relief to manage the loading of the distribution facilities. Traditional load
management tools have already been used by direct load control, peak shaving,
peak shifting and various load management programmes. Now, demand response
(DR) is taking on new realities with the push for energy conservation and demand-
side management as a key strategy of distributed resource management.
The role of the consumers in the electricity market will increase, paving the way
for DR in addition to the traditional tools of load management. Consumers will be
able to manage and adjust their electricity consumption in response to real-time
information and changing price signals. Information can be provided by dynamic
feedback programs designed by suppliers to let the consumers actively manage their
consumption. This alone is necessary but not enough to change electricity consump-
tion. It only becomes enough when consumers become more aware of the value of
shifting their electricity consumption. Meanwhile, grid tariffs should reflect the actual
costs in a manner that motivates electric markets to introduce attractive services.
Consumers’ awareness of their energy consumption can be increased by
installing smart meters, which will be the responsibility of DSOs, as well as
intelligent appliances and smart chargers. Smart meters as a kind of intelligent

metering are a part of the physical grid infrastructure, fall under the grid operator’s
domain and are managed by DSOs. They offer new advantages to both the utility and
the consumer by efficient use of energy, operational efficiencies, improved daily
service, performing majority of operations remotely, efficient use of resources and the
possibility for consumers to monitor and reduce their consumption. Therefore, the grid
will be ready to move from the traditional load-following operating strategy to a load-
shaping strategy in which demand-side resources are managed to meet the available
generation and the grid’s power delivery capabilities at any time [9].
Moreover, smart metering greatly contributes to optimizing network manage-
ment as it offers the following:
● remote connect/disconnect function to manage, for instance, peak shaving
programmes;
● two-way wire or wireless communication;
● advanced measurement capabilities;
● detailed monitoring of power quality;
● fault identification, fraud tempering and theft detection function; and
● automatic and self-contained interval metering capability.
1.4.2 DSOs’ role

DSOs need to play a more active and prominent role in applying the right tools and
to be following their basic tasks of maintaining supply reliability and service
quality. DSOs are responsible for the following:
● preventing bottlenecks in the most cost-effective way through coordination
with transmission system operators based on long-term network planning;
● integrating DG resources into the network in such a way that they maintain
voltage quality level;
● ability to obtain flexibility from DG resources and consumers to solve grid
constraints;
● application of desired technical tools to become real system operators – this
necessitates supervision and control of power flows using advanced protection
systems, monitoring and control strategies;
● dealing with new technologies in the grid to improve service quality and
reduce costs;
● monitoring the integration of the consumers in the most effective and eco-
nomical way to maintain adequacy and service level while satisfying overall
energy efficiency of the grid; and
● adapting information technology (IT) systems to probable change in the grid
structure.
1.4.3 Smart grid definition

New technologies are in progress on various fronts. At present, a lot of research is
devoted to incorporating renewable and non-renewable distributed generation
(MGs) as well as ESSs into the power grid. So, the power grid is a system that
encompasses multi-energy sources, transmission networks and distribution net-
works incorporating several MGs. This composition needs different components of
the power grid to be integrated together to constitute the infrastructure that enables
the power grid to function properly. New technologies such as digital commu-
nication systems, sensor technology, FACTS devices, supervisory control and data
acquisition (SCADA), phasor measuring units (PMUs), DR programs, IT and
intelligent metering technology can support this integration with a goal of satisfy-
ing the following [11]:
● procuring accurate demand forecast and renewable energy availability;
● real-time monitoring, operation and control of power system components
(remotely and local);
● real-time management of power demand and economic dispatch;
● desired power quality and security level;
● energy sustainability;
● achieving cleanliness and environmental friendliness as much as possible; and
● self-healing that needs the grid to rapidly detect, analyse, respond and restore.
With the implication of advanced technologies satisfying all or some of these
aspects, the power grid is being referred to as ‘smart grid’ (SG). Improvement of
the conventional power grid, to behave as an SG, is based on selected aspects that
each utility concerns itself with. In addition, as the new enabling technologies
(information, communication and automation) are exploited intensely in the power
grids, the power grid gets smarter. There is no clear or standard definition of an SG.
Everyone seems to have own vision and a framework for an SG.
According to an Alberta, Canada, government report, SG is a broad concept
that describes the integration of hardware, computer monitoring and control tech-
nologies, and modern communications networks into an electricity grid [1].
European Commission Task Force on Smart Grid defined SG as ‘an electricity
network that can intelligently integrate the action of all users connected to it –
generators, consumers and those that do both – in order to ensure economically
efficient and sustainable power system with low losses, high level of power quality,
security of supply and safety’ [12].
In the US Energy Independence and Security Act 2007, ‘smart grid’ refers to a
modernization of the electricity delivery system so that it monitors, protects and
automatically optimizes the operation of its interconnected elements – from the
central and distributed generator through appliances and other household devices.
In the absence of a formal definition of ‘smart grid’, a generic description, as
good as any, is as follows: ‘a smart grid is a digitally enabled electric grid that
gathers, distributes and acts on information about the behavior of all components
in order to improve the efficiency, reliability and sustainability of electricity
devices’ [1].
An example of SG perspective of the future is depicted in Figure 1.4, where the
communication, information and computer technologies provide predictive infor-
mation, recommendations, and real-time monitoring and control to utilities. Each of
SCADA for generation

automation
PMU
on
SCADA for MV/HV
ati
PMU
nic
substation
PMU
SCADA for transmission
mu
automation
com
and storage automation Satellite
ne
r li
we
rk
wo
Po
net
ion
iss
nsm
Tra
Main coordinator and

Compressed air control centre
SCADA for HV/MV
storage
substation
automation
PMU
Data
aggregator/
SCADA concentrator Fuel cell
Control centre Solar-thermal
Utility distribution PMU
network Smart
PCC meter
Plug-in electric
vehicle
Data aggregator/
FLYWHEEL concentrator
SCADA
Flywheel MOTOR/
GENERATOR
Control centre Load
storage
SMES
Microgrid
Smart
meter
PV array
Smart Power
meter converter
FACTS
Battery storage devices
Wind farm
Figure 1.4 SG perspective of the future
MG, distribution system, transmission system and central generation is locally

controlled by SCADA through measurements taken from data aggregator/
concentrator. Then, a coordinator and control centre unit receives data from the
local control centres through satellite for decision-making necessary to integrate
the function of overall system components and satisfy the aspects of SG. Some of
the major elements used are phasor measurement systems, PMUs and smart meters.
PMUs are used to measure in a complex form (magnitude and direction) the
electrical parameters at widely dispersed locations in the power system. The
measurements are synchronized by using the global positioning system (GPS) to

help the operators to specify the state of the power system and respond to system
conditions in a rapid and dynamic manner. Smart meters with the aid of wireless
networks are used for information exchange between users.
In addition, power line communication, by using wireless modems, is used to
communicate between the substations at the terminals of transmission lines.
Moreover, sensor technology is used for monitoring and measurement in order to
detect components’ failure and system faults to be isolated before they cause cas-
cading effects and lead to catastrophic breakdowns. However, from what has been
discussed and as reported in [13], it can be concluded that a smarter grid comprises
three systems: smart infrastructure system, smart protection system and smart
management system.
1.5 Power grid examination

Lack of control on the basic input source of energy, unlike that of the conventional
fossil or even hydro-based generation, and intermittence of generated electricity
requires special consideration while integrating RESs in conventional power grids.
System frequency and voltage control are important issues and must be studied.
This study needs the system to be modelled in steady state, particularly, to ensure
the rated voltage at each bus in the power grid. The integration of DG (renewable
and non-RESs) as a part of an MG that is connected to the local utility power grid in
addition to its asynchronous operation must be examined. Consequently, the load
flow analysis is required to compute the power line flows and bus voltages, and
hence, the bus impedance and admittance matrices should be determined. Also,
economic operation based on economic load dispatch of power grid incorporating
both DG and ESSs as a part of an MG should be examined.
On the other hand, power grid security during abnormal operating conditions
must be analysed as this grid accommodates different energy sources, ESSs,
complicated networks and several additional assets. Thus, many security risks
pertaining to this complexity as well as communication requirements, system
automation, new technologies and data collection may be introduced.
All these issues in addition to more detailed description of different energy
sources and energy storage systems are presented in the forthcoming chapters.
References
[1] Malik O. P. ‘Evolution of power systems into smarter networks’. Journal of

Control, Automation and Power Systems. 2013;24(1):139–47.
[2] Arefifar S. A. and Mohamed Y. A. ‘DG mix, reactive sources and energy
storage units for optimizing microgrid reliability and supply security’. IEEE
Trans. on Smart Grid. 2014;5(4):1835–44.
[3] Smith C. ‘Wind and solar energy – Variable generation across the land’.
IEEE Power & Energy Magazine. 2015;13(6):16–20.
[4] Krishnan V., Ho J., Hobbs B. H., et al. ‘Co-optimization of transmission and
other supply resources – concept, review and modeling approaches’. Energy
Systems. 2015;7(2):1–36.
[5] Basu A. K., Chowdhury S. P., Chowdhury S., Ray D. and Crossley P. A.
‘Reliability study of a micro-grid power system’. Universities Power
Engineering Conference 2008, UPEC, 43rd International. Padova,
Sep 2008, pp. 1–4.
[6] Ahshan R., Iqbal M. T., Mann G. K. I. and Quaicoe J. E. ‘Micro-grid system
based on renewable power generation units’. Electrical and Computer
Engineering (CCECE), 2010 23rd Canadian Conference on. Calgary, AB,
Canada, May 2010, pp. 1–4.
[7] Nehrir M. H., Wang C., Strunz K., et al. ‘A review of hybrid renewable/
alternative energy systems for electric power generation: Configurations,
control, and applications’. IEEE Trans. on Sustainable Energy. 2011;
2(4):392–401.
[8] Nakayama T., Yagai T., Tsuda M. and Hamajimam T. ‘Micro power grid
with SMES and superconducting cable modules cooled by liquid hydrogen’.
IEEE Trans. on Applied Superconductivity. 2009;19(3):2062–5.
[9] Albuyeh I. F. ‘Grid of the future’. IEEE Power & Energy Magazine. 2009;
7(2):52–62.
[10] Mallet P., Granstrom P. O., Hallberg P., Lorenz G. and Mandatova P.
‘European perspectives on the future of electric distribution’. IEEE Power &
Energy Magazine. 2014;12(4):51–64.
[11] Petinrin J. O. and Shaaban M. ‘Smart power grid: technologies and appli-
cations’. 2012 IEEE Int. Conf. on Power and Energy (PECon). Kota
Kinabalu Sabah, Malaysia, 2012, pp. 892–7.
[12] European Commission. ‘European smart grid technology platform’.
Luxemburg, Dec 2006.
[13] Fang X., Misra S., Xue G. and Yang D. ‘Smart grid – The new and improved
power grid: a survey’. IEEE Communications Surveys & Tutorials. 2012;
14(4):944–80.
Part I
Renewable energy sources
Chapter 2
Wind energy
One of the most mature technologies for renewable resources is the wind energy.
The wind is a motion of air masses caused by the different thermal conditions of
these masses. The energy produced by this motion can be converted into another
form of energy, such as electric energy. This approach, known as wind power
generation, has grown rapidly during the past few decades in many countries
around the world.
Major catalysts for this rapid growth have arisen in the recent past: continuous
technological advances in power electronics, controls and physical attributes (e.g.
tower heights and blades) as well as cost reduction and significant advancements in
understanding the access of wind power generation to the grid. Consequently, more
attention is paid to high penetration of wind power sources in power systems as
these sources are not only a means to reduce CO2 gas emissions but also an eco-
nomic alternative.
This chapter is focused on the description of wind system components and
their functions being combined to convert the wind power to electrical output
power. The basic relations necessary to design the wind turbine (WT) and evaluate
its performance are introduced. The recommended controllers are also explained as
the wind speed varies with time and yields a fluctuating character to the system.
Different terms may be used when talking about the wind system such as wind
generator, wind turbine generator, wind-driven generator and wind energy con-
version system. In this chapter and in most literature, the term ‘wind turbine’ is
used. It is to be noted that the power is the energy per unit time.
2.1 Wind turbines
The main components of a WT are blades, rotor shaft, gearbox and generator
(Figure 2.1). A simplified block diagram representing the WT system is shown in
Figure 2.2. It is to be noted that the components gearbox, yaw drive, coupling, main
shaft, etc. are represented as gear system model, whereas the generator model
includes the interface system and a transformer to step up the generated voltage to
the desired voltage. In addition, other parts, e.g. control panel, are all placed in a
compartment known as ‘nacelle’. It is to be noted that the nacelle may also contain
a transformer. Inflow of wind activates the blades and rotor which in turn spins the
Blades
Main Gearbox
Hub shaft Brake Generator
Wind
inflow High-speed
shaft
es
ad
Bl
Nacelle
Tower
Figure 2.1 WT outline configuration
Wind
inflow
Gear system Generator
model model
WT Utility grid
model Substation
Controller
Figure 2.2 A simplified block diagram of WT system
main shaft and generator resulting in electrical output power. WTs can be installed
close to the ground (onshore) or close to sea level (offshore) as shown in Figure 2.3.
Onshore WT installation is much cheaper than the offshore installation. The
rotor shaft can rotate about either a horizontal axis or a vertical axis. Based on the
axis of rotation, WT is classified into two types: horizontal axis wind turbine
(HAWT) and vertical axis wind turbine (VAWT).
2.1.1 Horizontal axis wind turbines

HAWT is the most common design of WTs. The axis of rotation is parallel to the
ground. The nacelle (with main rotor shaft and generator) is placed on the top of the
tower (Figure 2.4(a) and (b)). WT must be pointed into the wind so that the blades
facing into the wind provide full swept area. Small turbines are pointed by a simple
vane, whereas large turbines generally use a wind sensor coupled with a servo
Wind energy 19
Figure 2.3 Onshore HAWT (on the left side) and offshore HAWT (on the
right side)
Nacelle with
gearbox and Swept area
generator of blades
Rotor
diameter
Tower
Hub Rotor blade
height
Foundation
(a) (b)
Figure 2.4 HAWT: a schematic diagram. (a) Side view and (b) front view
Wind Wind
direction direction
Upwind Downwind
Figure 2.5 Upwind and downwind machines
motor. Most have a gearbox which turns the slow rotation of the blades into a faster
rotation more suitable to drive an electrical generator. Depending on the rotor
orientation, two types should be discussed: upwind machines and downwind
machines (Figure 2.5).
2.1.1.1 Upwind machines

Upwind machines have the rotor facing the wind. Upwind design has a basic
advantage of avoiding the wind shade behind the tower. On the other hand, the
wind starts bending away from the tower before it reaches the tower itself; that is,
some wind shade in front of the tower exists. Therefore, the power from the WT
drops slightly each time the rotor passes the tower. Due to this, the rotor needs to be
made stiff and placed a considerable distance in front of the tower to prevent the
blades from being pushed into the tower by high winds. Upwind machines need a
yaw mechanism to keep the rotor facing the wind.
2.1.1.2 Downwind machines

Downwind machines do not need an additional yaw mechanism for keeping the
blades in line with the wind. The wind itself controls the yaw, and so, it naturally
orients itself correctly with respect to wind direction. In addition, the rotor may be
Wind energy 21
made more flexible; that is, the blades can be allowed to bend at high wind speeds,
thus taking part of the load off the tower. Therefore, the basic advantage of the
downwind machine is that it may be built somewhat lighter than an upwind machine.
Downwind configuration has another advantage; it is easier to take advantage of
centrifugal forces to reduce the blade root flap bending moments. This is because the
blades are normally coned downwind, so centrifugal moments tend to counteract
moments due to thrust. On the contrary, the basic disadvantage of downwind machine
is the fluctuation in the wind power due to shadowing effects of the tower. This may
give more fatigue loads on the turbine than with an upwind design.
2.1.2 Vertical axis wind turbines

The main rotor shaft of a VAWT is arranged vertically as well as the generator
and gearbox can be placed down on the ground. The blades take the form as
those results from holding a rope at both ends and spinning it around a vertical
axis (Figure 2.6(a) and (b)). Principal advantages of this arrangement are the
following: (i) there is no need for any kind of yaw control to keep blades facing
into the wind to be effective; (ii) the tower itself need not be structurally as
strong as that of a HAWT; and (iii) WT components can easily be serviced and
maintained. On the other hand, several disadvantages of VAWT configuration
also exist such as the following: (i) the blades are relatively close to the ground,
where wind speeds are lower, so less wind energy is available for a given size of
turbine; (ii) wind near the surface of the ground is more turbulent, so the shear
stresses are more severe and the rotor experiences higher loads; (iii) because of
Rotor
diameter
Rotor
height
Fixed pitch
rotor blade
Gearbox Generator
(a) (b)
Figure 2.6 HAWT: (a) schematic diagram and (b) physical picture
turbulent flow, some problems can be introduced associated with vibrations such
as noise and bearing wear, so more frequent maintenance is required and
otherwise the service life may be shortened; and (iv) the blades on a VAWT
cannot be made to spill the high wind to protect the generator as easily as pitch-
controlled blades on a HAWT.
It is seen that the WT can be designed as a HAWT or a VAWT. For a HAWT
which is used in common, it will be either upwind or downwind. In all designs,
the rotor with blades is the WT component that is facing into the wind to convert
the wind power into rotational motion. So, the design decision relating to the
number of blades should be taken. Basically, multi-bladed design presents a
large swept area, and accordingly, it enables both low speed and high torque
operation. For instance, in case of a rotor with one blade the turbulence nega-
tively affects the blade efficiency but with fewer blades the turbine can spin
faster before the turbulence becomes excessive and consequently the generator
can be physically smaller in size. In the USA, WTs mostly have two blades,
whereas most European turbines have three blades. Three blades WT presents
smoother operation and lower noise, but on the other hand, the weight and cost
are considerably increased.
2.2 Basic relations

2.2.1 Power from wind
With a mass of air, m, moving at speed, v, its kinetic energy, Ke, is given by
Ke ¼ ½mv2 (2.1)
Thus, the power, P, produced by this mass through an area, A, can be con-
sidered as the energy per unit time, that is,
m 2
P¼½ _ 2
v ¼ ½mv (2.2)
T
where T is the time.
The flow rate of the mass, m, through the area, A, is defined as the product of
air density, r, speed, v, and cross-sectional area, A, that is,
dm D
¼m_ ¼ rAv (2.3)
dt
Incorporating (2.2) and (2.3) and considering the wind as a packet of air with
mass, m, the wind power in watts, Pw, can be given by
Pw ¼ ½rAv3 (2.4)
3
where r is the air density (kg/m ), v is the wind speed normal to A (m/s) and A is the
cross-sectional area (m2) through which the wind passes. It is to be noted that
r ¼ 1.225 kg/m3 at 15 C and 1 atm.
Wind energy 23
Hence, the wind power per unit area which is called ‘power intensity’ or
‘specific power’ Pspec, can be obtained in W/m2 as
Pw
Pspec ¼ ¼ ½rv3 (2.5)
A
which can be plotted as shown in Figure 2.7.
It is found from (2.4) that the wind speed is proportional to both the cube of
wind speed and the swept area of turbine rotor. On the other hand, wind energy, Ew,
obtained in time, T, can be expressed as
Ew ¼ Pw T ¼ ½rAv3 T (2.6)
Hence, decreasing the wind speed requires an increase in time taken to obtain
the same wind energy. The swept area of turbine rotor, A ¼ pD2 =4, where D is the
blade diameter (m), and so (2.4) can be rewritten as
Pw ¼ prD2 v3 =8 (2.7)
Thus, the wind power is proportional to the square of blade diameter as well.
This means that depending on increasing the blade diameter to increase the wind
power is more cost-effective as the turbine cost may be considered as a quantity
proportional to the blade diameter. More discussion of wind power dependency on
2,200
2,000
1,800
1,600
Specific power (W/m2)
1,400
1,200
1,000
800
600
400
200
0
0 2 4 6 8 10 12 14 16
Wind speed (m/s)
Figure 2.7 Variation of specific power (or power intensity) versus wind speed
those different variables is given in Example 2.1, in addition to the calculation of

average wind energy.
Example 2.1 From (2.4), it is seen that doubling the wind speed increases the
power by a factor of 8, whereas from (2.7), doubling the blade diameter increases
the power to four times its value.
With respect to the wind energy, it is the product of wind power and the time as
given by (2.6). So, for a turbine with constant swept area and assuming v ¼ 10 m/s,
the wind energy E ¼ ½rA(10)3T. If v is increased to 15 m/s, the corresponding
energy will be E ¼ ½rA(15)3T; that is, the energy E is increased by a factor of (1.5)
3
. So, the time T can be decreased by a factor of (1/1.5)3 to obtain the same wind
energy. Similarly, when v decreases and becomes 5 m/s (i.e. half its value), time T
is increased by a factor of 8 which equals (1/0.5)3.
Considering a period of 50 h, wind velocity is 10 m/s during the first 20 h, and
then increases to 15 m/s for the rest of the time. Obviously, the wind energy is the
sum of that obtained in 20 and 30 h. Assuming A ¼ 1 m2, then
E ¼ ½ð1:225Þð1Þð10Þ3 ð20Þ þ ½ð1:225Þð1Þð15Þ3 ð30Þ

¼ 12;250:0 þ 62;015:62
¼ 74;265:62 Wh
If the average wind speed is proposed to be used to calculate the average wind
energy for the total time of 50 h, wind energy is Eav ¼ ½(1.225)(1)(12.5)3
(50) ¼ 59,814.45 Wh, which is a wrong answer. Another proposed approach is to
use the average cubic value of speeds, that is, (v3)av ¼ ½[(10)3 þ (15)3] ¼ 2,187.5
and the corresponding energy is Eav ¼ ½(1.225)(1)(2,187.5)(50) ¼ 66,992.18 Wh
which is again a wrong answer.
The best way to solve this problem is to calculate the average cubic speed as a
summation of each cubic speed multiplied by its fraction of time, that is,
X
ðv3 Þav ¼ v3i ðfraction of hours at vi Þ
i
X (2.8)
¼ v3i ðprobability ðv ¼ vi Þ
i
Thus,

3 20 30
v av ¼ ð10Þ3 þ ð15Þ3 ¼ 400 þ 2;025 ¼ 2;425
50 50
and the corresponding Eav ¼ ½(1.225)(1)(2,425)(50) ¼ 74,265.625 Wh, which is
an accurate result. Another form of (2.8) can be written as
P 3
3 i vi ðhours at vi Þ
v av ¼ P (2.9)
hours
Wind energy 25
It is to be noted that the term ‘fraction of hours at vi’ in (2.8) represents the
probability of the wind having v ¼ vi. However, more details about statistics and
probability application are given in Section 2.2.3.
For a VAWT, the swept area may be approximated as two-thirds the area of a
rectangle with height equal to the rotor height (H) and width equal to the rotor
diameter (D). Thus,
2
A ¼ DH (2.10)
3
2.2.2 Factors affecting wind power

As seen in (2.4), the wind power is half the product of air density, swept area and
cube of wind speed. Therefore, it is of interest to discuss some physical factors
affecting the determination of air density as well as those having an impact on wind
speed [1].
2.2.2.1 Factors affecting air density

Factors affecting air density can be specified by examining the ideal gas law that is
PV ¼ nRT (2.11)
where
D
● P ¼ the absolute pressure (atm), 1 atm ¼ 101.325 kPa;
D
● V ¼ the volume (m3);
D
● n ¼ the mass (mol);
D
● R ¼ the ideal gas constant ¼ 8.2056 105 (m3atm)/(kmol);
D
● T ¼ the absolute temperature (K) and K ¼ C þ 273.15.
Assuming Mw stands for molecular weight of the gas (g/mol), it may equal
28.97 g/mol and air density can be expressed as
nMw
r¼ 103 kg=m3 (2.12)
V
Substitute V from (2.11) into (2.12) to obtain
PMw
r¼ 103 kg=m3 (2.13)
RT
Thus, both the pressure, P, and temperature, T, have an impact on determining the
air density, r.
Pressure impact
As the air pressure is a function of altitude and the WTs may be located at sites
above sea level, it is useful to determine the pressure at different heights in terms of
air density.
Considering a static column of air with cross section A and taking a slice of air
with thickness ds and density r, the difference of pressure between the top and
bottom of the slice is given by
dp ¼ grds
Hence,
dp
¼ gr (2.14)
ds
where g is the gravitational acceleration and equals 9.806 m/s2.
Substituting r as in (2.13) into (2.14) gives
dp gMw
¼ p 103 (2.15)
ds RT
Assuming T is constant throughout the air column, an acceptable assumption

as typically the rate of temperature drop is about 6.5 C/km, then, substituting the
constants and unit’s conversion factors at T ¼ 15 C, (2.15) can be rewritten as
dp
¼ 1:185 p 104 (2.16)
ds
and the solution can take the form as follows:
4
H
p ¼ po e1:18510 (2.17)
where Po is the pressure of one atm, and H is the height in meter (m).
Therefore, using (2.13) air density will be
po Mw 1:185104 H
r¼ e 103 kg=m3 (2.18)
RT
Temperature impact
As shown in (2.13) or (2.18), air density is inversely proportional to the tempera-
ture; that is, when the temperature increases, the air density decreases. The standard
value of air density is 1.225 kg/m3 at 15 C.
2.2.2.2 Factors affecting wind speed

WTs exploit the wind energy close to the ground. So, the wind stream is turbulent
and air movement experiences friction because of ground roughness. Not only due
to that but also due to irregularities such as forests and buildings, wind speed is
considerably slowed. On the other hand, it is very useful to increase wind speed as
the wind power is proportional to the cube of wind speed; that is, it is significantly
increased with an incremental increase in wind speed. Therefore, the way to get the
WT facing into the wind with high speed is only to mount the turbine on a
taller tower.
Wind energy 27
Impact of earth’s surface roughness on wind speed is commonly characterized

by an approximated expression to get wind speed variation with elevation as
follows:

v lnðH=zÞ
In Europe : ¼ (2.19)
vo lnðHo =zÞ
where
D
v ¼ wind speed;
D
vo ¼ wind speed at height Ho;
D
H ¼ rotor height;
D
Ho ¼ reference height (it is often a height of 10 m);
D
z ¼ roughness length (it equals 2 104 for water surface, 0.03 for open areas,
0.1 for farm land, 0.4 for urban districts and 1.6 for dense urban or forest) or in the
United States:
a
v H
¼ (2.20)
vo Ho
D
where a ¼ the friction coefficient and roughly equals 1/7 for open terrain (it
ranges from 0.1 for smooth ground or calm water to 0.4 for big city with tall
buildings).
From (2.4), the wind power, Pw, at height H, referred to reference wind power,
Pwo, at height Ho can be obtained as
3
Pw v
¼
Pwo vo

lnðH=zÞ 3
¼ European form (2.21)
lnðHo =zÞ
3a
H
¼ US form
Ho
Example 2.2
(a) Find the air density at 1 atm and 35 C.
(b) Find the air density at 20 C and an elevation of a height of 2,500 m. What is
its value at temperature 7 C? Also find the power density.
(c) A wind speed of 7 m/s is indicated by an anemometer mounted at height of
10 m. Assuming a temperature of 15 C, pressure of 1 atm and friction
coefficient of 0.2, find the wind speed and specific power at a height of 40 m.
Solution
(a) Using (2.13), air density, r, is
pMW 103 1 28:97 103

r¼ ¼ ¼ 1:146 kg=m3
RT 8:2056 105 ð273:15 þ 35Þ
It shows that air density decreases with increase in the temperature.

(b) Using (2.17), the pressure at 2,500 m height is
4 4
H
p ¼ po e1:18510 ¼ 1 e1:18510 2;500
¼ 0:744 atm
It shows that the pressure decreases with increase in height.
pMW 103 0:744 28:97 103

r¼ ¼ ¼ 0:896 kg=m3
RT 8:2056 105 293:15
It shows the combined effect of increasing both temperature and height.
If the temperature is decreased to 7 C, air density will be
pMW 103 0:744 28:97 103

r¼ ¼ ¼ 0:937 kg=m3
RT 8:2056 105 280:15
The corresponding power density is:
P 1 3 1
¼ rv ¼ 0:937 73 ¼ 170:47 W=m2
A 2 2
(c) Using (2.20) as it is common in the United States, the velocity at 40 m
height is
a 0:2
H 40
v ¼ vo ¼7 ¼ 9:1 m=s
Ho 10
and
Specific power ¼ ½ 1:225 9:13 ¼ 461:56 W=m2

The specific power at 7 m/s is ½ 1.225 73 ¼ 210.09 W/m2.
So, increasing the height from 10 to 40 m yields an increase of more than
the double in the specific power.
If (2.19) is used as it is common in Europe and assuming the roughness
length, z, equals 0.4, the velocity and specific power are as follows:
lnðH=zÞ lnð40=0:4Þ
v ¼ vo ¼7 ¼ 10 m=s
lnðHo =zÞ lnð10=0:4Þ
Specific power ¼ ½ 1:225 103 ¼ 612:5 W=m2
Wind energy 29
2.2.2.3 Rotor efficiency

It is very important to evaluate the wind power intensity (W/m2) to access the
economic feasibility when selecting a site location. The effective power extracted
from wind can be derived in terms of the airflow speed just as it arrives at the
turbine, v1, and the velocity just departing it, v2. Assuming the wind is passing
through swept area, A, of rotor blades at average speed, vav ¼ ½(v1 þ v2), then
using (2.3) gives
m_ ¼ rAvav ¼ ½rAðv1 þ v2 Þ (2.22)
The net wind mechanical power of the turbine is imposed by the difference in
the kinetic energy in the wind speed just reaching and just leaving the turbine.
Incorporating (2.2) and (2.22) obtains

1 1 v2 v2
Pw ¼ rAðv1 þ v2 Þ v21 v22 ¼ rAv31 1 þ 1 22 (2.23)
4 4 v1 v1
or
Pw ¼ ½CP rAv31 (2.24)

where

D v2 v22
CP ¼ power coefficient ¼ ½ 1 þ 1 2 (2.25)
v1 v1
Energy conversion from one form to another, e.g. from kinetic energy in the
wind to mechanical power, mostly has some constraints. The analysis necessitates
understanding what happens to the wind as it passes through a WT. A portion of
wind kinetic energy approaching the WT at speed, v1, is extracted by the turbine,
and consequently, the wind leaving the turbine has a lower velocity, v2, and
reduced pressure. This causes the air to expand downwind of the turbine. One may
think that the maximum power extracted from the wind can be achieved when the
total kinetic energy of the wind is already extracted and converted into mechanical
power, that is, v2 is zero. This is not true in reality as the air would come to a
complete stop behind the turbine and prevent any more of the wind to pass through
the rotor. In addition, it does not make sense for v1 equals v2; that is, downwind
velocity and upwind speed are the same and no energy is extracted by the turbine.
The question now is, what is the downwind velocity for maximum power extracted
by the turbine? The answer can be obtained by determining the value of CP, in
(2.25), that satisfies the condition of maximum power as in the following.
Assuming that l ¼ v2 =v1 , then (2. 23) and (2. 25) can be rewritten as
1
Pw ¼ rAv31 ð1 þ lÞ 1 l2 (2.26)
4

CP ¼ ½ð1 þ lÞ 1 l2 (2.27)
Comparison of (2.26) with (2.24) shows that the fraction of the wind’s power
extracted by the blades (in other words, rotor efficiency) is CP as given in (2.27).
CP versus l is depicted in Figure 2.8. It shows that the rotor efficiency reaches a
maximum when downwind speed is one-third of its wind speed.
Differentiating (2.27) with respect to l and equating to zero results in l ¼ 1/3
and the corresponding rotor efficiency (it is called ‘Betz limit’) is 59.26%. In
practice, rotor efficiency is not as high as 59%; a typical efficiency is between 35%
and 45%.
2.2.2.4 Tip-speed ratio

Tip-speed ratio (TSR) is defined as the ratio of rotor tip speed, vrotor, to the wind
speed, v:
vrotor rpm pD
TSR ¼ ¼ (2.28)
v 60v
where
● rpm is the rotor speed (revolutions per minute);
● D is the rotor diameter (m);
● v is the wind speed (m/s).
It is usual to determine rotor efficiency as a function of TSR that depends on
the rotor tip speed. If the rotor turns too slowly, its efficiency drops off as much
0.6
0.5
0.4
Rotor efficiency, Cp
0.3
0.2
0.1
0.0
0.0 0.2 0.4 0.6 0.8 1.0
l = Vd /v
Figure 2.8 Rotor efficiency versus ratio of downwind speed to upwind speed
Wind energy 31
upwind does not affect the blades. If the rotor turns too fast, its efficiency is
reduced as well because of the turbulence. Variation of rotor efficiency versus TSR
for turbines with different number of blades is shown in Figure 2.9 [2].
Example 2.3 Two-blade WT, 40-m rotor diameter, produces 400 kW at a wind
speed of 12 m/s. Using the standard air density value of 1.225 kg/m3 at 15 C and
1 atm, find the following:
(a) the rotor rpm assuming that it operates with a TSR ¼ 5.5;
(b) the rotor tip speed;
(c) the gear ratio needed to rotate the generator at 2,000 rpm;
(d) the overall efficiency of the WT system.
Solution
(a) Using (2.28),
TSR 600v 5:5 60 12
rpm ¼ ¼ ¼ 31:5 rpm
pD p 40
(it means about 2 s per revolution).
(b) Tip speed ¼ (31.5 p 40)/60 ¼ 65.97 m/s.
60
Betz limit
Theoretical turbine
50 High-speed
two-blade turbine
40
Three-blade
Rotor efficiency, Cp (%)
turbine
30
Multi-blade Darrieus turbine
turbine
20
10
0
0 1 2 3 4 5 6 7 8
Tip-speed ratio
Figure 2.9 Rotor efficiency versus TSR

(c) Gear ratio ¼ Generator speed 2;000

Rotor speed ¼ 31:5 ¼ 63:5.
(d) Using (2.4), the wind power is

p 402
Pw ¼ ½rAv3 ¼ Pw ¼ ½ 1:225 ð12Þ3 ¼ 1;330 kW
4
Thus, the overall efficiency, hovall, is
hovall ¼ 400=1;330 ¼ 30%
2.2.3 Average wind power

As seen in (2.8), the average of cube of wind speed can be calculated by summing
the product of each wind speed and its probability. In next two subsections, addi-
tional probability characteristics for the wind may be considered to compute the
average value of the cube of wind speed.
2.2.3.1 Probability density functions

Probability density function (pdf) is a continuous function that presents the discrete
wind histogram. Its variation against wind speed variation is shown in Figure 2.10.
It is to be noted that the area under the curve is equal to unity, whereas the area
under the curve between two wind speeds, v1 and v2, equals the probability of the
wind that has a speed within those two speeds. Relations next are used to, mathe-
matically, express these features:
g
ð1
Probability of wind speed ¼ f ðvÞdv ¼ 1 for 0 v 1
0
ð v2 (2.29)
¼ f ðvÞdv for v1 v v2
v1
where f (v) is the pdf of wind speed.
0.1
0.1
0.0
Total area under
f (v)
0.0 entire curve = 1
0.0
Probability
0.0 of wind speed v
v1 ≤ v ≤ v2
0.0
v1 v2 Wind speed
Figure 2.10 Probability density function versus wind speed

Wind energy 33
The number of hours in a period, T, at which the wind speed is within v1 and v2
can be computed by the product of its probability and the time, T, in hours, because
f (v) expresses the fraction of time the wind speed is between v1 and v2. Thus,
ð v2
Number of hours=T ðv1 v v2 Þ ¼ T f ðvÞdv (2.30)
v1
For instance, if T ¼ 8,760 h, (2.30) gives the number of hours per year, the
wind speed is between v1 and v2.
The average value of the cube of the wind speed using pdf can be found in the
same manner as that applied for the discrete approach in (2.8) and is given by
ð1
3
v av ¼ v3 f ðvÞdv (2.31)
0
2.2.3.2 Probability and statistics

To increase confidence in the data available, a large sample of wind should be
taken over many years as the wind movement is a random phenomenon. This is
always difficult and shorter periods are often used. Therefore, the data are averaged
over calendar months and can be described by what is called ‘Weibull pdf ’ for
characterizing the statistics of wind speed. This function is given as
k v k1 ðv=cÞk
f ðv Þ ¼ e for 0 v 1 % of hours=year=ðm=sÞ (2.32)
c c
where
● k is defined as the shape parameter and is practically presented for most sites
within the range of 1.5 to 2.5;
● c is known as the scaling factor and pertains to the number of days with high
wind speed.
The shape of the curve representing (2.32) varies with the value of the
parameter k as plotted in Figure 2.11 assuming a constant scaling factor, c.
At most sites around the world, k ¼ 2, and substituting this value into (2.32)
gives what is called ‘Rayleigh pdf’ as
2 v ðv=cÞ2
f ðv Þ ¼ e (2.33)
c c
Rayleigh pdf can be plotted at various values of the scaling parameter, c,

against wind speed variation as shown in Figure 2.12. It is noted that as c increases,
the curve shifts towards higher wind speed.
14 k3
k1 < k2 < k3
12 c = constant
k1
k2
10
f (v) × 10–2
0
0 2 4 6 8 10 12 14 16 18 20 22 24
Wind speed (m/s)
Figure 2.11 Weibull pdf at various values of k and constant c
12
c = 3 m/s
c = 5 m/s
10 c = 7 m/s
c = 9 m/s
Rayleigh pdf (%)
8 c = 11 m/s
c = 13 m/s
6
0
0 5 10 15 20 25
Wind speed (m/s)
Figure 2.12 Rayleigh pdf at various values of scaling factor, c
Substituting f (v) from (2.33) into (2.31), the average value of the cube of the
wind speed is
ð1 pffiffiffi
3 2v 2 3 p 3
v av ¼ v3 2 eðv=cÞ dv ¼ c (2.34)
0 c 4
Consequently, the average wind power can be calculated using
pffiffiffi
1 3 p
Pav ¼ rA v3 av ¼ rAc3 (2.35)
2 8
Wind energy 35
Example 2.4 Deduce the average wind speed using Rayleigh pdf and compare the
result with that obtained for the value of the cube of the wind speed. Examine the
impact on the average wind power.
Solution Like the equation of (v3)av in (2.31), the average speed can be written as
ð1
vav ¼ vf ðvÞdv (2.36)
0
Substituting Rayleigh pdf, f (v) as in (2.33) into (2.36) gives

ð 1 2 pffiffiffi
v ðv=cÞ2 p
vav ¼ 2 e dv ¼ c (2.37)
0 c 2
Hence,
2
c ¼ pffiffiffi vav (2.38)
p
Combining (2.34) and (2.38) gives
3 6
v av ¼ ðvav Þ3 (2.39)
p
Therefore, the average wind power in terms of the average of the cube of the wind
speed, (v3)av, is (6/p) times that in terms of average speed to power 3, that is, (vav)3.
For further explanation, it was found in Example 2.1 that using either the
arithmetic average speed ((v1 þ ... þ vn)/n) or the arithmetic average of the cube of
the speed ((v13 þ . . . þ vn3)/n) yields a wrong result and must be corrected by
multiplying each wind speed to the power of 3 by its probability.
This example emphasizes that even in case of using pdf, the cube of wind speed
must take the form of ðv3 Þav , not ðvav Þ3 , where the ratio between them is (6/p).
2.3 Wind characteristics

Basic relations of wind power estimation have been explained in Section 2.2. It can
be, simply, seen that the rotor blades extract energy, in the form of kinetic energy,
from the wind to rotate the rotor, then a gearbox is used to adapt the rotor speed at
the nominal rotating speed of the generator shaft and finally the generator delivers
the electrical power (Figure 2.13). In addition, the characteristics of a WT are
introduced in brief in the next subsections to specify the performance of the WT.
2.3.1 Airfoils
Airfoils are aerodynamic surfaces, for example, air plane wings, propeller blades
and wind rotor blades. Based on the shape of their cross section, they must generate
Blades
Three-phase
output
Rotor
Generator
Hub shaft
Gearbox Generator
Figure 2.13 Blades–rotor–gearbox–generator set
pL
A
pD
x
Wind direction
Figure 2.14 Drag and lift pressures on airfoil
a big lift with a corresponding small drag. The lifting force causes the turbine
blades to rotate. It is created as the pressure on the top of airfoil is less than that
under it. This difference of pressure is due to the fact that the air moving on the top
travels a longer distance before rejoining again the air passing immediately under
the airfoil [3].
Figure 2.14 depicts a cross section through an airfoil describing the forces on a
rotating turbine blade. An arbitrary reference plane is chosen and represented in x–y
frame by asymmetric axis A–B, where the region above the plane differs from that
under it. The air flows relative to the airfoil along x-axis, and so, a force is acting on
Wind energy 37
the foil. This force has two components: (i) drag component parallel to the wind
velocity, v, along the x-axis produced by pressure pD and (ii) lift component per-
pendicular to the velocity produced by pressure pL . The angle between wind
direction and the reference line is defined as the ‘angle of attack’ b.
Pressures pL and pD are determined experimentally under specified conditions
for each shape of the airfoil. It is found that the pressures are related to the dynamic
pressure, ½rv2 , by proportionality constants: lift coefficient CL and drag coefficient
CD , respectively.
pL ¼ ½rv2 CL
pD ¼ ½rv2 CD g (2.40)
Here, CL and CD are mainly functions of the angle of attack. A typical plot of both
CL and CD versus the angle of attack is shown in Figure 2.15. Increasing the angle
of attack improves lift component associated with an increase in drag component as
well. On the other hand, if the lift increase is too much, a phenomenon called ‘stall’
takes place. When a blade stalls, the airflow over the top does not stick to the
surface and the resulting turbulence destroys the lift.
Based on the effect of lift and drag forces, two designs of turbines are mostly
used to generate electricity from wind: drag turbines and lifting turbines. The
operation of drag turbine is based on the friction caused by wind on turbine blades.
The wind creates a force in the direction of its blowing. The turbine blade on which
the wind impinges cannot move faster than the wind itself.
Lifting turbines operate through the lifting effect produced by wind where the
wind generates a force normal to the direction of its blowing. The turbine blade can
move faster than the wind itself.
1.2
0.8 CD
0.4 CL
–0.4
–0.8
0 60 120 180 240 300 360°

Angle of attack
Figure 2.15 Typical lift and drag coefficient versus angle of attack
Power curve describes the relationship between wind speed and generator
electrical power output. It is of crucial importance to technically evaluate the per-
formance of a specific WT [4].
Power curve, in general form, encompasses four bands of wind speed: (i) cut-
in, (ii) normal operation, (iii) high and (iv) cut-out wind speeds (Figure 2.16). The
first band represents the cut-in wind speed band. It ranges from zero to what is
called ‘cut-in speed’, which is the minimum speed of generation, VC(in). Below this
speed, wind may not have enough power to supply both friction losses and gen-
erator field windings. At most, the power generated just supplies the friction losses.
Therefore, no power is generated at wind speed below VC(in).
The second band ranges from VC(in) to the rated speed, Vr. Starting from VC(in)
and increasing the wind speed, the power delivered by the generator tends to rise as
it is proportional to the cube of wind speed. Reaching the rated wind speed Vr, the
generator delivers its rated power in correspondence to its design. So, this band is
the normal operation maintained by controlling the blade position with respect to
the direction of wind attack.
The third band is the high wind speed. When the speed exceeds the rated value,
Vr, high-speed controllers can be actuated to maintain a maximum constant output
power limited by generator capacity. Pitch-control system to reduce the blade’s
angle of attack to shed some of wind or using stall control to increase the angle of
attack to create stall can be applied until a certain speed limit is reached, Vlimit. The
fourth band, above the speed limit, corresponds to the wind having excessive speed
Rated power
Generated power (kW)
VC(in) Vr VC(out)
Vlimit
First Fourth
Second band Third band
band band
Wind speed (m/s)
Figure 2.16 WT power curve

Wind energy 39
that is too strong and may cause mechanical damage to the generator, as well as
there is a real danger to the WT. At this speed, called ‘cut-out wind speed’, VC(out)
the machine must be shut down and no power is generated.
2.3.2 Capacity factor

Capacity factor, CF, is used to estimate the actual energy delivered by the WT as a
percentage of the rated energy. It can be calculated as
Actual energy delivered Average power

CF ¼ ¼ (2.41)
Pr 8;760 Rated power
where Pr is the rated power of the WT generator in kW.

It is noted that CF is a convenient measure of WT capability for producing
electrical power over a year or certain number of hours.
Example 2.5 A WT has 1,500 kW rated power, 64-m rotor diameter, cut-in wind
speed of 5 m/s, rated wind speed of 16 m/s and cut-out wind speed of 27 m/s. Using
Rayliegh statistics (k ¼ 2) and considering the scaling factor c equals 12, find the
following:
(a) The number of hours per year at which (i) wind speed is less than or equal to
cut-in speed and (ii) the WT should be shut down when the wind speed is
excessive.
(b) The energy generated per year at the rated power.
Solution
(a) As given by (2.37), the average wind speed is
pffiffiffi
p
vav ¼ c ¼ 10:6 m=s
2
From (2.29) and (2.32), the cumulative distribution function for Weibull sta-
tistics of wind speed equals to or is less than a specific value, V, and is the
integration of the function from zero to that speed, that is,
ð V k1
k v k
f ð0 v V Þ ¼ eðv=cÞ dv for 0 v V (2.42)
0 c c
Thus, it can be rewritten as

ðx
f ðv V Þ ¼ ex dx (2.43)
0
where
v k dx k v k1
x¼ ; then; ¼
c dv c c
Hence,
k
f ðv V Þ ¼ 1 eðv=cÞ (2.44)
In case of Rayleigh statistics, (2.44) is of the form
2
f ðv V Þ ¼ 1 eðv=cÞ (2.45)
Substitute c from (2.37) into (2.45) to get
2
f ðv V Þ ¼ 1 eðp=4Þðv=vav Þ (2.46)
Therefore, probability function of wind speed greater than V is the inte-
gration of the function from V to infinity, or in other words, it equals one
minus the probability of wind speed less than V as the summation of the two
probabilities represented by the total area under the curve of the function
against wind speed is unity. This means that
2 2
f ðv V Þ ¼ 1 ½1 eðp=4Þðv=vav Þ ¼ eðp=4Þðv=vav Þ (2.47)
Apply (2.46) to get the probability that the wind speed is VC(in) as
2 2
f v VCðinÞ ¼ 1 eðp=4Þðv=vav Þ ¼ 1 eðp=4Þð5=10:6Þ ¼ 0:16

The number of hours the wind v VCðinÞ ¼ 8;760 0:16 ¼
1;401:6 h=year.
Similarly, the hours when wind speed is greater than VC(out) can be obtained
using (2.47).
2 2
Hours v VCðoutÞ ¼ 8;760eðp=4Þðv=vav Þ ¼ 8; 760eðp=4Þð27=10:6Þ
¼ 53:6 h=year
(b) As is clear from the WT power curve, the rated power is delivered when the
wind speed is between rated speed, Vr, and cut-out speed, VC(out). Probability
(Vr v VC(out)) is the difference between probability (v Vr) and prob-
ability (v VC(out)), that is,

f Vr v VCðoutÞ ¼ f ðv Vr Þ f v VCðoutÞ
2 2
¼ eðp=4Þð16=10:6Þ eðp=4Þð27=10:6Þ
¼ 0:167 0:006 ¼ 0:161

Thus, hours Vr v VCðoutÞ ¼ 8;760 0:161 ¼ 1;410:36 h=year
Wind energy 41
Consequently,
The rated energy delivered by the wind turbine
¼ rated power no: of hours
¼ 1;500 1;410:36 ¼ 2:11554 106 kWh=year
2.4 Electrical generators in wind systems

WTs can be equipped with any type of rotating generators such as DC or AC generators.
DC generators require maintenance and are relatively bulky and inefficient. AC gen-
erators are mostly used in wind systems and can be either asynchronous (induction) or
synchronous type. Asynchronous generator may be squirrel cage induction generator
(SCIG) or doubly fed induction generator (DFIG), whereas permanent magnet syn-
chronous generator (PMSG) is commonly used as a synchronous generator.
Based on classifying the WTs as constant speed and variable speed turbines,
Figure 2.17 shows the schematic diagrams of the general configuration of the three
most commonly used generator–WT systems [4,5]. SCIG can be used for constant
speed WTs as SCIG speed varies only little with wind speed because of the gen-
erator slip. The generator and WT rotor are coupled through a gearbox as the
optimal rotor and generator speed ranges are different. Then, a transformer may be
used to directly connect the generator terminals with the utility grid and a shunt
capacitor bank may be used to feed the generator with the required reactive power,
particularly, in stand-alone operation (Figure 2.17(a)). If the wind speed increases,
automatically the angle of attack increases. At a certain wind speed, above the rated
value, the angle of attack will reach the value where stall occurs. Therefore, stall
control is used to keep the power produced by the turbine near the rated power.
For variable speed WT, either DFIG for limited variable speed or direct-drive
PMSG for full variable speed WT may be used. In case of using DFIG, a gearbox is
used. The stator is directly connected to utility grid and the rotor is connected to a
converter (Figure 2.17(b)). Speed variation ranges from about 60% to 110% of the
rated speed using variable speed capability to keep the TSR at the value resulting in
optimal energy. If the wind speed exceeds the rated value, a pitch control is applied
to reduce the power to be around its rated value.
When using PMSG, no gearbox is necessary as the generator rotates at very
low speed. The total power generated flows to the utility grid through a converter to
fix the varying generator frequency at the grid frequency (Figure 2.17(c)). At high
wind speed, above the rated speed, the power increase over the rated value is
reduced by pitching the blades. Accordingly, stall control and pitch control are
mainly used with constant speed and variable speed turbines, respectively.
2.4.1 Asynchronous induction generators

Induction generators (IGs) have outstanding operation as they have very robust
construction, mechanical simplicity and lowest cost, among generators. In addition,
Rotor
blades
Gearbox SCIG
Utility grid
Capacitor bank
(a)
Rotor
blades
Gearbox DFIG
Utility grid
Converter
(b)
Rotor
blades
PMSG Converter
Utility grid
(c)
Figure 2.17 Classic WT–generator systems. (a) Constant wind speed turbine with
SCIG. (b) Variable wind speed turbine with DFIG. (c) Variable wind
speed turbine with PMSG
IGs provide natural protection against short circuits and damping any current surge
by the magnetization path of the iron core without fear of demagnetization. On the
contrary, a major disadvantage is that the IG needs a reasonable amount of reactive
power as it does not contain permanent magnets (PMs) or other means to provide
Wind energy 43
magnetic excitation. Therefore, reactive power must be fed externally to establish

the magnetic field necessary to convert the input mechanical power into output
electrical power. External sources may be the grid in interconnected applications.
In stand-alone applications, reactive power may be supplied by the load itself, by a
capacitor bank connected across generator terminals, or by power electronic
systems.
2.4.1.1 Squirrel cage induction generators

The stator of this type of generator has a three-phase stator winding and the rotor
consists of a number of copper or aluminium bars shorted together at their ends
forming a cage. Applying a three-phase voltage to the stator winding, three-phase
currents will flow in the stator windings and produce a rotating magnetic field. Its
speed, called synchronous speed, is determined by the number of poles in the wind-
ings and the frequency of the current. When the rotor rotates at a speed other than the
synchronous speed, voltage is induced in the rotor cage winding caused by the rela-
tive speed (called slip) between the rotor and the rotating stator magnetic field. This
causes a current to flow in the rotor winding. Interaction of the associated magnetic
field of the rotor with the stator field results in a torque acting on the rotor [2].
Relations and equivalent electric circuit at steady state
For the induction machine to operate in generator mode, the rotor must spin faster
than the synchronous speed; that is, the operational speed is advanced with respect
to the stator magnetic field rotation. In quantitative terms, let ns and fs be the
synchronous mechanical speed (rpm) and synchronous frequency (Hz), respec-
tively. The induced voltage, resulting in output power, is proportional to the rela-
tive speed difference between the electrical synchronous rotation and the
mechanical rotation with a speed-slip factor, s, defined as
ns nr
s¼ (2.48)
ns
where nr is the rotor speed (rpm).

Denoting fs and fr as the electrical stator frequency and rotor frequency,
respectively, the interrelation between them can be expressed as
p
fr ¼ ðns nr Þ ¼ sfs (2.49)
120
where fs ¼ (p/120)ns and p is the number of poles.

It is to be noted that in case of no relative speed difference, that is, ns ¼ nr , the
rotor frequency is DC as s ¼ 0. For other speeds, s is a negative value if ns < nr ,
and the voltage induced on the rotor, Er, is directly proportional to s and transferred
to the stator windings. The stator and rotor are analogous to the primary and sec-
ondary of a transformer with the difference that the secondary is a rotating element.
So, in the following relations, the symbols of stator and rotor parameters referred to
the stator are subscripted by #1 and #2, respectively.
The voltage induced on the rotor, Er, is established for any speed with respect
to the blocked rotor voltage, Ero, as
E2 ¼ Er ¼ sEro (2.50)
and the stator voltage, E1, is

Er
E1 ¼ g (2.51)
s
where g is the turns ratio.

Similarly, the rotor current, I2, is obtained by
1
I2 ¼ I r (2.52)
g
and
Er sEro
Zr ¼ ¼ ¼ Rr þ jsXro (2.53)
Ir Ir
where Xro is the ‘blocked rotor’ reactance and Rr is the rotor winding resistance.
From (2.51), the rotor impedance can be calculated at any rotor speed from the
blocked rotor test as
Ero Rr Zr
Zro ¼ ¼ þ jXro ¼ (2.54)
Ir s s
Then, using (2.51) through (2.54), the rotor impedance referred to the primary can
be obtained by

E1 Rr
Z2 ¼ ¼ g2 þ jXro (2.55)
I2 s
and in terms of stator values, it can be rewritten as

R2
Z2 ¼ þ jX2 (2.56)
s
Consequently, the rotor current, I2, is

E1 E1 s
I2 ¼ ¼ (2.57)
Z2 R2 þ jX2 s
Accordingly, the equivalent circuit of the IG can be drawn as in Figure 2.18. It
is valid only for perfect sinusoidal excitation and balanced conditions. The para-
meters R1 and X1 are the resistance and leakage reactance of the stator; R2 and X2
are the resistance and leakage reactance of the rotor; and Rm and Xm are the
Wind energy 45
R1 jX1 jX2
Im I2
I1
Vt Rm jXm E1 R2/s
Figure 2.18 Equivalent circuit of the IG
equivalent core loss resistance and magnetizing reactance, respectively. The

equivalent per-phase impedance, Z, seen at the terminals is
Z2 Zm
Z ¼ Z1 þ (2.58)
Z2 þ Zm
where
Z1 ¼ R1 þ jX 1
R2
Z2 ¼ þ jX 2
s
1
Zm ¼ 1 1
Rm þ jXm
The rotor equivalent resistance, R2/s, depends on the rotor speed, which is a
major difference with respect to a transformer equivalent circuit as the rotor voltage
is subjected to a variable frequency causing Er, Rr and Xr to be variable. However,
the equivalent circuit shown in Figure 2.18 is usually considered for analysis.
Using (2.57), the apparent rotor power can be formulated by
E12 s
S ¼ E1 I2 ¼ (2.59)
R2 þ jX2 s
Therefore, it is seen that the rotor power factor depends on the slip, s, and other
parameters. In respect of voltage, the quadrature leading current component is
almost constant for all terminal voltages. So, the reactive power absorbed by the
equivalent circuit must be supplied by an external source such as utility grid,
capacitor banks or an electronic compensator.
Equivalent circuit in transient state
Machine model is necessary for examining the dynamic and transient behaviour of
the machine, to develop machine control and define the machine variables in
transient state as well as in steady state. Therefore, the dynamic model is used to
provide information to the planners to design enabling controllers for transition
from one state to another. Consequently, the dynamic model is generally repre-
sented by differential equations.
A reasonable representation of an induction machine can be obtained by con-

sidering the effects of mechanical transients and rotor electrical transients with a
single time constant. The effect of stator electrical transients on system response
usually can be neglected. The transient behaviour of an induction machine is
represented by the equivalent circuit shown in Figure 2.19. The rate of change of
the voltage behind transient reactance, X0 , can be described by the following dif-
ferential equation:
0
dE 0 1 0 0
¼ j2psE ½E jðX X ÞIt (2.60)
dt To
and the terminal current is given by
0 1
It ¼ ðVt E Þ (2.61)
R1 þ jX 0
where
To ¼ X22pfR
þ Xm
r
and is defined as the rotor open-circuit time constant in seconds.
X and X0 are calculated from the steady-state equivalent circuit by neglecting
rotor resistance. The values are
0 Xm X2
X ¼ X1 þ Xm and X ¼ X1 þ
X m þ X2
It is to be noted that the fixed speed turbine with SCIG directly connected to
the grid has a number of drawbacks. The reactive power and, consequently, the grid
voltage level are uncontrollable. The power varies due to blade rotation causing
voltage variations in the grid. Sensitivity to flicker is high at the resonance fre-
quency that the generator may have. Most of these drawbacks can be avoided by
using variable speed WTs with a doubly fed induction machine (DFIM) that can
operate in both generating and motoring mode. Obviously, in the normal operation
of a WT, it operates as a generator to deliver electric energy from the wind to
the grid.
2.4.1.2 Doubly fed induction machines

DFIM is a wound rotor induction machine (WRIM). The rotor has three-phase
windings connected to an external variable voltage and frequency source through
R1 X′
It
Vt E′
Figure 2.19 Simplified equivalent circuit of an induction machine for

transient analysis
Wind energy 47
slip rings using a power electronic converter. It is also possible to alter the rotor
reactance by modulating series inductors. The stator is directly connected to the
utility grid or an external AC voltage source to be supplied at constant frequency
and constant three-phase voltage amplitude [6].
Magnetic fields created in the air gap by both stator and rotor windings will
rotate at the same speed but with a phase shift, as a function of the torque. The
machine torque, active power, reactive power through the stator and rotor can be
controlled by adjusting the voltage (magnitude, phase and frequency) introduced in
the rotor.
Commonly, two level converters (rotor side converter ‘RSC’ and grid side
converter ‘GSC’) are used. RSC generates a three-phase voltage with variable
amplitude and frequency to control the machine torque as well as the reactive
power exchanged between the stator and the grid. GSC exchanges with the grid the
active power extracted or injected by RSC from the rotor. A schematic diagram
showing this configuration is depicted in Figure 2.20. It is noted that a transformer
is used to adapt the generated voltage to the point of common coupling (PCC).
Steady-state model
To ideally represent the DFIM by a simplified per-phase equivalent circuit, some
assumptions are considered: the stator windings are supplied by an external voltage
source or the grid at constant and balanced three-phase voltage amplitude and
frequency. Similarly, the rotor is supplied independently, by a source at constant
and balanced three-phase voltage such as power electronic converter. It is to be
noted that the frequency of the voltage and current of the stator differs from those
of the rotor. The stator frequency is fixed as the stator is directly connected to the
external source, whereas the frequency of rotor voltages and currents depends on
the machine speed. So, the rotor has controllable parameters such as current, vol-
tage, frequency and phase shift with respect to the stator voltage angle. Three
modes of operation for the DFIM, depending on the sign of the slip, are possible;
sub-synchronous operation at positive slip, synchronous operation at zero slip and
hyper-synchronous operation at negative slip.
The per-phase equivalent circuit representing the DFIM is drawn in
Figure 2.21, where the rotor parameters, current and voltages are necessarily to be
Utility grid or
AC voltage source
Rotor
blades
Transformer
Gearbox Slip
rings
WRIG
RSC GSC
Figure 2.20 Configuration of DFIM connected to grid

R1 jX1 jswsLℓr = jXℓr R2 I2

I1
V1 E1 E2 = s E1 V2
Figure 2.21 DFIM representation with different stator and rotor frequencies
(rotor current and voltage are referred to the stator)
referred to the stator. The parameters used in this representation are defined as in
the following. Vectors are written in bold fonts.
D
● Vs ¼ the applied voltage to the stator ¼ Vs ⦤ 0 ;
D
● Vr ¼ the applied voltage to the rotor ¼ Vr ⦤ j ;
D
● Is, Ir ¼ stator and rotor current, respectively;
D
● Es, Er ¼ induced emf in the stator and rotor, respectively;
D
● Rs, Rr ¼ stator and rotor resistance, respectively;
D
● Ns, Nr ¼ number of turns per phase of stator and rotor windings,
respectively;
D D
● g ¼ turns ratio ¼ Ns/Nr;
D
● L‘s, L‘r ¼ stator and rotor leakage inductance, respectively;
D
● X‘s, X‘r ¼ stator and rotor leakage impedance, respectively.
From this definition, the stator parameters are renamed as
V 1 ¼ V s ¼ V1 ⦤ 0 ; E1 ¼ Es ; I1 ¼ Is ; R1 ¼ Rs ; X1 ¼ Xls ¼ ws Lls
and consequently, the rotor parameters, current and voltages, referred to the stator
are
D
V 2 ¼ applied rotor voltage ¼ gV r ¼ V2 ⦤ j
D 1
I 2 ¼ rotor current ¼ I r ; E2 induced voltage in the rotor ¼ gEr ¼ sE1
g
D
R2 ¼ rotor resistance ¼ g2 Rr
D
Xlr ¼ rotor reactance at rotor frequency ¼ sws Llr
As shown in Figure 2.21, the following relations can be derived.
From the stator side:
V1 E1 ¼ I1 ðR1 þ jX1 Þ at stator frequency; fs (2.62)
Wind energy 49
From the rotor side:

V2 sE1 ¼ I2 ðR2 þ jXlr Þ at rotor frequency; fr (2.63)

V2 R2
E1 ¼ I2 þ jX2 at stator frequency; fs (2.64)
s s
where
Xlr
X2 ¼ ¼ jws Llr
s
Incorporating (2.62) and (2.64) gives

V2 R2
V1 I1 ðR1 þ jX1 Þ þ I2 þ jX2 ¼ 0 (2.65)
s s
According to (2.65), the per-phase equivalent circuit of the DFIM can be
derived as shown in Figure 2.22. The induced emf in the stator can be expressed as
E1 ¼ Zm ðI1 þ I2 Þ ¼ Zm Im (2.66)
where Zm is the equivalent impedance of core loss resistance, Rm, and magnetizing
reactance of the machine, that is,
1
Zm ¼ (2.67)
1=Rm þ 1=jXm
Steady-state phasor diagram

The phasor diagram can be deduced from the equivalent circuit shown in
Figure 2.22 by neglecting the mechanical power losses and using machine equa-
tions for specific operating conditions. The variables are expressed by their rms
values. Currents and fluxes of both stator and rotor are calculated by
Y1 ¼ Lm ðI1 þ I2 Þ þ L1 I1 ¼ Ls I1 þ Lm I2
Y2 ¼ Lm ðI1 þ I2 Þ þ L2 I2 ¼ Lm I1 þ Lr I2 g (2.68)
R1 jX1 jX2 R2/s

I1 I2
I1 + I 2
V1 E1 Rm Xm V2/s
Figure 2.22 Per-phase equivalent electric circuit of the DFIM

where Ls and Lm are the stator and rotor inductances given by

Ls ¼ L1 þ Lm and Lr ¼ L2 þ Lm (2.69)
Rewrite V1 and V2 in (2.62) and (2.64) in terms of their corresponding flux as
V1 R1 I1 ¼ jws Y1
(2.70)
V2 R2 I2 ¼ jsws Y2
It is easy to start the deduction of the phasor diagram for a DFIM operating in
motor mode at sub-synchronous speed and, from there, derive the phasor diagram
for the machine acting as a generator. In addition, the stator reactive power, Q1, is
assumed to be zero (the voltage, V1, and the current, I1, of the stator are in-phase).
Therefore, from (2.68) to (2.70) the phasor diagram can be drawn and is shown in
Figure 2.23 for negative slip. The stator variables are drawn in solid lines, whereas
for rotor are in dotted lines.
As shown in Figure 2.23(a), it is to be noted that Y 1 is determined by (2.70), I2
and Y 2 can be derived from (2.68), and then the rotor voltage, V2, is obtained by
(2.70). Finally, the phasor diagram indicating the phase shift angles between the
parameters of both stator and rotor is shown in Figure 2.23(b). The angles obtained
in this case show that (i) Y 1 leads Y 2 ; (ii) av
90 , ai ¼ av as Q1 ¼ 0 and
consequently P1 > 0, i.e. providing power to the stator; (iii) bv is very close to 90
as the difference is just due to R2I2; (iv) bi þ bv > 90 and V2 is leading to I2, i.e.
P2 < 0 (the rotor is absorbing power) and Q2 > 0 (the machine is magnetized
through the rotor); and (v) V2 is leading to Y 2 at sub-synchronous speed (alter-
natively, lagging at hyper-synchronous speed).
Therefore, if the machine works in generating mode as DFIG, the phasor dia-
gram may be one of two cases: (i) generating power, P1 < 0, with lagging power
factor (Q1 > 0), Figure 2.24(a); (ii) generating power, P1 < 0, with leading power
factor (Q1 < 0), Figure 2.24(b). It is noted that there is no change in the relative
position between Y 1 and Y 2 , whereas the change of magnitude exists.
I1 jw1Ѱ1 R1I1 I1
V1 V1
jsw1Ѱ2
αi
R2I2 αv
V2 V2
βv
LmI2 –LsI1 βi
θ
L1I2 Ѱ1 Ѱ1
Ѱ2
L mI 1 Ѱ2
I2 I2
(a) (b)
Figure 2.23 Phasor diagram of a DFIM at negative slip and zero stator reactive
power (a) deduced from model equations and (b) indicating phase
shifts between flux and both voltage and current for stator and rotor
Wind energy 51
I1
V1 V1
I1
Ѱ2
Ѱ1 Ѱ1 Ѱ2
(a) (b)
Figure 2.24 DFIG phaser diagrams: (a) Q1 > 0; (b) Q1 < 0
DFIM
P1, Q1 Stator Rotor P2, Q2
ωs, I1, V1 sωs, I2, V2
Tem
Pcu1 + Pcu2
Pmec
Figure 2.25 Inputs/outputs of the DFIM with motor convention
Power and torque relations

The positive sign of power flow into the machine corresponds to what is called
‘motor convention’ as shown in Figure 2.25.
Active power calculations. It can be seen that
P1 þ P2 ¼ Pmec þ Pcu1 þ Pcu2 (2.71)
where
D
P1 ¼ the three-phase active power through the stator; positive when receiving
and negative for delivering. It is given by

P1 ¼ 3Re V 1 I 1 ¼ 3jV 1 jjI 1 j cos ðav ai Þ (2.72)
D
P2 ¼ the three-phase rotor active power; positive for receiving and negative for
delivering. It is calculated by the relation

P2 ¼ 3Re V 2 I 2 ¼ 3jV 2 jjI 2 j cos ðbv bi Þ (2.73)
D
Pcu1¼ the three-phase stator copper losses
Pcu1 ¼ 3R1 jI 1 j2 (2.74)

D
Pcu2 ¼ the three-phase rotor copper losses
Pcu2 ¼ 3R2 jI 2 j2 (2.75)

Substituting (2.72) through (2.75) into (2.71) to obtain Pmec, which includes
core loss corresponding to Rm in Figure 2.22 as
Pmec ¼ 3jV 1 jjI 1 j cos ðav ai Þ þ 3jV 2 jjI 2 j cos ðbv bi Þ

(2.76)
3R1 jI 1 j2 3R2 jI 2 j2
As in (2.75), it is to be noted that R2, independent of the slip, is the electric
element which is responsible for rotor power losses. Referring to Figure 2.22, the
equivalent rotor resistance is equal to R2/s and, then, the difference R2(1 s)/s is
dependent on the slip and responsible for generating mechanical power. Similarly,
the rotor voltage, V2/s, can be represented by two sources; V2 and V2(1 s)/s as
shown in Figure 2.26. As the inductances are only pertained to the reactive power,
the elements dependent on the slip participate in the mechanical power. Therefore,
an alternative expression for calculating the three-phase mechanical power can be
written as
ð1 sÞ
Pmec ¼ 3 fR2 jI 2 j2 Re V 2 I 2 g (2.77)
s
Based on (2.72), (2.73) and (2.76), the machine efficiency, h, can be
obtained by
Pmec
h¼ for motoring operation; Pmec > 0
P1 þ P 2
(2.78)
P 1 þ P2
¼ for generating operation; Pmec < 0
Pmec
Consequently, the electro-magnetic torque in the shaft of the machine, Tem,
can be obtained in terms of the mechanical power, neglecting mechanical power
losses, as
p
Tem ¼ Pmec (2.79)
2wr
where wr is the angular frequency of the rotor (rad/s) and p is the number of poles.
Pmec can be substituted from either (2.76) or (2.79).
R1 jX1 jX2 R2 R2(1 – s)/s I2

I1
I 1 + I2 + –
V2(1 – s)/s
V1 E1 Xm V2
Figure 2.26 The electric equivalent circuit indicating the dependent elements
on the slip
Wind energy 53
Reactive power calculations. As shown in Figure 2.22 and incorporating

(2.68) and (2.70), the stator and rotor voltage can be calculated as
V 1 ¼ R1 I 1 þ jws Ls I 1 þ jws Lm I 2
(2.80)
V 2 ¼ R2 I 2 þ jsws Lm I 1 þ jsws Lr I 2
Hence, the stator reactive power, Q1, and rotor reactive power, Q2, can be
derived as follows:
Q1 ¼ 3ImfV 1 I 1 g ¼ 3ImfR1 jI 1 j2 þ jws Ls jI 1 j2 þ jws Lm I 2 I 1 g

¼ 3ws Ls jI 1 j2 þ 3ws Lm RefL2 I 1 g
(2.81)
Q2 ¼ 3ImfV 2 I 2 g ¼ 3ImfR2 jI 2 j2 þ jsws Lm I 1 I 2 þ jsws Lr jI 2 j2 g
¼ 3sws Lr jI 2 j2 þ 3sws Lm RefI 1 I 2 g
According to (2.72), (2.73), (2.77) and (2.81), the DFIM can operate at dif-
ferent modes (two generating and two motoring modes) depending on the power
and speed. Of course, the DFIM in wind energy application is used as a DFIG. The
power flow and the corresponding phasor diagram of each mode are depicted in
Figure 2.27 with Q1 > 0. Modes (a) and (b) show that the DFIG is operating at
hyper-synchronous speed and sub-synchronous speed, respectively.
Commonly, the magnitude of machine parameters (voltage, current, speed,
power, etc.) are normalized to be expressed as per unit values that are used for
DFIM
Stator Rotor
P1, Q1 P2, Q2
sωs, I2, V2 V1
ωs, I1, V1
V2 I2
Tem
I1 ψ2
Pmec ψ1
Power flow direction Phasor diagram

Mode (a) Hyper-synchronous speed, Pmec, P1 and P2 all with negative sign
DFIM V2
P1, Q1 Stator Rotor P2, Q2 V1

ωs, I1, V1 sωs, I2, V2
I2
Tem I1 ψ2
ψ1
Pmec
Power flow direction Phasor diagram

Mode (b) Sub-synchronous speed, Pmec and P1 are negative, P2 is positive
Figure 2.27 Generating modes of the DFIM with Q1 > 0

numeric calculations. To do that, base quantities are chosen. For instance, if the
stator rms values of voltages, currents and angular speed are chosen as basic base
values, then
Base voltage : Vbase ¼ jV 1 jrated
Base current : Ibase ¼ jI 1 jrated (2.82)
Base angular speed : wbase ¼ ws
where
D
● jV1 jrated ¼ line to neutral rated rms stator voltage (V);
D
● jI1 jrated ¼ rated rms stator current (A);
D
● ws ¼ rated stator angular speed (rad/s).
From (2.82), the other base values can be deduced such as
g
Vbase
Base impedance : Zbase ¼
Ibase
Vbase
Base flux : Ybase ¼
wbase
Ybase (2.83)
Base inductance : Lbase ¼
Ibase
Base power : Sbase ¼ 3Vbase Ibase
p Sbase
Base torque : Tbase ¼
2 wbase
From (2.82) and (2.83) and dividing the magnitude of each of stator parameters
and rotor parameters referred to the stator by the corresponding base value, the
relations (2.68), (2.70), (2.72), (2.73), (2.79) and (2.81) can be derived and are
g
rewritten in per unit as follows:
V 1pu ¼ R1pu I 1pu þ jY 1pu

V 2pu ¼ R2pu I 2pu þ jY 2pu
Y1pu ¼ L1pu I 1pu þ Lmpu I 2pu
Y2pu ¼ L2pu I 2pu þ Lmpu I 1pu
P1pu ¼ jV 1pu jjI 1pu j cos ðav ai Þ (2.84)
Q1pu ¼ jV 1pu jjI 1pu j sin ðav ai Þ
P2pu ¼ jV 2pu jjI 2pu j cos ðbv bi Þ
Q2pu ¼ jV 2pu jjI 2pu j sin ðbv bi Þ
Tempu ¼ ImfY2pu :I 2pu g
Wind energy 55
Dynamic model in d-q frame

Self-inductances of the stator and rotor windings are constant because the air gap
is smooth, whereas the mutual inductance between them varies with the rotor
displacement relative to the stator. This variation of the stator to rotor mutual
inductance makes the induction machine analysis complicated in terms of real
variables as the voltage equations become non-linear. In order to eliminate this
difficulty, and thus, facilitate analysis, a change of variables can be devised for
stator and rotor variables. This gives a fictitious magnetically coupled two-phase
machine in which the rotor circuits are not only made stationary but are also
aligned with the respective stator windings. In this way, all the inductances
become constant. These orthogonally placed balanced windings are known as the
d-q windings that may be considered fixed or moving with respect to the stator
[7–9].
A d-q frame of reference is chosen to model the DFIM. The model of the IM is
based on the two-axis representation commonly known as the ‘Park model’. A
synchronously rotating d-q frame of reference is used with the direct axis d-axis
oriented along the stator flux position. In this way, decoupled control between the
electrical torque and the rotor excitation current is obtained. The frame of reference
is rotating with the same speed as the stator voltage.
Referring to Figure 2.28, the transformation of stator state space variables
‘abc’ to fictitious ‘dqo’ variables can be obtained with the invariance of power
as the necessary criterion. The equivalent two-phase windings are assumed to
have H(3/2) times as many turns per phase as the three-phase winding. The ficti-
tious stator dqo variables are obtained from the stator variables abc through a
transform defined as
2 3
2p 2p
2 3 6 cos qs cos qs
3
cos qs þ
3 72 3
fd1 6 7 fa1
rffiffiffi6 7
6 7
6 fq1 7 ¼ 26
6 2p 2p 776 7
4 5 6 sin qs sin qs sin qs þ 764 fb1 7
5
36 3 3 7
6 7
fo1 6 7 fc1
4 1 1 1 5
pffiffiffi pffiffiffi pffiffiffi
2 2 2
(2.85)
where f is any of stator variables such as voltage, current and flux, and the com-
ponents along the axis denoted by ‘o’ in dqo frame of reference can be ignored in
balanced conditions (vo and io will be zero). On the other hand, the same trans-
formation is applied to the rotor quantities by replacing qs by (qs – qr). Therefore, a
set of equations, (2.86), results from applying this transformation to obtain the
Moving Moving
q-axis d-axis
Phase ‘a’
rotor axis
θs
θr Phase ‘a’
stator axis
Figure 2.28 Angles of reference axes
g
stator and rotor voltages in the d-q axes with using motor convention. In addition,
the corresponding electric equivalent circuit can be drawn as shown in Figure 2.29.
djd1
vd1 ¼ R1 id1 þ wjs q1
dt
djq1
vq1 ¼ R1 iq1 wjs d1
dt
djd2
vd2 ¼ R2 id2 þ swjs q2
dt
djq2
vq2 ¼ R2 iq2 swjs d2
dt
3 Lm p
Tem ¼ iq1 id2 id1 iq2 (2.86)
2 2
3 3
P1 ¼ Re v1 i1 ¼ vd1 id1 þ vq1 iq1
2 2
3 3
Q1 ¼ Im v1 i1 ¼ vq1 id1 vd1 iq1
2 2
3 3
P2 ¼ Re v2 i2 ¼ vd2 id2 þ vq2 iq2
2 2
3 3
Q2 ¼ Im v2 i2 ¼ vq2 id2 vd2 iq2
2 2
2.4.1.3 WT with DFIG control system

The basic configuration of a WT with DFIG connected to a three-phase grid as
explained in Section 2.4.1.2 is shown in Figure 2.30. The system incorporates a
power electronic interface (converter) between the generator rotor and the grid (or
the load). So, the electrical power delivered by the generator to the grid can be
dynamically controlled. The WT mechanical power, Pw, is converted into electric
power, Pg, by the DFIG. The rotor active power, Pr, can be imparted to the grid
Wind energy 57
ωsψ q1 sωsψq2
Id1 R1 – + L1 L2 + – R2 Id2
+ +
dψd1 dψd2
Vd1 Lm Vd2
dt dt –
– –
ωsψd1 sωsψd2
– L1 L2
Iq1 R1 + – + R2 Iq2
+ +
dψq1 dψq2
Vq1 dt Lm Vd2
dt
–
– –
Figure 2.29 d-q equivalent circuit of the DFIM
Rotor
blades
Pw Ps Pg Utility grid or
DFIG load
Gearbox
vw ωg/ωw
T w, ω w T g, ω g Transformer
Pr RSC GSC
Wind turbine control system
Figure 2.30 Basic configuration of WT with DFIG
(Pr < 0 at hyper-synchronous speed) or taken from the grid (Pr > 0 at sub-
synchronous speed) or equals zero at stationary operation modes. Therefore, a WT
control system is needed to regulate the rotor power at different wind speeds, which in
turn affects the power Pg. It is to be noted that the converter can directly be connected
between the stator circuit of the generator, and the grid has to be designed for the rated
power of the WT, which makes it more expensive than that connected between the
rotor and the grid where the converter rating is typically 25% of the total system
power as well as the speed range of the generator is 33% around the synchronous
speed.
To obtain a proper controlled WT system, the mathematical model of the WT

system must be formed. It usually contains the models of elements representing the
basic functional components: wind and WT aerodynamic, WT drive train, DFIG,
electric grid and WT control system. These elements are described briefly as
follows:
● Wind and WT: The wind facing the swept area of WT blades is represented by
its velocity, vw, which is expressed as the summation of four components
defined by
vw ¼ vwðavÞ þ vwðgstÞ þ vwðrmpÞ þ vwðnosÞ (2.87)
where all components are in m/s and

D
vwðavÞ ¼ average wind speed;
D
vwðgstÞ ¼ gust wind component;
D
vwðrmpÞ ¼ ramp wind component;
D
vwðnosÞ ¼ noise wind component.
Gust, ramp and noise components are the turbulent wind velocity compo-
nents. In addition, the first two components are deterministic turbulences,
whereas the noise wind component is a stochastic part to predict turbulence
occurrence and correlation. The WT model can be mathematically expressed
by (2.26) and (2.27) where the electric power is proportional to the cube of
wind velocity.
● WT drive train: It can be considered as the main part of the WT dynamic
structure that contributes to the interaction with the grid. WT-drive train
includes the inertia of the WT, generator and gearbox connecting the two
rotating shafts. When modelling the drive train, it is common practice to
neglect the dynamics of mechanical parts as their responses are considerably
slow in comparison to the fast-electrical dynamics on especially for machines
with large inertia. So, the two-mass rotational system may be expressed by a
single equation of motion given by
dwg Pw Pg
JWG ¼ (2.88)
dt wg
where
D
JWG ¼ WT mechanical inertia þ generator mechanical inertia;
D
Pw ¼ mechanical power produced by turbine;
D
Pg ¼ electric power delivered to the grid (or the load);
D
wg ¼ angular speed of the rotor.
Integrating (2.88) yields
ð w2 ð t2

JWG wdw ¼ Pw Pg dt
w1 t1
Wind energy 59
Hence,
ð t2

½J WG w22 w21 ¼ Pw Pg dt (2.89)
t1
It is seen from (2.89) that the instantaneous difference between the mechanical
wind power and electrical power changes the rotor speed. This method of speed
control has an advantage of not involving any mechanical action and is smooth in
operation. On the other hand, it implies a disadvantage of continuous fluctuation of
the power output to the grid due to the continuous control of the rotor speed, which
is undesirable for the power system.
● DFIG: Its d-q dynamic model has been explained in Section 2.4.1.2 and
described by (2.86). The active and reactive power of the DFIG can be
controlled by the q-axis and d-axis components of the rotor current,
respectively.
● Electric grid: At the PCC with the grid, the parameters such as current, voltage
and power can be considered as a mirror that reflects the response of the grid to
the operation of WT at normal conditions and in contingencies as well.
● Control system: The WTs require control systems. It is desirable to reduce
the drive train loads and protect the generator and associated equipment
from overloading by limiting the turbine power to the rated value up to the
furling speed. At gust speeds, the machine must be stalled. At low and
moderate wind speeds, the aim should be to capture power as efficiently as
possible. WTs can have two types of control: one for controlling mechanical
dynamics such as pitch angle control, stall control and yaw control, and the
other for controlling electrical dynamics using power electronics to control
generator speed and power. Electrical dynamics are much faster than
mechanical dynamics, and therefore, it is possible to control the machine in
a cascade structure.
However, the WT has four different speed bands for wind speed as described in
Section 2.3.2. So, it requires different speed strategies. A typical example of using
power electronics to control the active and reactive power of WTs is known as
machine vector control as schematically depicted in Figure 2.30. It shows the
power generation scheme using two bidirectional converters (RSC and GSC) and a
DC link capacitor. The capacitor is charged to a small value initially and the control
is initiated by injecting a small value of current in both the d- and q-axes of the
rotor current. While the d-axis current establishes the stator flux, the q-axis current
has a direct control of the DC link capacitor voltage. When the capacitor voltage
reaches a predetermined reference value, both converters are controlled in such a
way that the net power flowing to the capacitor is zero. At this point, the load
switch along with a filter is turned ON. At steady state, the system operates exactly
as a conventional slip power recovery drive except that the combined reactive
power demand of the machine and load is supplied by the two converters instead of
the grid.
Tw Tg i
vm Mechanical
WT rotor Generator Electric grid
ωw shaft ωg v
model model model
model
ωw Stator and rotor
Rotor
ωg voltage active power
WT control system
Stator active Rotor active

power reference power reference
Figure 2.31 Dynamic model of WT-DFIG system connected to the grid
According to what has been explained in Section 2.3.1, the dynamic model of a
WT system with DFIG can be established in compliance with the block diagram
shown in Figure 2.31.
2.4.2 Permanent magnet synchronous machines

The use of PM machines has become an attractive option for a wide variety of
industrial applications because of the benefits that can be gained such as follows:
(i) the excitation is provided by PMs, and so, brushes and slip rings are eliminated
resulting in simplification of construction and maintenance; (ii) because of the use
of PMs, the excitation is current-free and lossless leading to a substantial increase
in efficiency; (iii) these are higher in power density (power generated to weight
ratio) as compared to other electric machines; (iv) these have better dynamic per-
formance than machines with electromagnetic excitation (higher magnetic flux
density in the air gap); (v) it’s enabled machine speed control because of its unique
operation mode; (vi) there is adequacy to carry out precise speed and position
control on a PM machine as the armature field and rotor are always synchronized,
in addition to the facility of controlling the armature field precisely [10].
However, the synchronous nature of PM machines may cause problems during
start-up, synchronization and voltage regulation. The synchronous operation causes
a stiff performance in case of short-circuit occurrence and if the wind speed is
unsteady. In addition, the magnetic materials are sensitive to temperature and may
lose the magnetic qualities at high temperature, for instance, during a fault. Also,
application of PM machines is limited by the fixed excitation from PMs; that is, the
magnet flux is fixed. In wind energy application, it is required to operate the PM
machine over a wide range of speeds at constant power, which necessitates the use
of control laws for flux weakening [11,12]. There are different topologies of PM
machines presented in the literature [13]. The most common types are the radial
flux machine, the axial flux machine and the transversal flux machine.
2.4.2.1 Structure, equivalent circuit and phasor diagram

PM machines have a wound stator of a typical three-phase AC machine and a rotor.
The rotor is provided with a PM pole system and may have salient poles or may be
Wind energy 61
cylindrical depending on the magnet mounting style. Salient poles are more com-
mon in low-speed machines and may be the most useful version for wind generator
application. It is to be noted that the magnetic permanence along the direct axis
(d-axis) of the magnets in a PM machine is extremely low as compared to the other
types of AC machines due to the existence of PMs.
Equivalent circuit of a PM machine is shown in Figure 2.32. It is similar to that
of a conventional synchronous machine except that the internal induced voltage,
Eo, is due to magnets rotation rather than the rotation of the field winding; that is,
the source of the flux is different, as well as the direct-axis reactance, Xd, is much
smaller than the quadrature-axis reactance, Xq, because of the PMs. If the machine
is without damping winding and the zero-sequence current in the stator does not
exist (vo ¼ 0), the PM machine can mathematically be represented by a set of
equations as follows:
g
djd
v d ¼ ra i d þ wjq
dt (2.90)
djq
vq ¼ ra iq þ þ wjd þ E
dt
where Yq ¼ Lq iq ; Yd ¼ Ld id and E is the internal induced voltage due to the
magnet generated flux linking the stator windings. The power equation is
EV Xd X q
P¼ sin d þ V 2 sin ð2dÞ (2.91)
Xd Xd X q
where d is the angle between E and V.
ωψd ra ωψq Ld
Lq ra
– + + –
+ iq + id +
Eo Vq Vd
– – –
Figure 2.32 Equivalent circuit of a PM machine
jIdXd
jIqXq
Id
I
δ
Iq E
Figure 2.33 Steady-state phasor diagram of a PM machine

1
Bo
2
Bm
Rated d-axis
I
armature reaction H
Ho Hm Hc
Reversible Irreversible
range range
Figure 2.34 Demagnetization curve of a PM
Rotor
blades
PM generator Utility grid
AC/DC DC/AC
Step-up
transformer
ESS
DC bus
Figure 2.35 PM generator connected to utility power grid
The steady-state operation of a PM machine can be characterized by utilizing

the phasor diagram shown in Figure 2.33. In addition, the characteristics of the
PMs, given in Figure 2.34, help to describe the PM machine. The demagnetizing
curve of the magnet can be divided into three ranges: no-load, rated-load and
excessive load bounded by lines #1, #2 and #3, respectively. So, if the demagne-
tizing current is near or beyond line #3, the PM is in danger and may be damaged.
In transient studies, it is common to neglect the effect of speed variations, and
then, the terms wYq and wYd in (2.90) can be eliminated.
Wind energy 63
2.4.2.2 PM machines connected to power grid

PM machines can be used as variable speed generators in variable speed wind
energy systems. This means that the drive, supplying mechanical power to the
generator, has a variable speed. So, the connection of PM generator to a local
power grid can be implemented using full AC/DC and DC/AC power converters as
shown in Figure 2.35. In this case, the variable speed PM generator produces power
at variable voltage with variable frequency. Therefore, the generated power cannot
directly be injected into the power grid. It needs, first, to match its frequency and
voltage to those of the grid. For this reason, the variable AC frequency voltage is
rectified by the AC/DC rectifier. The DC power may be used to charge an energy
storage system through a DC bus, and meanwhile, it is converted to AC power at
grid frequency and voltage using DC/AC inverter to be imparted to the grid through
a step-up transformer.
References
[1] Masters G. M. Renewable and efficient electric power systems. New York:
Wiley-Interscience, John Wiley & Sons, Inc.; 2004.
[2] Farret F. A. and Simoes M. G. Integration of alternative sources of energy.
New York: IEEE Press, Wiley-Interscience, John Wiley & Sons, Inc.; 2006.
[3] Da Rosa A. V. Fundamental of renewable energy processes. 2nd edn.
Burlington, USA: Elsevier Academic Press; 2009.
[4] Ackermann T. Wind power in power systems. NY: John Wiley & Sons. Inc.;
2005.
[5] Polinder H., de Hann S. W. H., Dubois M. R. and Slootweg J. G. ‘Basic
operation principles and electrical conversion systems of wind turbines’.
EPE Journal: European Power Electronics and Drives. 2005;15(4):43–50.
[6] Abad G., Rodriguez J. M. A., Marroyo L. and Iwanski G. Doubly fed
induction machine. New York: IEEE Press, John Wiley & Sons, Inc.; 2011.
[7] Huan-ping L. I. and Jin-ming Y. (eds.). ‘The performance of large scale wind
farm connected to external power grid’. Power Electronics Systems and
Applications, PESA 2009, 3rd Int. Conf. on. Hong Kong, May 2009, pp. 1–5.
[8] Smajo J. ‘Wind turbine system with doubly-fed induction generator
and rotor power feedback control’. WSEAS Trans. on Systems. 2006;12
(5):2860–7.
[9] Morren J. and de Haan S. W. H. ‘Ridethrough of wind turbines with doubly-
fed induction generator during a voltage dip’. IEEE Trans. on Energy
Conversion. 2005;20(2):435–41.
[10] Gieras J. F. and Wing M. Permanent magnet motor technology – Design and
applications. 2nd edn. New York: Marcel Dekker, Inc.; 2002.
[11] Xu L., Ye L., Zhen L. and El-Antably A. ‘A new design concept of perma-
nent magnet machine for flux weakening operation’. IEEE Trans. on
Industry Applications. 1995;31(2):373–8.
[12] Amara Y., Lucidarme J., Gabsi M., Lecrivain M., Ben-Ahmed A. H. and
Akemakou A. D. ‘A new topology of hybrid synchronous machine’. IEEE
Trans. on Industry Applications. 2001;37(5):1273–81.
[13] Luo X. and Lipo T. A. ‘A synchronous/permanent magnet hybrid AC
machine’. IEEE Trans. on Energy Conversion. 2000;15(2):203–10.
Chapter 3
Solar energy
3.1 Introduction
Solar is the Latin word of the sun and solar energy is radiant light and heat from the
sun that is a powerful source of energy. Many applications may use solar energy
such as heating, cooling, ventilation, illumination, transport, cooking, water heat-
ing, water treatment, fuel production, electricity production, energy storage systems
(ESSs) and buildings [1,2]. It is found, statistically, that the amount of solar energy
from the sun falling on the earth in one hour is more than that used by everyone in
the world in one year [3]. So, it is an important source of renewable energy, and
solar technologies are broadly characterized as either active solar or passive solar
depending on how they capture and distribute solar energy or convert it into solar
power [4].
Active solar technologies are the most commonly used. The sun delivers its
energy in two main forms: light and heat. Correspondingly, there are two main types
of solar power systems, namely, solar photovoltaic (PV) and solar thermal systems
(Figure 3.1). In case of PV, when the sun shines onto a solar panel, photons from the
sunlight are absorbed by the cells in the panel. It creates an electric field across the
layers and causes electricity to flow. Solar thermal technology uses the sun’s energy,
rather than fossil fuels, to generate low-cost, environmentally friendly thermal
energy. This energy is used to heat water or other fluids and can also power solar
cooling systems. It comes in many shapes and sizes and is used in a wide variety of
commercial applications. Commonly seen types of solar thermal systems with lim-
ited size include domestic hot water systems, swimming pool-/hot tub-heating sys-
tems and radiant space heating systems. Large-scale solar thermal systems,
concentrating solar power (CSP) technology, can be used for electricity production.
CSP is used primarily in very large plants and its technology uses mirrors to
reflect and concentrate sunlight onto receivers that collect solar energy and convert
it into heat to produce steam. The steam drives a turbine and generates power in the
same way as conventional power plants. Other concepts are being explored and not
all future CSP plants will necessarily use a steam cycle. Therefore, businesses and
industry use these technologies to diversify their energy sources, improve effi-
ciency and save money. In addition, developers and utilities are also using these
technologies to produce electricity on a massive scale to power cities and small
towns [3].
Su
nl
at
ig
he
ht
n
Su
Convert sun energy into
Convert sunlight directly

into electrical energy
thermal energy
to heat a fluid
Thermal solar system PV solar system
Figure 3.1 Thermal and PV solar systems
However, solar thermal systems differ from solar PV systems which generate
electricity directly rather than heat.
Passive solar technologies include orienting a building to the sun, selecting
materials with favourable thermal mass or light-dispersing properties and designing
spaces that naturally circulate air.
Consequently, active solar technologies increase the supply of energy and are
considered as supply-side technologies, whereas passive solar technologies reduce
the need for alternate sources and are generally considered as demand-side
technologies.
To study and analyse a utility grid supported by such active solar energy, it
entails understanding how the solar energy is converted into electricity especially
when using PV or CSP technologies. Characteristics of both solar PV and CSP
technologies are explained in the following sections.
3.2 Solar PV technology

3.2.1 Basics
PV materials and devices (also known as solar PV) convert sunlight directly into
electrical energy. A single PV device is known as a PV cell or solar cell. PV gets its
name from the process of converting light (photons) to electricity (voltage), which
Solar energy 67
n-type
Barrier N-side
P-side
p-type
Figure 3.2 Effect of electric field in a PV cell
is called PV effect [5,6]. It is a non-mechanical device usually made from silicon

alloys. Sunlight is composed of photons, or particles of solar energy. These photons
contain various amounts of energy corresponding to the different wavelengths of
the solar spectrum. When photons strike a PV cell, they may be reflected, pass right
through or be absorbed. Only the absorbed photons provide energy to generate
electricity.
Silicon (Si) crystals are all electrically neutral. In n-type Si, extra electrons are
balanced out by the extra protons in the phosphorous. In p-type Si, missing elec-
trons (holes) are balanced out by the missing protons in the boron. However, when
the holes and electrons mix at the junction between n-type and p-type silicon, that
neutrality is disrupted. Do all the free electrons fill all the free holes? No. If they
did, then the whole arrangement would not be very useful. Right at the junction,
however, they do mix and form a barrier, making it harder and harder for electrons
on the N side to cross to the P side. Eventually, equilibrium is reached, and an
electric field separating the two sides (A and B) exists as schematically shown in
Figure 3.2 [7].
This electric field acts as a diode, allowing (and even pushing) electrons to
flow from the P side to the N side, but not the other way around. It’s like a hill –
electrons can easily go down the hill (to the N side) but cannot climb it (to the
P side). So, the electric field is acting as a diode in which electrons can only move
in one direction.
When light, in the form of photons, hits the solar cell, its energy frees electron–
hole pairs. Each photon with enough energy will normally free exactly one electron
and result in a free hole as well. If this happens close enough to the electric field, or
if free electron and free hole happen to wander into its range of influence, the field
will send the electron to the N side and the hole to the P side. The P side accu-
mulates holes and the N side accumulates electrons, which creates a voltage that
can be used to deliver current to a load. This causes further disruption of electrical
neutrality. If electrical contacts are attached to the top and bottom of the cell, an
external current path is provided and electrons will flow out of the N side into the
path and back to their original side (the P side) through the load to unite with holes
that the electric field sent there. The electron flow provides the current, and the
cell’s electric field causes a voltage (Figure 3.3). With both current and voltage,
power, which is the product of the two, is obtained.
So, solar cells, like batteries, generate direct current (DC), which is generally
used for small loads (electronic equipment). When DC from PV cells is used for
commercial applications or sold to electric utilities using the electric grid, it must
be converted to alternating current (AC) using inverters, solid-state devices that
convert DC power to AC [8].
An individual PV cell is usually small, typically producing about 1–2 W of
power [9]. Modules or panels can be formed as larger units by connecting PV cells
together in chains to boost the power output of PV cells. They can be used indi-
vidually, or several can be connected to form arrays. One or more arrays are then
connected to the electrical grid as part of a complete PV system. Because of this
modular structure, PV systems can be built to meet almost any electric power need,
small or large. So, PV has evolved from small-scale applications towards becoming
a mainstream electricity source with efficiency varying from 3% to 31% depending
on the technology, the light spectrum, temperature, design and material of the
solar cell.
With a goal of improving the efficiency and achieving cost-effective alter-
native, there are variety of technologies using different semiconductor materials in
solar cells such as [9] those given below:
1. Crystalline silicon. It is the most common material used in solar cells, repre-
senting approximately 90% of the modules sold today. Crystalline cells are
made from ultra-pure silicon raw material such as those used in semiconductor
chips. They use silicon wafers that are typically 150–200 mm (one fifth of a
millimetre) thick. Typical mono-crystalline silicon and poly-crystalline silicon
modules are shown in Figure 3.4(a) and (b), respectively.
2. Thin-film PVs. The solar cell is made by depositing one or more thin layers of
semiconductor material on a supporting material such as glass, plastic or metal.
Electrical contacts
Photons
Accumulated negative charge e−
−
n-type
N-side
External load
Barrier V
p-type
P-side
+
Accumulated positive charge I
Bottom contact
Figure 3.3 Operation of a PV cell

Solar energy 69
The two main types of thin-film PV semiconductors commonly used, so far,

are cadmium telluride (CdTe) and copper indium gallium diselenide (CIGS)
(Figure 3.4(c) and (d)). Both materials can be deposited directly onto either the
front or the back of the module surface. This technology, after silicon, enables
low-cost manufacturing processes.
3. Organic PVs. The solar cell is composed of carbon-rich polymers and can be
tailored to enhance a specific function of the cell, e.g. sensitivity to a certain
type of light. This technology has the theoretical potential to provide electricity
at a lower cost than silicon or thin-film technologies, but it is less efficient and
has shorter lifetimes.
4. Concentration PV. It focuses sunlight onto a solar cell using a mirror or lens.
Thus, less PV material is required because of focusing sunlight onto a small
area as well as PV materials become more efficient at energy conversion as the
light becomes more concentrated. So, the highest overall efficiencies are
obtained with concentration PV cells and modules. However, more expensive
materials, manufacturing techniques and tracking are required. A round
figure about conversion efficiency for different PV technologies is given in
Table 3.1 [5].
(a) (b) (c) (d)
Figure 3.4 Common PV module technologies [6]: (a) mono-crystalline silicon,

(b) poly-crystalline silicon, (c) CdTe thin film and (d) CIGS thin film
Table 3.1 Conversion efficiencies of different module technologies
Module technology Module efficiency (%)

Mono-crystalline silicon 12.5–15
Poly-crystalline silicon 11–14
CdTe 9–12
CIGS 10–13
Amorphous silicon (a-si) 5–7
In addition, as crystalline silicon is the most common used in PV cells, dif-

ferent attempts can be applied in order to reduce the cost of manufacturing.
Single-crystal silicon is not the only material used in PV cells. Poly-crystalline
silicon is also used to cut manufacturing costs, although resulting cells are not as
efficient as the cells resulting from single-crystal silicon. Amorphous silicon, which
has no crystalline structure, is also used, again to reduce production costs. Other
materials used include gallium arsenide, copper indium diselenide and CdTe. As
different materials have different band gaps, they seem to be ‘tuned’ to different
wavelengths, or photons of different energies. One-way to improve the efficiency is
to use two or more layers of different materials with different band gaps. The
higher band-gap material is on the surface, absorbing high-energy photons while
allowing lower-energy photons to be absorbed by the lower band-gap material
beneath. This technique can result in much higher efficiencies. Such cells, called
multi-junction cells, can have more than one electric field.
There are a few more steps before using the PV cell. Silicon happens to be a
very shiny material, which means that it is very reflective. Photons that are
reflected cannot be used by the cell. For that reason, an antireflective coating is
applied to the top of the cell to reduce reflection losses to less than 5%. The final
step is the glass cover plate that protects the cell from the elements (Figure 3.5).
PV modules are made by connecting several cells (usually 36) in series and
parallel to achieve useful levels of voltage and current and putting them in a sturdy
frame complete with a glass cover and positive and negative terminals on the back.
The series-parallel connection of cells allows the design of solar panels with high
Sunlight
Cover glass n-type Si
Antireflective coating p-type Si
Contact grid Back contact
Figure 3.5 Basic structure of a generic silicon PV cell

Solar energy 71
currents and voltages in the range of kilovolts. To connect a PV system to the grid,
it is necessary to include power electronic equipment, energy storage and mon-
itoring systems.
Power electronic circuits may be used for smoothing the DC output of PV
system or directly supplying a DC bus to feed an ESS and DC loads (if any). In
addition, inverters are used to convert the DC current generated by the PV system
into AC electricity, which may be connected to the utility grid through a step-up
transformer or feeding any local AC loads (Figure 3.6). PV systems have either one
inverter that converts the electricity generated by all the modules or micro-inverters
that are attached to each individual module. A single inverter is generally less
expensive and can be more easily cooled and serviced when needed. The micro-
inverter allows for independent operation of each module, which is useful if some
modules might be shaded. For instance, it is expected that inverters will need to be
advanced inverters, or ‘smart inverters’, and allow for two-way communication
between the inverter and the electrical utility. This can help balance supply and
demand either automatically or via remote communication with utility operators.
Allowing utilities to have this insight into (and possible control of) supply and
demand allows them to reduce costs, ensure grid stability and reduce the likelihood
of power outages.
The ESS, for example, batteries, allows storing the solar PV energy to be used
at night or when weather elements keep sunlight from reaching PV panels. Such
storage system can be used not only in homes but also for playing a crucial
important role for utilities as it will help modernize and stabilize the electric grid.
A typical example of application of PV systems to buildings is shown in
Figure 3.7. PV modules can be mounted on the buildings or on the ground if the
land is not a constraint. For buildings, commonly, PV modules are mounted either
on the roof or displacing another building component, e.g. window glass or roof/
wall cladding. The building has two parallel power supplies, one from the PV
system and the other from the power grid. The combined power supply feeds all the
loads connected to the main AC distribution board. The ratio of PV system supply
to power grid supply varies, depending on the size of the PV modules. Whenever
the PV supply exceeds the building’s demand, excess electricity will be exported
PV array AC bus Utility grid

Local AC loads
Inverter
DC/DC DC/AC
Step-up
Local DC loads transformer
DC bus
ESS
Figure 3.6 PV system connected to utility grid

PV modules
Data
communication
ESS
Inverter
AC distribution
board
Substation
Smart meter
Utility power grid
Figure 3.7 Combined PV-power grid powering a house
into the grid. When there is no sunlight to generate PV electricity at night, the ESS
will supply the building’s demand if it has enough energy to do that. Otherwise,
ESS may share the power grid or, totally, the power grid supplies all demand. Of
course, this operation needs monitoring and management systems that can be
implemented using data communication channels and smart meters.
It is to be noted that PV arrays must be mounted on a stable, durable structure
that can support the array and withstand wind, rain, hail and corrosion over dec-
ades. These structures tilt the PV array at a fixed angle determined by the local
latitude, orientation of the structure and electrical load requirements. To obtain the
highest annual energy output, modules in the northern hemisphere are pointed due
south and inclined at an angle equal to the local latitude. Rack mounting is cur-
rently the most common method because it is robust, versatile and easy to construct
and install. More sophisticated and less expensive methods continue to be devel-
oped [5].
For PV arrays mounted on the ground, tracking mechanisms automatically
move panels to follow the sun across the sky, which provides more energy and high
returns on investment. One-axis trackers are typically designed to track the sun
from east to west. Two-axis trackers allow for modules to remain pointed directly
at the sun throughout the day. Naturally, tracking involves more upfront costs and
sophisticated systems are more expensive and require more maintenance. As sys-
tems have improved, the cost-benefit analysis increasingly favours tracking for
ground-mounted systems.
Solar energy 73
However, solar systems have some advantages and disadvantages as follows:

Advantages of solar systems:
● Conversion from sunlight to electricity is direct, so that bulky mechanical
generator systems are unnecessary.
● PV arrays can be installed quickly and in any size, required or allowed.
● The environmental impact is minimal as the PV system is inexhaustible, does
not pollute the atmosphere and has great reliability.
● There is emission-free operation and no fossil fuel consumption.
● Solar systems have excellent modularity (nearly any building-related load can
be matched well by multiple units).
Disadvantages of solar energy:
● The amount of sunlight that arrives at the earth’s surface is not constant. It
depends on location, time of day, time of year and weather conditions.
● Because the sun does not deliver that much energy to any one place at any one
time, a large surface area is required to collect the energy at a useful rate.
● The price of delivered power exceeds other distributed generation resources.
● Temporal match of power produced to load is imperfect; batteries or other
systems are often needed.
● The scheme may also fail to satisfy the emerging energy needs of the area and
constrain industrial development and/or expansion.
3.2.2 Energy loss in a solar cell

The performance of a PV array is dependent upon sunlight. Climate conditions (e.g.
clouds and fog) have a significant effect on the amount of solar energy received by
a PV array and, in turn, its performance.
The amount of solar energy reaching the ground, as well as its spectral dis-
tribution, depends very much on how much atmosphere it has had to pass through to
get there. Recall that the length of the path taken by the sun’s rays through the
atmosphere to reach a spot on the ground, divided by the path length corresponding
to the sun directly overhead, is called the air mass ratio, AM. Thus, an air mass ratio
of one (designated ‘AM1’) means that the sun is directly overhead. By convention,
AM0 means no atmosphere; that is, it is the extraterrestrial solar spectrum. For most
PV work, an air mass ratio of 1.5, corresponding to the sun being 42 degrees above
the horizon, is assumed to be the standard. For an AM1.5 spectrum, 2% of the
incoming solar energy is in the ultraviolet portion of the spectrum, 54% is in the
visible and 44% is in the infrared. The solar spectrum at AM1.5 for a silicon PV cell,
based on ERD/NASA (1977), is depicted in Figure 3.8(a). It illustrates that photons
with wavelength longer than 1.11 mm do not have enough energy to excite electrons
(20.2% of the incoming solar energy); those with shorter wavelengths cannot use all
their energy, which counts for another 30.2% unavailable to a silicon PV cell.
Visible light is only part of the electromagnetic spectrum. Electromagnetic
radiation is not monochromatic; it is made up of a range of different wavelengths,
UV energy = 2% of the incoming

solar energy
Visible energy = 54% of the
incoming solar energy
IR energy = 44% of the incoming

1,400
solar energy
Unavailable energy > Eg
1,200 30.2%
Radiant power (W/m2 μm)
1,000
800
Energy available
49.6%
600
Unavailable energy < Eg
20.2%
400
Band-gap wavelength
1.11 μm
200
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
Wavelength (μm)
(a)
Silicon
9
Photons with more than Photons with not
8 enough energy enough energy
Photon energy (eV)
7
Photon energy
6
5
Lost energy > Eg
4
3
Usable energy Lost energy < Eg
2
Eg = 1.12
1
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
(b) Wavelength (μm)

1.11
Figure 3.8 Solar spectrum at AM1.5 for a silicon PV cell: (a) radiant power versus
wavelength indicating the percentage of available and unavailable
photon energies; (b) solar spectrum illustrating usable and lost energies
Solar energy 75
and therefore energy levels. Since the light that hits the cell has photons of a wide
range of energies, it turns out that some of them would not have enough energy to
form an electron–hole pair. They will simply pass through the cell as if it
were transparent. Still other photons have too much energy. Only a certain amount
of energy, called ‘band-gap energy’, Eg, measured in electron volts (eV ¼
1.6 1019 joule) and defined by the cell material (about 1.11 eV for crystalline
silicon), is required to knock an electron loose. If a photon has more energy than the
required amount, then the extra energy is lost (unless a photon has twice the
required energy and can create more than one electron–hole pair, but this effect is
not significant). These two effects alone account for the loss of around 70% of the
radiation energy incident on the cell. Figure 3.8(b) shows that for a silicon PV cell,
photons with wavelength greater than 1.11 mm have energy less than the 1.12-eV
band-gap energy needed to excite an electron. None of those photons create hole–
electron pairs capable of carrying current, so all their energy is wasted. It just heats
the cell. However, photons with wavelengths shorter than 1.11 mm have more than
enough energy to excite an electron. As one photon can excite only one electron,
any extra energy above the 1.12 eV needed is also dissipated as waste heat in the
cell [10].
A material with a low band gap could not be chosen, so more of the photons
can be used. Unfortunately, the band gap also determines the strength (voltage) of
the electric field, and if it is too low, then extra current by absorbing more photons
is produced but at a small voltage which is not useful as power is product of voltage
and current. The optimal band gap, balancing these two effects, is around 1.4 eV for
a cell made from a single material.
There are other losses as well. Electrons must flow from one side of the cell to
the other through an external circuit. The bottom can be covered with a metal,
allowing for good conduction, but if the top is completely covered, then photons
cannot get through the opaque conductor and all the current is lost (in some cells,
transparent conductors are used on the top surface, but not in all). If the contacts are
only placed at the sides of the cell, then the electrons must travel an extremely long
distance (for an electron) to reach the contacts. It should be noted that silicon is a
semiconductor; that is, it is not nearly as good as a metal for transporting current.
Its internal resistance (called series resistance) is high, and high resistance means
high losses. To minimize these losses, the cell is covered by a metallic contact grid
that shortens the distance that electrons must travel while covering only a small part
of the cell surface. Even so, some photons are blocked by the grid, which cannot be
too small or else its own resistance will be too high.
3.2.3 Electricity generation

PV cells are made of a semiconductor material which can be described as a p-n
junction diode. The conventional p-n junction diode has the voltage–current (V–I)
characteristic as shown in Figure 3.9. Applying a voltage Vd across the diode
terminals leads to flowing a forward current Id through the diode from the p side to
the n side; but in reverse direction, only a very small saturation current Is will flow
Id
Id = Is (eqVd/ηkT − 1)
Vd
+ −
+ −
Id
p-side n-side Is
Vd
(a) (b)
Figure 3.9 p-n junction diode: representation and characteristic curve.

(a) p-n junction diode. (b) V–I characteristic curve
(1012 A/cm2). Is is the result of thermally generated carriers with the holes being
swept into the p side and the electrons into the n side. The voltage drop across the
diode in the forward direction is only a few tenths of a volt.
V–I characteristic can mathematically be expressed by Shockley diode equa-
tion as
Id ¼ Is ðeqVd =hkT 1Þ (3.1)

where
D
Id ¼ forward diode current (A);
D
Is ¼ reverse saturation current of the diode (A), typically 100 pA for the
silicon cell;
D
k ¼ Boltzmann constant (1.381 1023 J/K);
D
q ¼ electron charge (1.6021 1019 C);
D
Vd ¼ diode terminals voltage from p side to n side (V);
D
T ¼ absolute temperature (K);
D
h ¼ empirical constant to account for different mechanisms responsible for
moving carriers across the junction (¼ 1 in PV application).
Substituting the values of q, k, h and T at 25 C (¼ 25 þ 273 þ 25 ¼ 298.15 K)
into (3.1) obtains

Id ¼ Is e38:9Vd 1 (3.2)
3.2.4 PV cell equivalent circuit

According to what is explained in Section 3.2.1 and depicted in Figures 3.2 and 3.3,
the equivalent circuit of a PV cell can simply be deduced as a real diode in parallel
Solar energy 77
with a current source driven by absorbed photons and directly proportional to the
solar insolation as shown in Figure 3.10.
Based on two conditions, both the current flowing into the load, I, and
the open-circuit voltage, VOC, across the terminals can be determined. First, when
the terminals are short-circuited, there is no current flowing in the diode, that
is, Id ¼ 0, as the voltage Vd ¼ V ¼ 0 and all the ideal source current, ISC, will
flow through the short-circuited terminals. Consequently, the magnitude of the
ideal current source must be equal to ISC, and then, the current equation for the
equivalent circuit can be written as
I ¼ ISC Id (3.3)
Incorporate (3.2) and (3.3) to get
I ¼ ISC Is ðeqV =hkT 1Þ (3.4)

where V ¼ Vd.
It can be seen from (3.4) that the current I is equal to ISC added to the diode
equation with negative sign. It is also noted that when ISC ¼ 0, i.e. there is no solar
insolation (dark state), the current I is just the negative amount of diode current Id.
The plot of V–I relationship for both cases, with and without insolation, is shown in
Figure 3.11.
+
I
Id
ISC
V
Load
Wires connecting
PV cell model
an external load
Figure 3.10 Equivalent circuit for a PV cell

ISC
With solar insolation
VOC
0 V
Without solar insolation
Figure 3.11 Current–voltage relationships for a PV cell
Second, if the diode terminals are opened, then the current I is equal to zero
and the open-circuit voltage, VOC, can be obtained by solving (3.4) as follows:

kT ISC
VOC ¼ ln þ1 (3.5)
q IS
At 25 C, (3.4) and (3.5) are written as
g
I ¼ ISC IS ðe38:9V
1Þ
ISC (3.6)
VOC ¼ 0:0257 ln þ1
IS
Example 3.1
(a) A p-n junction diode has a reverse saturation current of 1010 A. Find the
voltage drop across the diode when carrying (i) no current and (ii) currents of
1, 5 and 10 A, respectively.
(b) If the model of a 100 cm2 PV cell consists of the diode described in (a) and the
cell produces short-circuit current of 50 mA/cm2, find the open-circuit voltage
at full sunlight, and at 50% sunlight.
Solar energy 79
Solution
(a) Using (3.2), it can be found that

1 Id
Vd ¼ ln þ1
38:9 Is
(i) At no current, Id ¼ 0 and accordingly Vd ¼ 0

(ii) At diode current of 1 A, the voltage across the diode is

1 1
Vd ¼ ln þ 1 ¼ 0:592 V
38:9 1010
Similarly, the voltage across the diode when Id ¼ 5 A is given by

1 5
Vd ¼ ln þ 1 ¼ 0:633 V
38:9 1010
and at Id ¼ 10 A, Vd will be

1 10
Vd ¼ ln þ 1 ¼ 0:651 V
38:9 1010
It is to be noted that, when the diode current changes from 1 to 5 A,

the voltage across its terminals changes just by an amount of 0.041 V.
Also, for current changing from 1 to 10 A, the terminal voltage changes
by 0.059 V. This means that a little change of Vd corresponds to a large
change of diode current, Id. So, Vd is commonly assumed to be 0:6 V
for circuit analysis.
(b) At full sunlight, ISC ¼ 50 103 100 ¼ 5 A
Using (3.6),

5
VOC ¼ 0:0257 ln þ 1 ¼ 0:633 V
1010
At 50% sunlight, ISC ¼ 25 103 100 ¼ 2:5 A as it is directly proportional

to solar intensity. Thus,

2:5
VOC ¼ 0:0257 ln þ 1 ¼ 0:615 V
1010
So, V–I curves for the two cases, full sunlight and 50% sunlight, can be plotted as
follows:
I
Isc = 5 A
Full sunlight
Isc = 2.5 A
50% sunlight
VOC = 0.615 V VOC = 0.633 V
The equivalent circuit of a PV cell can be modified by adding a parallel

resistance, RP, and a series resistance, RS (Figure 3.12). RP can represent the effect
of shading and internal losses or leakage current across the diode (its value ranges
from 200 to 300 W). RS is usually very small (0.05–0.10 W) and might be the
contact resistance and the resistance of the semiconductor itself. However, RS is
reflected directly on the manufacturing quality of the PV cells.
RS
I +
Id IP
ISC
RP
V
Figure 3.12 Modified equivalent circuit of a PV cell

Solar energy 81
As shown in Figure 3.12, the following relations (at 25 C) can be derived.

I ¼ ISC Id IP (3.7)
Using (3.2) and substituting IP ¼ Vd =RP , (3.7) becomes
Vd
I ¼ ISC Is e38:9Vd 1 (3.8)
RP
Vd ¼ V þ IRS (3.9)
Incorporating (3.8) and (3.9) obtains
V þ IRS
38:9ðV þIRS Þ1Þ
I ¼ ISC Is ðe RP (3.10)
The open-circuit voltage can be obtained as

VOC ¼ Vd jI¼0 ¼ ðISC Id ÞRP ¼ ISC Is e38:9Voc 1 RP (3.11)
3.2.5 Solar modules

An individual PV cell produces only 0:6 V. So, it is impractical to find an
application using just a single cell. Instead, a group of PV cells in series, known as
a module, is used, and consequently, a schematic line diagram of a module of n-PV
cells can be drawn as shown in Figure 3.13(a). Assuming that the cells are identical
and have the same characteristic, all cells carry the same current whereas the
ISCn ISC2 ISC1

ISCn ISC2 ISC1 ISC1
ISCn ISC2
Idn RSn Id2 RS2 Id1 RS1
V
RPn IPn I RP2 IP2 I RP1 IP1 I
(a)
ISC ISC ISC
Id Id Id RS RS RS
V
RP IP RP IP RP IP I
(b)
Figure 3.13 Line diagram of a module consisting of n-PV cells: (a) Module of
n-PV cells. (b) Module of similar n-PV cells
corresponding voltages are added. In addition, there are no circulation currents

among the cells, all series resistances are equal and parallel resistances are equal
too. Accordingly, the schematic line diagram can be redrawn as shown in
Figure 3.13(b) [11].
To obtain the equivalent circuit of a module consisting of n-PV cells, the line
diagram in Figure 3.14(a) can be formed as shown in Figure 3.14(b). It contains
n-series current sources having the same current ISC, n-parallel resistances con-
nected in series and n-series resistances connected in series as well. Thus, the
equivalent circuit is a current source carrying a current of ISC connected in parallel
with a diode of voltage nVd and a resistance nRP and all in series with a resistance
nRS as shown in Figure 3.14(b).
From the equivalent circuit, it is noted that
V ¼ nðVd Rs I Þ (3.12)
where Vd is the diode terminal voltage for an individual cell. It can be obtained
from (3.1) as

kT Id
Vd ¼ ln þ1 (3.13)
q Is
ISC ISC ISC

ISC
Id Id RS RS RS
V
RP IP RP IP RP I
(a)
nRS
ISC Id IP I +
ISC
nVd
nRP V
(b)
Figure 3.14 Equivalent circuit of a PV module consisting of n-PV cells.

(a) Equivalent line diagram of n series PV cells. (b) Compact
equivalent circuit for a PV module
Solar energy 83
Thus, (3.12) becomes

nkT Id
V¼ ln þ 1 nRS I (3.14)
q Is
and the current through the diodes can be obtained by Shockley’s equation as

qðV þ nRS I Þ
Id ¼ IS exp 1 (3.15)
nkT
3.2.6 Solar arrays

Arrays are made of some combination of series and parallel modules to increase
both voltage and current, which in turn increases the power. First, as analogous to
the PV cells connected in series, it is required to calculate the voltage and currents
by determining the equivalent circuit for two cells connected in parallel.
In case of connecting m-PV cells in parallel across common terminal, there is
no direct interaction between current sources or individual voltages because they
are considered the same. Therefore, the equivalent source current is m times the
individual source current whereas the voltage across the terminal is the same as the
voltage across the terminals of the individual cells and the equivalent resistance
(series or parallel) is equal to its value for the individual cell divided by the number
of cells m. Thus, the following relations can be written, and the corresponding
compact equivalent circuit is as shown in Figure 3.15 [11].

ISC ¼ mISCi ; V ¼ Vi ; I ¼ mIi ; RP ¼ RP =m; RS ¼ RS =m
(3.16)
Id ¼ mIdi ; IP ¼ mIPi
where i denotes the individual cell.
RSi/m
+
mIPi
mIdi
mISCi I = mIi
mISCi
Vd
V
RPi/m
Figure 3.15 Compact equivalent circuit for m-PV parallel cells

The array is composed of a combination of series and parallel modules. For

modules in series, V–I curves are simply added along the voltage axis, whereas they
are added along the current axis for parallel modules. For instance, consider an
array consisting of two parallel strings and each string is composed of three mod-
ules in series as shown in Figure 3.16. The total V–I curve is the sum of the indi-
vidual module curves (Figure 3.17). The number of modules in series and the
number of strings are determined according to the power needed.
3.2.7 PV cell power curve

Referring to V–I curve of a PV module, it is seen that the short-circuit current, ISC,
will flow when the cell terminals are short-circuited, and the output voltage will be
zero. Also, when the cell is exposed to the sun and its terminals are opened, an
open-circuit voltage is produced with no flowing current. So, in both cases there is
no power delivered as the power is the product of current and voltage, and conse-
quently, no power is received by the load. With connecting a load, some combi-
nation of current and voltage will result, and power is delivered. Therefore, at each
point on the V–I curve the power delivered by the module can be obtained by
calculating the voltage times its corresponding current. The module power curve
will be as shown in Figure 3.18. It is noted that the maximum power point (MPP) is
that point near the knee of the V–I curve where the product of current, Im, and
voltage, Vm, reaches its maximum. It can be located by finding the biggest possible
rectangle fitted beneath the V–I curve where its sides will be Im and Vm or analy-
tically MPP can be obtained as follows.
String #1
Module #1 Module #2 Module #3
String #2
– V +
Figure 3.16 An array consists of two parallel strings; each is composed of three
series modules
Solar energy 85
The array (string #1 + string #2)
Module #1 Module #2 Module #3
Figure 3.17 Current–voltage curve for the array in Figure 3.16
Pm
r
we
Po
Current (A)
Power (W)
ISC
Im
MPP
0
0 Vm VOC
Voltage (V)
Figure 3.18 I–V and power output curves for a PV module

Assuming a load with resistance RL is connected to the cell terminals shown in

Figure 3.12, the output current, I, can be obtained from
RP
I¼ ISC IS e38:9Vd 1 (3.17)
RP þ R S þ R L
where
Vd ¼ I ðRS þ RL Þ ¼ V ð1 þ RS =RL Þ (3.18)
The output power of the cell, P, is the product of the output current, I, and the
terminal voltage, V, i.e.
VRP h i
P¼ ISC IS ðe38:9V ð1 þ RS =RL Þ 1Þ (3.19)
RP þ RS þ R L
Differentiating (3.19) with respect to V and equating the derivative to zero, the
terminal load voltage at maximum power of the cell, Vm, must satisfy
ISC þ IS
IS e38:9Vm ð1 þ RS =RL Þ ¼ 38:9ð1 þ RS =RL Þ
(3.20)
1 þ Vm
and the maximum power, Pm, may be given by
Vm ðISC þ IS Þ
Pm ¼ (3.21)
1 þ 0:0257=Vm ð1 þ RS =RL Þ
The conversion efficiency for the maximum power, hmc, can be obtained in
terms of the incident solar power, PSP, as
hmc ¼ Pm =PSP (3.22)
In addition to MPP, another quantity is also used to characterize the module
performance [8]. It is known as the fill factor (FF), which is the ratio of the power at
MPP to the product of VOC and ISC, i.e.
Power at MPP V m Im
FF ¼ ¼ (3.23)
VOC ISC VOC ISC
The deduction of output power curve depends on the form of V–I curve. So, it
is important to investigate the factors having prominent effects on the characteristic
curve (V–I curve) of a PV module. These factors are mainly cell temperature,
insolation intensity and shading.
3.2.7.1 Temperature and insolation impact

V–I curves shift as cell temperature and insolation change. Typical such changes
are shown in Figure 3.19. With increasing cell temperature, the short-circuit current
is slightly increased whereas the open-circuit voltage is substantially decreased.
Therefore, the MPP slides slightly upward and towards the left yielding a decrease
in the amount of maximum power (e.g. for crystalline silicon cells ISC increases by
Solar energy 87
ISC1
ISC2
T3
T 3 > T2 > T1 T2 ISC3
Current (A)
Current (A)
T1
Constant insolation
ISC4
ISC1 > ISC2 > ISC3 > ISC4

Constant cell temperature
Voltage (V) Voltage (V)

Cell temperature impact Insolation impact
Figure 3.19 I–V curves at different cell temperatures (in the left) and insolation
levels (in the right)
0.05%/ C, VOC drops by 0.37%/ C and Pm decreases by 0.5%/ C). Hence, the
effect of module temperature should be considered when estimating the perfor-
mance because of its significant impact.
It is to be noted that cell or module temperature differs from the ambient
temperature. Cell temperature varies not only because of ambient temperature
change, but also because of the change of insolation on the cell. A small fraction of
insolation is converted into electricity and the rest is absorbed and converted into
heat. In most cases, manufacturers provide an indicator known as nominal operat-
ing cell temperature (NOCT) to help designers in cell performance estimation with
temperature. NOCT is the cell temperature in a module at ambient temperature of
20 C, irradiation of 0.8 kW/m2 and wind speed of 1 m/s. For other ambient tem-
peratures, Tamb ( C), the cell temperature, Tcell ( C), may be obtained by the
expression given as follows [8]:

NOCT 20
Tcell ¼ Tamb þ S (3.24)
0:8
where S is solar insolation (kW/m2).
Another expression for estimating cell temperature can be used, if NOCT is not
given, as follows [8]:

Insolation
Tcell ¼ Tamb þ g (3.25)
1 kW=m2
where g is a proportionality factor. It depends on wind speed and the state of
module ventilation when installed. Commonly, it ranges from 25 to 35 C.
On the other hand, as the insolation drops, short-circuit current drops in direct
proportion whereas the open-circuit voltage is reduced by modest changes.
Example 3.2 In a 36-cell PV module, all cells are identical. Each cell with 1-sun
insolation (1 kW/m2) has a short-circuit current ISC ¼ 7:45 A, reverse saturation
current IS ¼ 4 1010 , parallel resistance RP ¼ 7:2 W and series resistance
RS ¼ 0:006 W.
(a) If the junction voltage of each cell, Vd, is 0.5 V, find the output voltage,
current and power delivered by the module.
(b) Considering the open-circuit voltage, VOC, of the module at the standard
temperature of 25 C is 21.5 V, the maximum power delivered, Pm ¼ 120 W,
and the ambient temperature ¼ 30 C. Estimate cell temperature, VOC, and the
maximum power delivered.
Solution
(a) From (3.8), the current can be obtained as
Vd
I ¼ ISC Is e38:9Vd 1
RP
0:5
¼ 7:45 4 1010 e38:90:5 1 ¼ 7:27 A
7:2
The voltage produced by 36-cell PV module as in (3.12) is
Vmodule ¼ nðVd RS I Þ ¼ 36ð0:5 0:006 7:27Þ ¼ 16:43 V
Thus, the power delivered, P, will be
P ¼ IVmodule ¼ 7:27 16:43 ¼ 119:44 W
(b) Using (3.25) and assuming g ¼ 30 C, the cell temperature is

Insolation
Tcell ¼ Tamb þ g ¼ 30 þ 30 ¼ 60 C
1 kW=m2
That means the cells are getting hotter than the ambient by 30 C.
As explained in Section 3.7.2.1, the open-circuit voltage drops by about
0.37% per degree centigrade. Thus, VOC drops, at 60 C, to
VOC ðat 60 CÞ ¼ VOC ðat 60 CÞ½1 0:0037ðTcell Tstd Þ
¼ 21:5½1 0:0037ð60 25Þ ¼ 18:716 V

and the percentage of open-circuit decrease is about 12.9%.
Solar energy 89
Similarly, the power delivered drops by about 0.5% per degree centigrade.
Then,
Pm ðat 60 CÞ ¼ Pm ðat 25 CÞ½1 0:005ðTcell Tstd Þ
¼ 120½1 0:005ð60 25Þ ¼ 99 W
that is, the maximum power delivered by the module is decreased by
about 17.5%.
3.2.7.2 Shading impact

The output voltage of a PV module can reduce extremely when one or more cells
are shaded. To illustrate that, consider a module of n-PV cells and one of these cells
is shaded (e.g. the cell on the top), whereas the other n 1 cells are in the sun as
shown in Figure 3.20. The current source of the shaded cell is reduced to zero. The
voltage drop across RP due to current flow through it causes the diode to be reverse
biased, and thus, the diode current will also be zero. Hence, the entire current
flowing through the module must flow through both RP and RS in the shaded cell to
complete its way to the load. That means the shaded cell reduces the output voltage
rather than adding to it. The output voltage of the module with one cell shaded will
drop to become
Vsh ¼ Vn1 I ðRP þ RS Þ (3.26)
Assuming the output voltage in case of all cells in the sun is V and carrying a
current I, then the voltage of the n 1 cells can be obtained by

n1
Vn1 ¼ V (3.27)
n
nth cell – shaded

ISC = 0
Id = 0
Vn−1 RS
n − 1 cells in the sun
I I
RP
I
I
I
– + Vsh
Figure 3.20 A PV module with (n 1) cells in the sun and nth cell in the shade
+ Blocking diodes ∑I
Vsh
I +
I=0 I I I
Bypass diode V
I –
–
(a) (b)
Figure 3.21 Mitigation of shading impact: (a) shaded cell in a module with
bypass diode and (b) shaded string in an array with blocking diodes
Incorporate (3.26) and (3.27) to give

n1
Vsh ¼ V I ðRP þ RS Þ (3.28)
n
Thus, the voltage drop caused by the shaded cell, DV, at any given current I is
given by

n1 V
DV ¼ V Vsh ¼ V V þ I ðR P þ R S Þ ¼ þ I ð R P þ R S Þ (3.29)
n n
Considering RP RS , (3.29) can be approximated and rewritten as

V
DV ffi þ IRP (3.30)
n
Different efforts are made by manufacturers or designers to mitigate shade
problems such as using bypass diodes for a partial shaded PV module or block
diodes to isolate the shaded string in an array (Figure 3.21). More details about that
can be found in [10].
Example 3.3 Suppose that a 36-cell module has the following data for the standard
conditions (1 kW/m2, 25 C, AM 1.5): effective illumination area ¼ 0.136 m2, and
the voltage and current at the maximum power are 15.6 V and 1.4 A, respectively.
Solar energy 91
The parallel and series resistance equal 6.6 and 0.005 W, respectively. It is required
to find the following:
(a) Module efficiency for the maximum electrical power.
(b) If one cell is shaded, calculate the module output voltage, voltage across the
shaded cell and the power dissipated in this cell.
Solution
(a) Module efficiency for the maximum power, h, will be
Pelectrical 15:6ð1:4Þ 21:84
h¼ ¼ ¼ ¼ 16:06%
Psolar 1; 000ð0:136Þ 136
(b) Using (3.29), the voltage drop across the module will be
V 15:6
DV ¼ þ IRp ¼ þ 1:4ð6:6Þ ¼ 9:673 V
n 36
Thus, the output voltage, Vsh ¼ 15:6 9:673 ¼ 5:927 V and the corre-
sponding power, Psh ¼ Vsh I ¼ 5:927ð1:4Þ ¼ 8:298 W.
As well, the efficiency is reduced and becomes
8:298
h¼ ¼ 6:1%
136
That means, the decrease in efficiency T ¼ 16:06 6:1
16:06 ¼ 62:02%.
The voltage drop across the shaded cell, Vcellsh ¼ I Rp þ Rs ¼
1:4ð6:605Þ ¼ 9:247 V. The shaded cell subtracts more than 9 V from the
module voltage, 15.6 V. Power dissipated in the shaded cell,
Pcellsh ¼ Vcellsh I ¼ 9:247ð1:4Þ ¼ 12:945 W. It is converted to heat causing
a local hot spot that may cause damage in the plastic layer enclosing the cell.
3.3 Concentrating solar power

CSP technology is used primarily for utility-scale projects. It uses mirrors or lenses
and tracking systems to reflect and focus a large area of sunlight onto a receiver
that converts the solar energy to thermal energy through heating a working fluid.
Then, the working fluid (heat transfer fluid) can be used as a heat source for con-
ventional power plant or energy storage.
Therefore, the CSP process, as shown in Figure 3.22, mainly includes
(i) concentrators supplemented with tracking system to track the sun and focus it
onto thermal receivers, (ii) receivers to absorb the solar energy and convert it to
heat, (iii) working fluid to be heated and used as a heat source and (iv) power plant
or ESS to use the heat source for producing electricity.
Conversion of sunlight energy to heat Conversion of the heat to

electricity in the power plant
Concentrators Receivers Heat source

Power plant
Storage
system
Tracking
system
Figure 3.22 Concentrating solar energy process
A wide range of concentrating energy technologies are applied. Three main

most developed are (i) linear concentrator technology, which produces high-
pressure superheated steam, (ii) dish/engine technology, which can directly gen-
erate electricity in isolated locations and (iii) power tower technology, which
produces air above 1,000 C or synthesis gas turbine operation [12].
3.3.1 Linear concentrator systems

The linear concentration is the technology where the concentration can be done on
a line. There is not a precise point of focusing, but a line. Linear concentrator
systems collect the sun’s energy using mirrors. Only one coordinate needs to follow
the sun with an optimized collector adjustment. The mirrors are tilted towards the
sun to reflect and concentrate sunlight onto linear receivers (or tubes) located along
the focal line of the mirrors. The reflected sunlight heats a fluid flowing through the
tubes. The hot fluid then is used to boil water in a conventional steam-turbine
generator to produce electricity. There are two major types of linear concentrator
systems: parabolic trough systems and linear Fresnel reflector systems.
For parabolic trough systems, the collectors are installed in parallel rows (each
row consists of a number of concatenated parabolic-shaped mirrors) aligned on a
north-south horizontal axis. Using a tracking system, they rotate from east to west
to track the sun throughout the day. The receiver tubes are located along the foci of
the parabolas (Figure 3.23).
A typical example is shown in Figure 3.24 to illustrate the process of gen-
erating electricity. The solar collector field consists of an array of parabolic trough
solar concentrators to reflect sunlight onto the receiver (absorber tube). The fluid
flowing through the absorber tube, usually synthetic oil or water/steam, transfers
the heat to a conventional steam-turbine power cycle.
For linear Fresnel reflector systems, a developed collector with segmented flat
mirrors following the principle of Fresnel is used. One receiver tube is positioned
above several mirrors to allow the mirrors greater mobility in tracking the sun
Solar energy 93
Working fluid
Sunrise
Sunset
Reflector
Sunlight
N
E
Receiver
W
Concentrator Receiver
S
Working fluid
(a) (b)
Figure 3.23 Structure of parabolic trough solar concentrating system:

(a) singular parabolic trough concentrator; (b) array of
parabolic trough collector system
Solar field Thermal energy storage Steam power plant connected to the grid
Grid
Reheater
Solar
collector Super-heater
field Hot tank
Turbine Generator
Storage
Evaporator
Condenser Cooling
Cold tank
tower
Heat transfer fluid pump Reheater Feed pump
Figure 3.24 Schematic diagram of a steam cycle power plant with parabolic
trough collector and energy storage [12]
(Figure 3.25). The linear Fresnel system also shows a good potential for low cost
steam generation and provides a semi-shaded space below, which may be particu-
larly useful in desert climates. However, further modelling and experimental work
is still required to determine under what conditions it may be more cost-effective
than the parabolic trough system with direct steam generation [13].
Sunlight
Absorber tube
Fresnel
reflector
Figure 3.25 Fresnel concentrator structure (on the left) and a typical example
(on the right)
Linear Fresnel reflectors are optically analogous to parabolic troughs. They are
concentrating reflectors with linear focus, where the parabolic reflective surface is
obtained by an array of linear mirror strips which independently move and
collectively focus on absorber lines suspended from elevated towers [14,15].
3.3.2 Dish-engine systems

Dish-engine systems use dish-shaped parabolic mirrors, like a very large satellite
dish, as reflectors to direct and concentrate sun’s rays onto a thermal receiver. To
minimize costs, the mirrored dish is usually composed of many smaller flat mirrors
formed into a dish shape (Figure 3.26). The receiver is mounted above the dish at
the dish focal point. It absorbs the sunlight energy and converts it into thermal
energy that can be either transported to a central generator for conversion or con-
verted directly into electricity at a local generator coupled to the receiver.
Dish-engine systems are the most efficient collector systems using dual-axis
tracking system to keep the collector pointing at the sun. Different types of engines
have been considered for using dish-engine system such as Rankine cycle engines,
Brayton cycle engines and sodium-heat engines, but greatest attention has been
paid to Stirling engine system. This system uses the fluid heated by the receiver to
move pistons and create mechanical power which can be used to run a generator to
produce electricity. In addition, a dish-Stirling system can be stand-alone power
plant. It does not need access to fuel lines or sources of cooling water as well as it is
ideal for producing electricity in sunny desert areas. To increase the reliability and
guarantee the service continuity of dish-Stirling systems, some units have been
designed to operate in a hybrid mode in which fuel is burned to heat the engine at
the time of solar inadequacy.
Solar energy 95
Receiver
coupled to engine
t
ligh
Sun
Parabolic dish
Figure 3.26 Dish-Stirling engine structure; parabolic mirror (on the left)
and multi-flat mirrors formed into a dish shape (on the right)
[http://www.csiok.cz/assets/docs/esap-CSP-02.pdf]
3.3.3 Power tower systems

Power tower systems use a large field of flat, sun-tracking, computer-controlled
mirrors known as heliostats to concentrate sunlight onto a receiver located on the
top of a tall tower (Figure 3.27). A heat transfer fluid in the receiver passes
optionally to storage and finally to generate steam, which in turn, is used in a
conventional turbine-generator to produce electricity and supplies it to the grid
(Figure 3.28). Water/steam may be used as a heat transfer fluid for some towers,
whereas other advanced designs use molten nitrate salt because of its superior heat
transfer and energy storage capabilities. The energy storage capability adds an
advantage of allowing the system to continue to dispatch electricity during cloudy
weather or at night in addition to the simplicity, scalability, quick start-up, high
efficiency due to high achievable temperatures. Therefore, the power tower system
that can also be named as a central receiver system is composed of five main
components: (i) heliostats including their tracking system, (ii) receiver, (iii) heat
transport and exchange, (iv) thermal storage and (v) control system.
The heliostat can be constructed as a group of several distributed flat mirrors
positioned in a manner that the reflected rays are collected at the receiver on the top
of the tower (Figure 3.29).
An overview of some of the technical parameters of the different CSP tech-
nologies is given in Table 3.2. Parameters in the fifth and sixth columns are defined
as follows:
Net power generation
Solar efficiency ¼ (3.31)
Incident beam radiation
Solar operating hours per year
Capacity factor ¼ (3.32)
8;760 hours per year
Solar tower
Central receiver
t
ligh
Sun
Mirrors
Figure 3.27 Power tower system basics
Molten salt is The heated salt is

heated in the then pumped
tower, then Hot through a steam
Large mirrors pumped into a generator
focus the sun’s storage tank
rays on a central
collection tower
The steam is used
to power a
Cold
turbine generator
Cooled salt is
returned to a Grid
second storage
tank to be sent
back through the
cycle Generator
Use of molten salt in solar power generation to store heat until it is needed
Figure 3.28 Power tower system connected to the grid [16]

Solar energy 97
Steam condenser
Electricity
Receiver
Feedwater
reheater
Generator
Turbine
Steam drum
Heliostats
Figure 3.29 Power tower system with distributed flat mirrors [17]
Parabolic troughs, linear Fresnel systems and power towers can be coupled to
steam cycles of 10–200 MW of electric capacity, with thermal cycle efficiencies of
30%–40%. The values for parabolic troughs, by far the most mature technology,
have been demonstrated in the field. Today, these systems achieve annual solar-to-
electricity efficiencies of about 10%–15% and may reach about 18%. The values
for other systems are, in general, projections based on component and prototype
system test data, and the assumption of mature development of current technology.
Overall solar-electric efficiencies are lower than the conversion efficiencies of
conventional steam or combined cycles, as they include the conversion of solar
radiation energy to heat within the collector and the conversion of the heat to
electricity in the power block. The conversion efficiency of the power block
remains essentially the same as in fuel-fired power plants [12].
Because of their thermal nature, each of these technologies can be ‘hybri-
dized’, or operated with fossil fuel as well as solar energy. Hybridization has the
potential to improve dramatically the value of CSP technology by increasing its
power availability and dispatch ability, decreasing its cost (by making more
effective use of the power block equipment), and reducing the technological risk
by allowing conventional fuel use if, for example, the collector must be repaired.
Solar heat collected during the daytime can be stored in concrete, molten salt,
ceramics or phase-change media. It can be extracted at night from storage to run
Table 3.2 Performance data for CSP technologies
Class Type of Capacity Concentration Peak solar Annual solar Thermal cycle Capacity factor
concentrator (MW) efficiency (%) efficiency (%) efficiency (%) (solar) (%)
Line Trough 10–200 70–80 21 (d) 10–15 (d) 30–40ST 24 (d)
concentration
17–18 (p) 25–70 (p)
Fresnel 10–200 25–100 20 (p) 9–11 (d) 30–40ST 25–70 (p)
Central point Dish-Stirling 0.01–0.4 1,000–3,000 29 (d) 16–1 (d) 30–40ST 25–70 (p)
concentration
18–23 (p) 20–30GT
Power tower 10–150 300–1,000 20 (d) 8–10 (d) 30–40ST 25 (p)
35 (p)
15–25 (p) 45–55CC
(d) ¼ demonstrated; (p) ¼ projected; ST ¼ steam turbine; GT ¼ gas turbine; CC ¼ combined cycle.
Solar energy 99
the power block. Fossil and renewable fuels such as oil, gas, coal and biomass
can be used for co-firing the plant, thus providing power capacity whenever
required.
Moreover, solar energy can be used for co-generation of electricity and
heat. In this case, the high-value solar energy input is used with the best
possible efficiencies of up to 85%. Possible applications include the combined
production of electricity, industrial process heat, district cooling and seawater
desalination.
References
[1] Sayigh A. A. M. Solar energy engineering. 2nd edn. Orlando, FL: Academic
Press; 2013.
[2] Chwieduk D. Solar energy in buildings. San Diego, CA: Academic Press;
2014.
[3] National Renewable Energy Laboratory. Solar energy basics [online]. 2016.
Available from https://www.nrel.gov/research/re-solar.html [Accessed 20
June 2016].
[4] Wikipedia, the free encyclopedia. Solar energy [online]. 2019. Available
from https://en.wikipedia.org/wiki/Solar_energy [Accessed 31 January
2020].
[5] Gevorkian P. Large scale solar power systems – Construction and eco-
nomics. Cambridge, UK: Cambridge University Press; 2012.
[6] Kiatreungwattana K., VanGeet O. and Stoltenberg B. Facility scale solar
photovoltaic guide book. Report NREL/Bureau of reclamation, 2016.
[7] Ehrlich R. Renewable energy: A first course. Boca Raton, FL: CRC Press,
Taylor & Francis Group; 2018.
[8] Masters G. M. Renewable and efficient electric power systems. New York:
Wiley-Interscience, John Wiley & Sons; 2004.
[9] Department of Energy’s Office of Energy Efficiency & Renewable Energy.
Solar photovoltaic technology basics [online]. 2013. Available from
http://energy.gov/eere/energybasics/articles/photovoltaic-technology-basics
[Accessed 31 January 2020].
[10] Krauter S. C. W. Solar electric power generation – Photovoltaic energy
systems. Berlin Heidelberg: Springer Verlag; 2006.
[11] Felix A. and Farret F. A. Integration of alternative sources of energy.
New York: Wiley-IEEE Press; 2006.
[12] Steinhagen H. M. and Trieb F. Concentrating solar power – A review of the
technology [online]. Stittgart, German: Institute of Technical Thermodynamics,
German, Aerospace centre; 2003. Available from https://www.researchgate.
net/publication/224797493_Concentrating_solar_power_-_A_review_of_
the_technology [Accessed 31 January 2020].
[13] Mills D. R. and Morrison G. L. ‘Compact linear Fresnel reflector solar
thermal power plants’. Solar Energy. 2000;68(3):263–83.
[14] Cau G. and Cocco D. ‘Comparison of medium-size concentrating solar

power plants based on parabolic trough and linear Fresnel collectors’. 68th
Conference of the Italian Thermal Machines Engineering Association,
ATI2013; Bologna, Italy, Sep 2013. Energy Procedia, Elsevier. 2014;45:
101–10.
[15] El-Gharbi N., Derbal H., Bouaichaoui S. and Said N. ‘A comparative study
between parabolic trough collector and linear Fresnel reflector technolo-
gies’. Energy Procedia, Elsevier. 2011;6:565–72. Available from www.sci-
encedirect.com [Accessed 31 January 2020].
[16] Vignarooban K., Xu X., Arvay A., Hsu K. and Kannan A. M. ‘Heat transfer
fluids for concentrating solar power systems – A review’. Applied Energy,
Elsevier. 2015;146:383–96.
[17] U.S. Department of Energy. ‘Power tower system concentrating solar power
basics’. Solar Energy Technologies Office’; 2013. Available from https://
www.energy.gov/eere/solar/articles/power-tower-system-concentrating-solar-
power-basics [Accessed 31 July 2020].
Chapter 4
Ocean energy: tidal energy
4.1 Introduction
Ocean space has a huge amount of energy. Useful electric energy, generally known
as ‘ocean energy’, can be derived from the ocean. It is an alternative renewable
energy source (RES), like solar and wind energy. The important feature of ocean
energy resources is that they have the highest density among the other sources of
renewable energy.
Ocean energy comes from three distinct sources of water energy with different
origins and conversion technologies as in the following:
● Tidal energy: It is derived from gravitational forces of earth–moon–sun system
as explained in this chapter.
● Waves and swells energy: It is derived from the wind energy and is the subject
of Chapter 5.
● Ocean thermal energy conversion: It is derived from the solar energy that is
stored in a form of heat in ocean surface layers or hydrothermal energy at
submarine volcanic centres. Chapter 5 focuses on this kind of energy.
Large amounts of renewable energy can be harnessed by these three sources
because of the oceans’ large area.
4.2 Physics of tidal phenomena

Tidal phenomena are characterized by the currents and variations in the sea level
caused by the gravitational forces of both the moon and the sun on the earth. This
movement of vast quantities of water around the oceans and seas producing tides
every day can be classed as an alternative RES. The large amount of energy in the
tides may be considered as a form of hydropower that can be converted into elec-
trical energy. The sun does not always shine, and the wind does not always blow.
However, what the tides will be doing at any instant in time can be reliably cal-
culated for millennia in advance. Thus, tidal energy is more predictable compared
to wind and solar energy, but its cost is relatively high as well as the availability of
sites with sufficient flow velocities is limited. However, much attention is currently
being devoted to technological developments aimed at getting both higher avail-
ability and lower cost.
As the earth, moon and sun rotate around each other in space, the gravitational
movement of the moon and the sun with respect to the earth causes millions of litres
of water to flow around the earth’s oceans creating periodic shifts in these moving
molecules of water. These vertical shifts of water are called ‘tides’.
The attractive force exercised by the moon has the larger impact on the water
molecules (two or three times greater) than that by the sun because of the shorter
distance between the earth and moon. Depending on how the earth, moon and sun
are positioned relative to each other, the illuminated part of the moon as seen from
the earth varies and passes through different phases. So, the phases of the moon are
the different appearances of the illuminated moon from a point on the earth. The
moon itself does not change, but the part of it that is reflecting sunlight is visible to
a different degree. For example, when the earth is located between the moon and
the sun, one side of the moon can be seen fully illuminated by the sun. It looks
perfectly round and is called the ‘full moon’. When the moon is situated between
the earth and the sun, the illuminated side of the moon faces away from the earth
and this is what is called the ‘new moon’. It cannot be seen, but a day or two after
the new moon, a skinny crescent-shaped moon appears. It gets bigger and bigger
until it reaches the full moon, then it wanes again, becomes a crescent and finally
goes back to be a new moon. This cycle (new moon, full moon, new moon) is
called ‘lunar cycle’ and takes approximately 29 days from one new moon to
another.
The lunar cycle mostly coincides with a calendar month, but not completely.
This means that a new moon will typically be seen once a month, but sometimes
there will be two in a month. The second new moon in a month is called a ‘black
moon’. Like a blue moon (the second full moon in a month), a black moon occurs
every two and a half years or so. The black moon does not get as much attention as
a blue moon, which makes sense as the black moon cannot be seen.
The moon’s position relative to the earth and the sun as viewed from the outer
space, above the solar system, as well as the moon as seen from earth are shown as
inner and outer sequence, respectively, in Figure 4.1 [1]. For instance, the waxing
gibbous (upper left) as a large part of the illuminated moon is seen on the right side.
A waning crescent (lower right) as a small slice of the moon illuminated would
have the left side of the moon illuminated. The physical pictures of the moon
phases illustrating the gradual change of the illuminated moon are shown in
Figure 4.2.
At any instant, there are two high tides and two low tides around the earth
because of the moon’s gravitational attraction. One high tide is between the moon
and the earth and closest to the moon (the sublunar point), and another on the
opposite side of the earth and furthest from the moon (the antipodal point). The
points in between are the resulting low tides (Figure 4.3). The difference in height
between a high tide and low tide is defined as the tidal range [2].
The height variation of these tides is an effective factor to consider when
producing tidal power. The height is affected by the relative positions of the moon
and the sun. When the moon is in perfect alignment with the earth and sun, the
combined effects of the lunar and solar tides result in a greater tidal bulge.
Ocean energy: tidal energy 103
First Outer sequence

quarter
xing
Wa
Gibbous Crescent
Inner sequence
Sun
Full New
Earth
Wanin
g
Gibbous Crescent
Third
quarter
Figure 4.1 Moon phases diagram (the lunar cycle) [2]
New First First Full

moon quarter quarter moon
(a) (b)
Full Third Third New
moon quarter quarter moon
(c) (d)
Figure 4.2 Physical pictures of the moon appearance in phase; (a) from new
moon to first quarter, (b) from first quarter to full moon, (c) from full
moon to third quarter, and (d) from third quarter to new moon [3]
Meanwhile, the tides have a larger tidal range (maximum) and become much
stronger than normal. The high tides become very high and the low tides become
very low during each tidal cycle. These tides occur during the full or new moon
phase and are known as ‘spring tides’.
Other neap tides having minimum height arise when the sun, moon and earth
are not in alignment during first and third quarter moon phases (the pull of the
moon is at 90 to the sun). The lunar and solar tides, therefore, occur in different
places, so the overall gravitational effect on the water is less. The net result is a
Tidal bulge towards moon

Tidal bulge opposite moon
Moon
Antipodal point Sub-lunar point
Earth
Figure 4.3 Basics of tide formation [2]
Quarter
moon 90° alignment
The moon is in perfect alignment with the earth and the sun
Full or Earth Sun

Earth Sun
new moon
(a) (b)
Figure 4.4 Alignment of moon and sun on (a) spring tides and (b) neap tides
smaller pulling action on the seawater creating much smaller differences between
the high and low tides, thereby producing very weak tides (Figure 4.4).
4.2.1 Types of tides

Tidal phenomena are, therefore, periodic, though the exact nature of the periodic
response varies depending, mainly, on the combined effect of the lunar and solar
gravitation, and the relative positions of the earth, sun and moon in addition to
other geographical considerations. There are three main types of tides at different
locations around the globe as described in the following [4]:
1. Semidiurnal tides: These tides are regular resulting in two high and two low
tides each day with the level of the oceans constantly moving between a high
tide and a low tide. The rotational movement of the earth is 180 around its
axis each 12 h, whereas for the moon around the earth is 6 in the same period,
12 h. So, the time taken for a tidal cycle to happen is about 12 h and 24 min
(called the ‘diurnal cycle’) between two consecutive high tides. The lunar
behaviour dominates this type of tide, which is observed on the non-
Mediterranean stretches of the French coastline.
The amplitude of the tide varies through the lunar month, with the tidal
range being greatest at the full moon or new moon – that is, when earth, sun
First Full Third New

quarter moon quarter moon
12 h and 24 min 7 days
Time
x
Wa Neap
Spring Wan tide
tide
Figure 4.5 Semidiurnal tides monthly cycle
and moon are aligned and the pulls of both gravitational forces combine to
produce a spring tide. The tidal range is at its lowest at the first and third
quarters of the moon – that is, when the pull of the moon is orthogonal to the
sun producing neap tides (Figure 4.5) [4].
2. Diurnal tides with monthly variation: The tidal period of this type corresponds
to a full revolution (24 h and 50 min). The ‘day-side’ gravitational attraction
outweighs the ‘night-side’ attraction because of the variation in the distance
from the axis of rotation, and also the tidal range varies with the monthly lunar
cycle. This type of tide is found in the China Sea and Tahiti.
3. Mixed tides: This type combines the characteristics of semidiurnal and diurnal
tides, either of which may display monthly or bimonthly variation. The laws
governing water level variation in mixed tide areas can be highly complex.
Examples of mixed tides are those observed in the Mediterranean and at Saigon.
4.3 Tidal energy estimation
Throughout the year, the position of the earth and the moon with respect to the sun
changes. So, the potential energy of the water contained in the daily movement of
the rising and falling sea levels can be exploited to generate electricity. Electricity
generation from tides is similar in many ways to hydroelectric generation. The
difference is that the water flows in and out of the turbines in both directions rather
than in just one forward direction.
Tidal energy, just like hydro energy, transforms water in motion into a clean
energy. The motion of the tidal water, driven by the pull of gravity, contains large
amounts of kinetic energy in the form of strong tidal currents called ‘tidal streams’.
The daily ebbing and flowing, back and forth of the oceans’ tides along a coastline,
and into and out of small inlets, bays or coastal basins, is little different to the water
flowing down a river or stream.
The movement of the seawater is harnessed in a similar way using waterwheels
and turbines to that used to generate hydroelectricity. Because of the bi-direction of
seawater flow in a tidal energy system, electricity can be generated when the water
is flowing in and also when it is ebbing out. It means that the tidal generators are
Earth axis
ω
Moon
P
Earth’s equatorial Y X
O O
plane L
r d
Figure 4.6 Physical explanation of the semidiurnal tide
designed to produce power when rotor blades are turning in either direction and,
consequently, the reversible electrical generators cost more than the single direc-
tion generators.
4.3.1 Tide generating forces

Tidal energy is derived from the gravitational forces of attraction, which operate
between the earth and the sun and between the earth and the moon. The gravitational
force, F, which mutually attracts any two bodies of mass m1 and m2, respectively, is
directly proportional to the product of their masses and inversely proportional to the
square of the distance, d, separating them, that is, F / m1 m2 =d 2 [5].
Thus, the attractive force exercised by the sun and the moon, respectively, on a
water molecule at the surface of the earth, FS and Flu , can be given by
FS ¼ GMS ME =dS2 and Flu ¼ GMlu ME =dlu2 (4.1)
where
D
G ¼ gravitational constant;
D
MS ¼ mass of the sun;
D
ME ¼ mass of the earth;
D
Mlu ¼ mass of the moon;
D
dS ¼ distance between the centres of mass of the earth and the sun;
D
dlu ¼ distance between the centres of mass of the earth and the moon.
Accordingly, and as shown in Figure 4.6, the gravitational forces FSX and FluX ,
at point #X, the nearest to the sun or the moon, are
FSX ¼ GMS ME =ðdS rÞ2 and FluX ¼ GMlu ME =ðdlu rÞ2 (4.2)
where r is the equatorial radius of the earth.

Similarly, at point #Y, the furthest from the sun or the moon, the gravitational
forces, FSY and FluY , are
FSY ¼ GMS ME =ðdS þ rÞ2 and FluY ¼ GMlu ME =ðdlu þ rÞ2 (4.3)
and at the centre of mass of the earth, O, the gravitational forces, FSO and FluO , are
FSO ¼ GMS ME =dS2 and FluO ¼ GMlu ME =dlu2 (4.4)

The tide-generating force is equal to the difference between the gravitational
force at the centre of mass of the earth and that at other points on the earth. So, its
value at point #X and point #Y, FSX -tide and FSY -tide , the nearest to and furthest from
the sun, respectively, are
FSX -tide ¼ FSX FSO ¼ GMS ME =ðdS rÞ2 GMS ME =dS2

¼ GMS ME 2rdS 1 r2 =2rdS =r4 ðignoring the small value r=dS Þ

¼ G 2MS ME r=dS3 ðwhere r=2dS is very smallÞ
(4.5)
and

FSY -tide ¼ FSO FSY ¼ G 2MS ME r=dS3 (4.6)
From (4.5) and (4.6), the tide-generating force on the earth due to the sun,
FS -tide , can generally be expressed as

FS -tide ¼ G 2MS ME r=dS3 (4.7)
Similarly, the tide-generating force at the surface of the earth (at the points #X
and #Y) nearest to and furthest from the moon, FluX -tide and FluY -tide , respectively,
can be given by

FluX -tide ¼ FluX FluO ¼ G 2Mlu ME r=dlu3 (4.8)

FluY -tide ¼ FluO FluY ¼ G 2Mlu ME r=dlu3 (4.9)
Therefore, the general form of the tide-generating force on the earth due to the
moon can be written as

Flu-tide ¼ G 2Mlu ME r=dlu3 (4.10)
From (4.7) and (4.10), the tide-generating force ratio (TGFR), which is the
ratio of the tide-generating force by the moon to that by the sun, can be given by

2Mlu ME r dS3
TGFR ¼ ¼ ðMlu =MS ÞðdS =dlu Þ3 (4.11)
dlu3 2MS ME r
The sun is 27 106 times greater in mass than the moon, and the moon is 389
times nearer to the earth than the sun. So, as given by (4.11), the moon is 2.16 times
more influential in causing the tide than the sun.
It is to be noted that although the gravitational force varies inversely as the
square of the distance between the two bodies, the tide-generating force varies
inversely as the cube of the distance. This is very important as it accounts for the
moon being 2.16 times more influential in causing the tide than the sun.
Furthermore, the ranges of the lunar tide Rlu and solar tide RS are proportional to the
difference in force, that is,
Rlu
¼ TGFR ¼ 2:16 (4.12)
RS
As explained in Section 4.2.1, the solar tide moves in and out of phase with the
lunar tide. When the sun, earth and moon are aligned, the lunar and solar tides are
in phase and produce tides of maximum range, named as ‘spring tides’. These tides
occur twice per lunar month at times of both full and new moons, Figure 4.5. When
the sun/earth and moon/earth directions are perpendicular, the tides of least range,
named as ‘neap tides’ are produced and occur twice per lunar month.
Considering the spring tide results from the sum of lunar and solar tides, and the
neap tide from their difference, the ratio of spring to neap ranges can be calculated as
Rspring 1 þ ð1=2:16Þ
¼ ¼ 2:72 (4.13)
Rneap 1 ð1=2:16Þ
As an alternative development of using centrifugal force, the procedure is the
same because the centrifugal force everywhere is equal (but of opposite sign) to the
gravitational force at the centre of mass of the earth to maintain the earth-moon
separation. Assuming the centre of revolution is at point #O at L from the centre of
the earth, as shown in Figure 4.6:
GMlu ME
¼ Mlu ðdlu LÞw2 ¼ ME Lw2 (4.14)
d2
D
where w ¼ the lunar frequency.
The force causing the tides can be estimated as the resultant of two forces:
(i) the centrifugal force about O at the lunar frequency w and (ii) the gravitational
lunar force. So, for a water of mass m at the positions X and Y, the following
relations can be written:
GmMlu
FX ¼ þ mðr LÞw2 (4.15)
ðdlu rÞ2
GmMlu
FY ¼ mðr þ LÞw2 (4.16)
ðdlu þ rÞ2
where FX and FY are the force causing tides at points X and Y, respectively.
At position O with definition of L, the relation in the following can be
obtained:
GmMlu
¼ mLw2 (4.17)
dlu2
It is to be noted that r << d, and so,

1 1 2r
¼ 1 (4.18)
ðdlu rÞ2 dlu2 dlu
Incorporating (4.15) to (4.18) gives

2L
FX ¼ FY ¼ mrw2 1 þ (4.19)
dlu
Therefore, two lunar tidal ranges of equal amplitude are expected each day
when the moon is in equatorial plane of the earth. For instance, and as shown in
Figure 4.6, point #P experiences two equal tides each day (semidiurnal tide).
At some point #P with polar coordinates of re and q, as shown in Figure 4.7
and assuming, for simplicity, that the earth is covered by water, the gravitational
potential (y) for a unit mass of water due to the moon is given by [6]
y ¼ GðMlu =d Þ (4.20)
where d is the distance from P to the centre of the moon, and 1/d can be expanded
as follows:
1 1
¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
d d 2 þ r2 2d r cos q
lu e lu e
1

1 2re re2 2
¼ 1þ cos q þ 2
dlu dlu
dlu
1 re re2 3 1
¼ 1 þ cos q þ 2 cos2 q þ
dlu dlu dlu 2 2
(4.21)
Substituting (4.21) in (4.20) gives

GMlu re r2 3 1
y¼ 1 þ cos q þ e2 cos2 q (4.22)
dlu dlu dlu 2 2
It is to be noted from (4.22) that the first term can be ignored as it does not
yield a force. The second term represents a constant force, GMlu re cos q=dlu2 ,
P
Earth covered
by water
re d
h Moon
θ Y
X
dlu
Figure 4.7 Lunar-tidal effect at some point on the surface of the earth
directed towards the centre of the moon. It acts on the earth and is balanced by the
centrifugal force due to the rotation of the earth–moon system. The third term
GM r2
describes the variation of the moon’s potential around the earth, d 3lu e 32 cos2 q 12 .
lu
The surface profile of the water is an equipotential surface due to the combined
effects of the moon and the earth. The potential of the unit mass of water to the
earth’s gravitation is gh, where h is the height of the water above its equilibrium
level and g ¼ GME =r2 is the acceleration due to gravity at the earth’s surface.
Hence, the height of the tide hðqÞ can be obtained from

GMlu re2 3 2 1
ghðqÞ cos q ¼0
dlu3 2 2
That is,

Mlu re4 3 2 1 3 2 1
h ðqÞ ¼ cos q ¼ hmax cos q (4.23)
ME dlu3 2 2 2 2
The maximum height of the tide, hmax , occurs at points #X and #Y, where q ¼ 0
and p, respectively. For Mlu =ME ¼ 0:0123; d ¼ 384;400 km and r ¼ 6;378 km,
the value of hmax 0:36 m, which is around the observed mean tidal height. More
analysis and details about the rotation of celestial bodies and their orbits can be
found in [7–9].
4.3.2 Tidal wave

Tidal wave is the form of a particular type of the ordinary tidal movement of the
sea. The wavelength l is much greater than the mean depth of the sea, h. It is
assumed also that the amplitude of the wave is small compared to the depth, in
which case the vertical acceleration is small compared to the acceleration due to
gravity, g. A slice of water (column of water) of thickness Dx, depth h, width b and
surface level above the mean position z is considered. Thus, the change in surface
level over the thickness Dx is ð@z=@xÞDx, which is small compared to z. The side
area of the water slice is A ¼ hb (Figure 4.8).
Considering Newton’s equation of motion for the slice and the requirement for
conservation of water mass, the form of the wave is obtained as follows [5].
The pressure difference across each side of the slice, Dp, arising from the small
change in height of the surface is

@z
Dp ¼ rg Dx (4.24)
@x
Considering force ¼ mass acceleration, the equation of motion of the slice
of velocity u is

@z @u
rg Dx hb ¼ rhbDx (4.25)
@x @t
z
z
z=0 Water surface
Depth
∆x Velocity u
h
b
y
idt
W
z = –h Seabed x
l Thickness
Figure 4.8 Water motion in a tidal wave
Hence,
@u @z
¼ g (4.26)
@t @x
The difference between the flow of water into and out of the slice must be
accounted for by a change in volume V of the slice with time, and this conservation
of water mass leads to the equation of continuity:

@ @V @ ½ðA þ bzÞDx
ðA þ bzÞu Dx ¼ ¼ (4.27)
@x @t @t
As A ¼ hb is constant and much greater than bz,
@u @z
A ¼b
@x @t
Or
@u 1 @z
¼ (4.28)
@x h @t
From (4.26) and (4.28),
@2u @2z @2u 1 @2z

¼ g 2 ¼ ¼ (4.29)
@t@x @x @x@t h @t2
Thus,
@2z @2z @2z

2
¼ gh 2 ¼ c2 2 (4.30)
@t @x @x
So, (4.30) is the equation of a tidal wave speed c, where
pffiffiffiffiffiffiffiffiffi
c ¼ ðghÞ (4.31)
Another alternative proposed formula for the tidal wave speed as a function of
the spring tide speed, neap tide speed and tide coefficient (C) is given by [10]
ðC s 45Þ þ ðcst cnt Þ
ctide ¼ cnt þ (4.32)
95 45
where
D
ctide ¼ tidal current speed;
D
cnt ¼ neap tide speed;
D
cst ¼ spring tide speed;
D
C s ¼ constant (¼ 95 for spring, 45 for neap tide).
4.3.3 Tidal resonance

The only possibility for enhanced motion is for the natural tidal motion to be in
resonance with the solar and the lunar forces. For instance, the Atlantic Ocean has a
width of about 4,000 km and an average depth of about 4,000 m [6]. The speed of
the tide wave as given by (4.31) is
pffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
c ¼ ðghÞ 10 4;000 200 m=s
The tidal frequency ftide is about 2 105 s1 , so the wavelength l can be
calculated by
c
l ¼
ftide
(4.33)
200
m 104 km
2 105
That is, the wavelength is about twice the width of the Atlantic and so resonance
occurs. In addition, the amplitude builds up as the time taken for the water wave to
make the round trip, reflecting off both shores, is about the same as the tidal period.
The wave amplitude also increases on the continental shelf and can reach about
3 m at the shores. River estuaries can also exhibit large tidal resonance if the length
and depth of the estuary are favourable. From (4.31), the time taken for a wave to
propagate the length of the channel, L, and back to the inlet is given by
2L pffiffiffiffiffi
t¼ ¼ 2L= gh (4.34)
c
If this time is equal to half the time between successive tides then the tidal
range is doubled.
4.4 Tidal energy extraction

Two primary methods to extract energy from the tides are (i) estuaries into which
large amounts of ocean water flows in due to high tidal range, water is captured
behind barrages and the turbines are rotated by exploiting the potential energy of
the stored water, and (ii) the kinetic energy of moving water can be used to extract
energy like the principle of extraction of wind energy. Each method has its own
advantages and disadvantages. It may be also possible to employ pumping strate-
gies for barrages to obtain better efficiency and to match electricity demand.
Accordingly, the devices used in energy generation are classified into three
types: (i) tidal barrages that store tidal flow and generate power through discharge,
(ii) tidal stream devices such as (a) current devices, which are fixed or moored
within a tidal stream, which is formed by the horizontal flowing volumes of water
or (b) fences that block a passage and extract energy in either or both directions of
tidal flow.
4.4.1 Tidal barrage

A tidal barrage is a type of tidal power generation that involves the construction of
a low dam wall across the mouth of the estuary through which the water flows in
and out of the basin. This dam has many underwater tunnels cut into its width
allowing seawater to flow through them in a controllable way using sluice gates.
The water flows into the bay during high tide and is retained by closing the sluice
gates at the beginning of low tide. The barrage gates are controlled according to the
tidal range of the location and operating it at right times of the tidal cycle. As the
tide ebbs and flows, seawater can flow in or out of the basin through a one-way
underwater tunnel system. This flow of tidal water back and forth causes the water
turbine generators located at the sluice gates to rotate producing tidal energy with
special generators used to produce electricity on both the incoming and the out-
going tides (Figure 4.9).
The main disadvantage of tidal barrage generation is that it can only generate
electricity when the tide is flowing either ‘in’ or ‘out’ as during high and low tide
times the tidal water is stationary. However, because tides are totally predictable,
other power stations can compensate for this stationary period when there is no
Sluice gates
Road
High tide level
Tidal Tidal
basin Head
height barrage
Low tide
Incoming tide Outgoing tide

Turbine tunnel
Figure 4.9 Schematic diagram of the tidal barrage

Open gate Sea level rising
Basin level
rising
Open gate
Closed gate Head
Turbine
Ebb
Basin Dam Ocean Dam Ocean
Basin
(a) (b)
Figure 4.10 Ebb generation: (a) filling the tidal basin and (b) generating
electricity
Sluice
gates
High Head height
closed
tide sufficient
Basin C
level B Head
Ebb Basin Sluice gates
Water level
tide empties closed

A
Time
A
Basin Basin
fills up Flood D fills up
tide E
Head height
too small
Low tide
Generation time
Figure 4.11 Ebb generation cycle
tidal energy being produced. Another disadvantage of the tidal barrage system is
the environmental and ecological effects that long concrete dams may have on the
estuaries they span. Different schemes of the tidal barrage system can be applied
for power generation as explained in the forthcoming subsections.
4.4.1.1 Ebb generation

Tidal barrage ebb generation uses the energy of an outgoing or falling tide. The
tidal basin is filled through the sluices until high tide. Then the sluice gates are
closed and at this stage pumping may be used to raise the level further. As shown in
Figure 4.10(a), the turbine gates are kept closed until the sea level falls to create
sufficient head across the barrage, and then are opened so that the turbines generate
until the head is again low (Figure 4.10(b)). Then the sluices are opened, turbines
are disconnected and the basin is filled again, and the cycle repeats itself as shown
in Figure 4.11. When there is sufficient head height between two sides of the
barrage, the sluice gates are opened at point C releasing the trapped water back to
the sea and power generation starts. The tidal basin continues to empty driving the
tidal generators until the level of water in the basin reaches point D. At point D, the
head height of the seawater across the barrage is no longer adequate to drive
the turbines so the sluice gates are closed, and generation stops until the level of the
seawater reaches point E again and the whole process starts over.
Ebb generation is also known as ‘outflow generation’ as the generation takes
place when the tidal direction changes [11].
4.4.1.2 Flood generation

Tidal barrage flood generation uses the energy of an incoming rising tide as it
moves towards the land. The water is filled into the basin through the turbines
which will generate once at tide flood. The tidal basin is emptied through sluice
gates or lock gates located along a section of the barrage, and at low tide, the basin
is affectively empty. As the tide turns and starts to come in, the sluice gates are
closed and the barrage holds back the rising sea level, creating a difference in
height between the levels of water on either side of the barrage dam.
The sluice gates to the entrances to the dam’s tunnels can either be closed as
the seawater rises to allow for a sufficient head of water to develop between the sea
level and the basin level before being opened generating more kinetic energy as the
water rushes through, turning the turbines as it passes. Or may remain fully open,
filling up the basin more slowly and maintaining the same water level inside the
basin as out in the sea. The tidal basin is therefore filled up through the turbine
tunnels which spin the turbines generating tidal electricity on the flood tide and is
then emptied through the opened sluice or lock gates on the ebb tide. Then a flood
tidal barrage scheme is a one-way tidal generation scheme on the incoming tide
with tidal generation restricted to about 6 h per tidal cycle as the basin fills
up. Once the tidal basin is full of water at high tide, all the sluice gates are opened
allowing all the trapped water behind the dam to return to the ocean or sea as it
ebbs away.
The movement of the water through the tunnels as the tidal basin fills up can be
a slow process, so low-speed turbines are used to generate the electrical power.
Generally, this is not efficient if compared to that of ebb generation as the
basin’s upper half has more volume than the lower half, during the flood genera-
tion, if filled thoroughly. The utmost importance for the turbine to produce power is
the level difference that is available between the barrage’s sea side and the basin
side, and it will reduce faster. Rather than ebb generation enhancement, water from
the rivers flowing into the basin will reduce the potential of the energy [12].
4.4.1.3 Pumping at high tide

Pumping at or soon after high tide is like pump-storage generation for ebb-
generation scheme. When the sluice gates are being closed, the turbines are oper-
ated in reverse so that they act as pumps moving water from the sea into the basin.
Pumping stops when the sea has ebbed enough to increase the head across the
turbines to an uneconomical height. With more water stored in the basin, generation
will start earlier than would be the case for simple ebb generation scheme. The
additional water in the basin is released through the turbines at the higher head than
that through which it was pumped, so that there should be a net gain in energy. The
turbine operated as a pump is supplied by the grid energy to raise the basin’s level
of water. The energy produced during generation is more than is returned; this is
because the output power is strongly related to the head, as well as the correlation
between the potential energy is not in a linear relationship, but rather is related to
the square of the tidal height variation.
4.4.1.4 Two-way generation scheme

It is seen that both ebb tidal barrage and flood tidal barrage installations are ‘one-
way’ tidal generation schemes. So, to increase the power generation time and thus
improve the efficiency, special double effect turbines that generate power in both
directions, tide ebbs and flows, should be used. Two-way electrical generation
requires a more accurate control of the sluice gates keeping them closed until the
differential head height is sufficient in either direction before being opened. As the
tide ebbs and flows, seawater flows in or out of the tidal reservoir through the same
gate system. This flow of tidal water back and forth causes the turbine generators
located within the tunnel to rotate in both directions producing electricity.
However, this two-way generation is in general less efficient than one-way
flood or ebb generation as the required head height is much smaller which reduces
the period over which normal one-way generation might have otherwise occurred.
Also, bi-directional tidal turbine generators designed to operate in both directions
are generally more expensive and less efficient than dedicated unidirectional tidal
generators.
One way of improving the operating time and efficiency of a two-way tidal
barrage scheme is to use individual one-way unidirectional tidal turbines installed
along the barrage. By controlling their individual sluice gates, one set of turbines
can be made to operate on the flood tide between points E and B and the other set
operate on the ebb tide between points C and D in Figure 4.11. While this method
increases the total number of tidal turbines located along the tidal barrage, it has the
advantage that the generation period is greatly extended.
In practice, this scheme that is only used for spring tides, affords only a slight
increase in output, though it does make it possible to alter the pattern of distribution
throughout the day. In addition, double action mode does not have any effect on the
output variations produced by the fortnightly tide cycle [4].
Investment in double-action capability would only appear to be economically
warranted in situations where there is a high tidal range and where accurate long-
term forecasts are available as to relative energy values at different times of day.
Efficient use of two-way scheme operation requires accurate information, several
days in advance, on energy values throughout the day.
It is to be noted that the different schemes of tidal barrage system can use
double-basin concept instead of one basin. Here the turbines are located on a dyke
separating two basins: (i) upper or high-level basin ‘A’, which fills at high water,
and (ii) lower or low-level basin ‘B’, which empties at low water as shown in
Upper basin
Seawater ‘A’
Sluice gate
Turbine
Return water
Sluice gate ‘B’
Lower basin
Figure 4.12 Two-basin system
Figure 4.12. At the beginning of the flood tide, the turbines are shut down. Basin A
fills and B remains empty. As soon as the head difference of A and B is large
enough the water flows from A to B and the turbines start. The separation can be
controlled to continuously operate from upper to lower basin. However, as water
head between basin varies during each tidal cycle, power generation can improve
situation in using off peak power from tidal generator or from alternate system to
pump water from low to high basin, and increased head is available from tidal
power at peak times.
4.4.1.5 Energy calculations

The energy per tide can be obtained by considering the following: (i) the water is
trapped at high tide in the basin and allowed to run out through the turbine, (ii) the
basin has a constant surface area A, which remains covered in water at low tide, and
(iii) the trapped water has a mass of rAR at a centre of gravity R/2 above the low-
tide level. Therefore, the potential energy available per tide when all the water falls
through R/2 is
R
Energy per tide ¼ ðrARÞg (4.35)
2
neglecting the small changes in density from the value of the seawater and taking
r ¼ 1:025 103 kg=m3 .
The average potential power P for one tidal period, t, becomes
rAg 2
P¼ R (4.36)
2t
The range R through the month varies sinusoidally from the highest value
Rspring for the spring tides to the lowest for neap tides, Rneap, with a period of half
the lunar month as explained in Section 4.1.2.1 (Figure 4.13).
At any time, t, after a mean high tide within the lunar month of period
T ¼ 29:53 days, the range is given by

R Rspring þ Rneap Rspring Rneap
¼ þ sin ð4pt=T Þ (4.37)
2 4 4
7 days
(Rspring – Rneap)/4 Mean high tide

Rspring/2
Rneap/2 (Rspring + Rneap)/4
Mean sea level
Figure 4.13 Sinusoidal variation of tidal range
Hence,
Rspring
R¼ ½ð1 þ aÞ þ ð1 aÞ sin ð4pt=T Þ (4.38)
2
where Rneap ¼ aRspring :
The power can be obtained from the mean square range, R2 ; as
ÐT
R2spring t¼0 ½ð1 þ aÞ þ ð1 aÞ sinð4pt=T Þ2 dt
R ¼
2
ÐT (4.39)
4 dt t¼0
Thus,
R2spring
R2 ¼ 3 þ 2a þ 3a2 (4.40)
8
The mean power produced over the month is
rAg R2spring
P month ¼ 3 þ 2a þ 3a2 (4.41)
2t 8
where t is the intertidal period.
The value of a is about 0.5, and so, (4.41) differs little from the approximations
that are commonly used in the literature by either
rAg 2
P R (4.42)
2t
where R is the mean range of all tides, or

rAg R2max þ R2min
P (4.43)
2t 2
where Rmax and Rmin are the maximum and minimum ranges.
Example 4.1 Typical values of mean tidal power are given as Rspring ¼ 5 m ¼ Rmax ,
Rneap ¼ 2:5 m ¼ Rmin ; A ¼ 10 km2 , r ¼ 1:025 103 kg=m3 , g ¼ 9:81 m=s2 and
t ¼ 4:47 104 s ( 12 h and 25 min). Therefore, a ¼ 0:5 and R ¼ 5 þ2 2:5 ¼ 3:75 m.
6
Applying (7.40) gives R2 ¼ 1:84, rAg
2t ¼ 1:125 10 .
Then, P can be calculated as given by (4.41), (4.42) and (4.43) to see the
difference that results from the approximation as follows:
From (4.41),
P ¼ 16:70 MW
From (4.42),
P ¼ 15:82 MW
From (4.43),
P ¼ 17:58 MW
It is to be noted that the value resulted from (4.41) is the average value of those
resulted from (4.42) and (4.43) as approximated values.
4.4.2 Tidal stream

The tidal stream generation (TSG) system exploits the major tidal flows and ocean
currents (e.g. gulf stream) to extract its tidal energy using turbine generators
beneath the surface of the water. In principal, TSG looks like wind power genera-
tion, except that the water currents flow across turbine rotor blades rather than wind
currents turning the blades for wind power turbines. So, TSG areas on the sea can
look just like underwater wind farms.
TSG system includes a turbine-generator set as shown in Figure 4.14(a). The
ocean’s currents turn the turbine blades, powering the generator, and then elec-
tricity can be transmitted to the shore through submarine cables. Tidal turbines can
range from about 5 to 30 m in diameter and are designed to swivel 180 degrees or
reverse rotation as tides ebb and flood. The stream generator, often referred to as a
tidal energy converter (TEC), is the machine that converts the kinetic energy pro-
vided by the turbine into electrical energy. TSG system may include several
turbine-generator sets as shown in Figure 4.14(b). The set can be supported on the
ocean floor by either tripod or concrete block or sunken pile.
Unlike offshore wind power, which can suffer from storms or heavy sea
damage, tidal stream turbines operate below the sea surface or are fixed to the
seabed. Tidal streams are formed by the horizontal fast-flowing volumes of water
caused by the ebb and flow of the tide as the profile of the seabed causes the water
to speed up as it approaches the shoreline.
Tidal streams’ turbines have much smaller diameters and higher tip speed rates
than those of wind turbine because the water is denser than air (1,000 times greater than
air) and has a much slower flow rate. They generate tidal power on both the ebb and
Sea level
Sea level
Tidal
stream
Current
Concrete block
Sunken Gravity
Tripod pile
Ocean floor
Seabed
(a) (b)
Figure 4.14 Tidal stream generators: (a) single turbine generator and (b) set of
multi-turbine generator
flow of the tide. One of the disadvantages of TSG is that as the turbines are submerged
under the surface of the water they create hazards to navigation and shipping.
Other forms of tidal energy include tidal fences, which use individual vertical-
axis turbines mounted within a fence structure, known as the caisson, to completely
block a channel and force water through them. An alternative way of harnessing
tidal power is using an oscillating tidal turbine by which the tidal stream flowing
past it causes its giant wing to oscillate up and down with the movement of the tidal
current. This motion is used to generate electricity.
4.4.2.1 Tidal energy calculation

Tidal energy is the energy due to the tidal stream (kinetic energy). It is also a fact
that the increase in tidal stream energy yields, to a large extent, an increase in
energy extraction [13,14].
Various turbine designs have varying efficiencies and, therefore, varying
power output. If the efficiency of the turbine, x; is known, the power output of a
turbine is given by
rAv3
P¼x kW (4.44)
2
where r ¼ the density of the water (1,025 kg/m3), A ¼ the sweep area of the turbine
(m2) and v ¼ the velocity of the flow (m/s).
Tidal velocities vary with time and can approximately be determined by [5]
¼ vmax sin ð2pt=tÞ (4.45)
where vmax is the maximum speed of the current and t is the period of the natural
tide (12 h and 25 min for a semidiurnal tide).
In practice, x may approach a maximum of 40%. Thus, the average electrical

power output is
ð t=4
sin3 ð2pt=tÞdt
0:4
P rAv3max t¼0 ð t=4
2 (4.46)
dt
t¼0
0:1rAv3max kW
Example 4.2 For a tidal current maximum velocity of 4 m/s and a sweep area of a
circle with a radius of 5.6 m, the power output of the wind turbine can be calculated
as follows:
The sweep area is

A ¼ pð5:6Þ2 100 m2
From (4.44), the power is
1;025 102 ð4Þ2

P ¼ 0:4 1:3 MW
2
From (4.46), the average power is P 0:1 1;025 102 43 6:6 MW.
4.4.2.2 Resource assessment

It is important to recognize how best to estimate the resource of a site. The real
amount of power generated will depend on the hydrodynamic and other char-
acteristics of the turbine used. It is worthy to make an upper estimate of the
resource having a detailed knowledge of tidal design.
It is found that early assessment of tidal energy in a channel focuses on cal-
culations using kinetic energy flux model. It is, of course, a useful measure but has
been shown not to be related to the maximum extractable power. Physically, sig-
nificant complicated limitations of tidal power generation exist. For instance, the
maximum possible energy extraction from a strait connecting two large basins is
given to within 10% by [11]
P ¼ 0:22rgDHmax Qmax (4.47)
where DHmax is the maximum differential water surface elevation across the channel
and Qmax is the maximum volumetric flow rate through the channel.
To understand this limit, considering the absence of turbines retarding the flow
in a site, the current is a maximum, but no power can be generated as no work is
done. Meanwhile, for a site so blocked by turbines that the flow is small, the
extracted power is going to be low. This yields the existence of an optimal condi-
tion in between, at which the power is maximum. In practice, the maximum
extractable power is likely to be significantly higher than that can be realized from
Tidal current Kinetic Electrical To the grid

energy energy energy
Tidal current Electrical Power

turbine generator converters
Figure 4.15 Typical structure of tidal current energy conversion system
a site because of two key reasons: (i) large losses from the wake mixing behind
turbines, from the turbine support structures, in hydrodynamic inefficiency in the
turbine, and in generator and transmission losses; (ii) in case more rows of turbines
are placed at a site, there is a diminishing return – that is, when adding a second row
of turbines in addition to the first one, it will not double the power but will lead to
both rows of turbines producing less than the original one row [15].
4.5 Tidal current energy conversion devices
A typical structure of tidal energy conversion system is shown in Figure 4.15 as a

chain of the devices: turbine, generator and converters. A brief description of both
is in the next subsections, and more details are explained in Chapter 2, Section 2.3,
for turbines and Section 2.4 for generators.
4.5.1 Tidal current turbines

The tidal current turbine is a device used to harness the kinetic energy in a tidal
form and then transforming the energy into mechanical power, which can then
drive a generator. The technology of tidal current turbine design to harness tidal
energy is the same as that used to design many wind turbines to harness the wind
power as both tidal current and wind are in the form of fluid. Most tidal turbines
can be characterized as belonging to three fundamental types [16–19].
4.5.1.1 Horizontal axis turbines

This type of turbine rotates around a horizontal axis which is parallel to the current
stream and is the most common type used to date. To generate greater starting torque
and reduce balancing problems encountered with single-blade turbines, multi-bladed
turbines are favoured. The number of blades should be decided with taking into
consideration that hydrodynamic losses are increased with greater number of blades.
Fixed pitch or variable pitch blades can be used depending on the turbine design.
4.5.1.2 Vertical axis turbines

The rotation speed of both vertical axis turbines and horizontal axis is very low
(15 rpm). However, vertical axis turbines can transmit the torque directly to the
water surface without the need of complex transmission systems or underwater
nacelle and they permit the harnessing of tidal flow from any direction.
4.5.1.3 Oscillating hydrofoil turbines

This kind of turbine induces hydrodynamic lift and drag forces due to a pressure
difference on the foil section caused by motion of the tidal current. Consequently, a
resultant tangential force to hydraulic fluid to drive an electrical generator is pro-
vided. It has been found that linear generators are the superior power take-off choice
for this kind of turbine, which is normally used for wave energy applications.
4.5.2 Tidal current generators

The generator is the device that converts the input mechanical power delivered by
the turbine through its shaft into electric power. As the tidal current turbine
topologies are like wind turbine, many wind turbine generator topologies can be
used for tidal current turbines.
4.5.2.1 Squirrel cage and wound rotor induction generators

The two types of induction generators, squirrel cage (SCIG) and wound rotor
(WRIG), are widely used as they are relatively inexpensive, robust and require low
maintenance. SCIG is mounted on the same shaft of the turbine through a multiple-
stage gearbox and is connected to the grid through a full-scale power converter and
a transformer. WRIG has a similar topology but with controllable rotor resistance.
4.5.2.2 Doubly fed induction generator

It has a stator directly connected to the grid through a transformer, whereas the
wound rotor is connected through a power electronic converter. The rating of the
converter is in the range of 25%–30% of the generator capacity and the generator
can operate in a wide range of speed variation (e.g. 30%) around the synchronous
speed. So, doubly fed induction generator (DFIG) is promising and widely used
from the economic point of view for wind integration but it is probably not the case
in tidal turbine applications in comparison with the permanent magnet synchronous
generator (PMSG).
4.5.2.3 Permanent magnet synchronous generator

It is commonly used in direct drive train with full-scale power converter connected
to the grid. PMSG has high potential for the tidal current turbines because of its
reduced failure, increased energy yield and reliability. It is favourable for oscil-
lating hydrofoil systems with horizontal turbines.
4.5.2.4 Direct drive permanent magnet generator

The direct drive synchronous generator with permanent magnet (DDPM) is the
most attractive because it has the highest yield of energy even if it is more
expensive than the generator systems with gearbox. There are many different
alternatives for the design of DDPM as suggested for tidal current turbine and wind
turbine as well. Basically, DDPM can be divided into radial-flux machine, axial-
flux machine and transverse flux machine according to the flux direction in the air
gap [20].
4.5.3 Drawbacks of tidal stream generators

Because of high start-up and construction costs due to limited experience of
installing, operating and maintaining plants, contractors’ perceptions of risk are
likely to be reflected in higher costs.
Like the wind, waves and tidal streams are variable RESs. Their intermittent
generation has implications for large-scale grid integration.
Because the amount of energy of the tidal streams naturally varies over time, the
power output of wave energy converters and tidal stream generators will also vary.
This has implications for grid integration, particularly balancing supply and demand.
4.6 Tidal power plants connected to utility grid

Traditionally, to facilitate power flow from the generating stations to the load centres
through the transmission system, the electrical networks are designed and operated in
such a way that both frequency and voltage are fixed. Generation technologies such as
hydro, thermal and nuclear are supplied with synchronous generators to produce bulk
power at constant frequency and constant voltage by monitoring and controlling the
active power and reactive power, respectively, in a reliable and cost-effective manner.
Renewable energy plants are deployed in areas in which the network avail-
ability is limited (e.g. remote areas) or to assist the power network for facing the
demand increase in addition to the other benefits of green energy sources.
However, tidal power plants can be developed in areas with network accessibility
as most of the high-energy tidal channels are in naturally articulated channels
already being used by shipping, navigation and neighbouring populated areas.
In the case of grid integration, some issues should be investigated, such as
capacity limits, power quality and system analysis [21].
4.6.1 Capacity limits

For areas with limited network strength, reinforcement of transmission line thermal
rating and fault protection mechanisms may be needed, as well as limitations in
transmission capacity because of the technical and cost issues associated with it.
However, high currents may flow into the distribution network and its feeders in
fault conditions, and that makes the planners to pay more attention to the switch-
gear components.
4.6.2 Power quality

Tidal generation source behaves as a time-varying generation source, which in turn
leads to a complex voltage profile in the steady state, and flicker because of resource
intermittency, start-up and shutdown conditions.
Tidal power generation systems with induction machines may consume sig-
nificant amount of reactive power. Therefore, compensation techniques may be
essential to reduce system losses and enhance voltage stability. In switching con-
ditions such as sudden disconnection of generators when fault occurs or turn on/off
actions of intermittent energy converters coupled with induction machines and

transformers, it may cause large, sudden and sustained alteration of network voltage.
Another important aspect of power quality is ‘Harmonics’, which may be
injected into the network in some cases such as use of PM generators through
direct-drive or gear-coupled arrangements and power electronic systems (full-sized
for PM generators and partially rated for DFIGs) used in grid-interfacing schemes.
4.6.3 System analysis

To analyse a system incorporating tidal energy sources, load flow studies are
required prior to integration to a network. That necessitates the availability of
adequate and accurate system performance data for applying numerical modelling.
Typical with any other generation or load condition, tidal energy systems are
exposed to faults or failures in system elements (e.g. internal connection circuits,
site-to-shore cables and onshore network supplying power to the grid) due to harsh
sea conditions, human activities (e.g. fishing, anchors and leisure activities) and
short-circuit cables. For short-circuit calculations with a goal of protection design
and service personnel safety, grounding is an important issue.
For tidal energy systems planning, some studies are essential such as (i) sta-
bility (transient stability, small signal stability, voltage stability and frequency
stability), (ii) scheduling and unit commitment to maintain the balance between
electricity demand and supply and (iii) load forecasting [22]. Therefore, dynamic
model development and statistical data must be determined.
Intermittency and variability of tidal current systems can be mitigated through
long-term storage techniques, which have high cost at present.
4.7 Environmental impacts
Tidal energy as a renewable resource does not result in any emissions. So, it is
beneficial to the environment because its use reduces the demand on fuels (coal and
oil) and it is attractive to many because of its reliance on existing technology and
knowledge. In contrast, some of the non-monetary costs associated with barrage
systems include destruction of habitat, interruption of organisms’ travel routes and
potential both electromagnetic interference and acoustic pollution. Moreover,
hydroelectric systems contribute to killing fish, and the waste heat that ends up in
the water reduces its capacity to store dissolved oxygen, harming not only fish but
all organisms in the affected area. Tidal energy systems also lead to a reduction in
the area of intertidal foreshore, which affects feeding habits of the large numbers of
wading birds that use the estuary [23].
4.8 Tidal power pros and cons

Though the development of the tidal power is at the nascent stage, it is estimated to
have a vast potential globally. It is generated from converting energy from the force
of tides into power. Tidal energy production is more predictable compared to wind
energy and solar power. Despite this, tidal power is still not exploited widely even
though the world’s first large-scale tidal power plant became operational in 1966.
However, increased focus on generating power from renewable sources is expected
to accelerate the development of new methods to exploit tidal energy.
4.8.1 Pros of tidal power

Renewable: As it is derived from tidal waves, which result from the gravitational
pulls from the sun and the moon, and the earth’s rotation around its axis, the energy
is classified as a renewable source. Compared to limitations in power generation
from finite fossil fuel reserves, tidal energy does not have any such constraints.
Taking this into account, the tidal energy generation may last for thousands of years
as long as the tides remain.
Environmentally friendly: As the tidal power generation emits no harmful
gases into the atmosphere, it is an environmentally friendly energy source. Further,
the tidal power projects do not need much space compared to solar farms which
require a large area of land for installation of modules. Tidal power installations
also occupy less space compared to other modes of power generation.
Predictability: As it is easy to know the occurrence of tidal currents, the scope
of a tidal energy project can be measured accurately. The construction of the pro-
jects can also be carried out with correct dimensions based on the formation of high
and low tides in certain cycles. Assessment of these factors can help easily deter-
mine the size and capacity of a tidal plant at an area.
Cost-effective: Given the longer lifespan of tidal power plants compared to
those that depend on fossil fuels, they are considered as cost-effective. Also, the
tidal plants can be effective at low tidal speeds and produce power to bring down
the overall cost.
4.8.2 Cons of tidal power

Site availability constraints: Limited availability of sites has remained a major
constraint in the development of tidal power projects. The limitation has restricted
the growth of tidal power projects worldwide. However, recent innovations in
design and turbine technology for tidal power are expected to offer a potential to
increase the installed capacity of tidal plants. The technological developments are
also expected to result in reducing economic and environmental costs of the pro-
jects, thereby increasing their availability.
Sites close to land: Tidal power plants must be constructed near to the land, as
high tidal currents are needed to produce power. A huge variation in tides and tidal
velocities can significantly increase the potential of a site for tidal electricity gen-
eration. Though the places where there is variation in tides offer greater potential
for tidal projects, scientists are trying to develop technology which can generate
enough energy even in low-tide environment.
Environmental effects: The impact of tidal power projects on the environment
is yet to be ascertained precisely. Tidal power can damage marine life, as tidal
turbines with their rotating blades may lead to death of living creatures in a sea.
Noise from the rotation of the turbines may also impact fish habitations in the tidal
power locations. Tidal energy can also impact the quality of water and sediment
processes.
Tidal barrage: Installing a tidal barrage may result in a change to shoreline
within a bay, damaging an ecosystem that relies on tidal flats. Flushing of the bay
may also be affected by the prevention of the flow of water in and out of the bay.
As a result, there may be reduction in saltwater, impacting marine life.
References
[1] Schatz D., Allan P. and Cooper D. The moon and its phases: Modeling moon
phases [online]. 2003. The Pacific Science Center. Available from https://
www.southampton.ac.uk/~sj2y10/moonPhasesSheetA3.pdf [Accessed 2 Feb
2020].
[2] Richardson C. J. The ‘ins and outs’ of tides and their use as a renewable
energy source [online]. 2015. Institute of Physics – Environmental Physics
Essay Competition. Available from https://www.iop.org/activity/groups/
subject/env/prize/winners/file_65752.pdf [Accessed 2 Feb 2020].
[3] Scherrer D. Teaching moon phases. 2015. Stanford Solar Centre. Available
from http://solar-center.stanford.edu/activities/MoonPhases/Teaching-Moon-
Phases.pdf [Accessed 2 Feb 2020].
[4] Frau J. P. ‘Tidal energy: Promising projects’. IEEE Transactions on Energy
Conversion. 1993;8(3):552–8.
[5] Blokdyk G. Tidal power [online]. E-book, Chapter 13, pp. 429–52. Available
from http://documents.routledge-interactive.s3. amazonaws.com/978041558
4388/S12-1_(RER2e-ch13-tidal_power).pdf [Accessed 2 Feb 2020].
[6] Bhatia S. C. Advanced renewable energy systems. New Delhi: Woodhead
Publishing India PVT Ltd.; 2014.
[7] Kawalik Z. and Luick J. The oceanography of tides: Tidal forces. Alaska:
University of Alaska Fairbanks; 2013, Chapter 1, pp. 5–40.
[8] Micks S. D. Understanding tides. US Dept. of Commerce: National Oceanic
and Atmospheric Administration (NOAA); 2006, Chapters 5–9, pp. 21–50.
[9] Gjevik B. Lectures on tides. Oslo: University of Oslo, UNIS Longyearbyn;
2011.
[10] Aly H. H. H. ‘Dynamic modeling and control of the tidal current turbine
using DFIG and DDPMSG for power system stability analysis’. Electrical
Power and Energy Systems. 2016;83:525–40.
[11] Tousif S. R. and Taslim B. S. ‘Tidal power: An effective method of gen-
erating power’. International Journal of Scientific and Engineering
Research. 2011;2(5):1–5.
[12] Bohari Z. H., Sulaima M. F., Baharom, M. F. and Bukhari W. M. ‘Are tidal
power generation as the future generation for Malaysian climate and
location: A technical review?’. International Journal of Applied Engineering

Research. 2016;11(10):7095–9.
[13] Vikas M., Ran S. and Seelam J. K. (eds.). ‘Tidal energy: A review’.
Proceedings of International Conference on Hydraulic, Water Resources
and Coastal Engineering (Hudro2016); Pune, India, Dec 2016, pp. 2320–8.
[14] Al Yusuf S., Gupta S. D., Ali T., Arnab I. Z. and Begum J. ‘An introduction
to tidal power generation and analyzing the potentiality in Bangladesh’.
Sustainable Energy Technologies (ICSET), 2012 IEEE 3rd Int. Conf. on;
Kathmandu, Sep 2012, pp. 316–20.
[15] Adcock T. A. A., Draper S. and Nishino T. ‘Tidal power generation – A
review of hydrodynamic modelling’. Proc. IMechE Part A: Journal of
Power and Energy. 2015;229(7):755–71.
[16] Rourke F. O., Boyle F. and Reynolds A. ‘Marine current devices: Current
status and possible future applications in Ireland’. Renewable and
Sustainable Energy Reviews. 2010;14(3):1026–36.
[17] Zhang J., Moreau L., Machmoum M. and Guillern P. E. ‘State of the art in
tidal current energy extracting technologies’. First International Conference
on Green Energy ICGE. Sfax, Tunisia, March 2014, pp. 1–7.
[18] Mohammadi S. and Jam S. S. ‘Usage of tidal power as power station’.
Electrical Power Distribution Networks (EPDC), 2012, Proc. of 17th Conf.
on. Tehran, Iran, May 2012, pp. 1–7.
[19] Chen H., Aı̈t-Ahmed N., Zaı̈m E. H. and Machmoum M. ‘Marine tidal cur-
rent systems: State of the art’. Industrial Electronics (ISIE), 2012 IEEE Int.
Symposium on. Hangzhou, China, May 2012, pp. 1431–7.
[20] Chen Y., Pillay P. and Khan A. ‘PM wind generator topologies’. IEEE
Trans. on Industry Applications. 2005;41(6):1619–26.
[21] Khan J., Bhuyan G., Moshref A., Morison K., Pease Jr. J. H. and Gurney J.
‘Ocean wave and tidal current conversion technologies and their interaction
with electrical networks’. PES General Meeting, 2008 IEEE-Conversion and
Delivery of Electrical Energy in the 21st Century. Pittsburgh, PA, July 2008,
pp. 1–8.
[22] Jones A. T. and Westwood A. ‘Recent progress in offshore renewable energy
technology development’. PES General Meeting, 2005 IEEE. San Francisco,
USA, June 2005. IEEE 2017–22.
[23] Baker A. C. ‘Tidal power’. IEE Proceedings Pt. A. 1987;134(5):392–8.
Chapter 5
Ocean energy: wave and thermal energy
Useful electric energy can be derived from the ocean waves and is known as ‘ocean
wave energy’. Physics of waves and their different forms as well as estimation of
energy and its conversion into electrical energy are explained in the forthcoming
sections.
5.1 Wave energy concept
Wave energy is the energy contained in the waves that are caused by winds blowing
across the surface of the sea. Friction transfers some of the wind energy to the water
thus forming waves. This energy is stored in the waves as (i) potential energy in the
mass of water displaced from the mean sea level and (ii) kinetic energy in
the motion of water particles. Thus, the stronger the wind is, the more energy in the
waves and the larger the distance the waves travel. The wave energy is proportional
to the square of the height of the wave, which is inversely proportional to the depth
of the water. This means that the waves near the coast carry the most energy.
Wave energy is not tidal energy even though both fall under the category of
ocean energy. While the tidal energy consists of long period oscillations, sea waves
of short period are generated by the action of the wind. The total theoretical wave
energy resource is very high, roughly twice the global electrical energy consump-
tion as the oceans cover about 70% of the surface of the earth and are considered to
be one of the greatest sources of renewable energy. Therefore, it is worthy to
capture the wave energy to be transferred into useful energy, such as electricity.
The devices used to capture wave energy and convert it into electrical energy are
known as ‘wave energy converters’ (WECs).
5.2 Ocean waves creation
Radiation from the sun to the earth and ocean surface creates air circulation, which
is wind. When wind blows over the surface of the oceans, waves will be generated
by the effect of both gravity and friction (Figure 5.1(a)). Energy in the wind is
transferred to the surface water of the sea or ocean producing waves of size that
depends on the magnitude of that energy, which is a function of the wind speed, the
duration the wind blows and the area affected by the wind (known as ‘fetch’). The
Wind generation
Wind–sea interaction under the

influence of gravity
Wave
propagation
Seabed
(a)
Wave-formed Fully developed Changing to

ripples sea waves swell waves
Direction of wave advance

Wind
Fetch length = distance travelled

by wind across open water
(b)
Figure 5.1 Ocean waves: (a) wave creation and (b) fetch area created by
wind blows
fetch is specified by its length, where the fetch length, is expressed as the distance
travelled by the wind across open water. When the wind blows across the water
surface, wave-formed ripples or capillary waves may take place as well as con-
tinuous flows in the direction of wave advance. Thus, the wave is fully developed
and can change to swell (Figure 5.1(b)) [1].
5.2.1 Ripples
In sedimentology, wave-formed ripples or wave-formed ripple marks are a feature
of sediments (sandstones, limestone, siltstones) and dunes. These ripple marks are
often characterized (and thus distinguished from current ripples) by symmetric
Ocean energy: wave and thermal energy 131
Equal slopes
Waves
Figure 5.2 Symmetrical ripple
cross sections and long relatively straight crests, which may commonly bifurcate.
Commonly, these crests can be truncated by subsequent flows. Their wavelength
(periodicity) depends on the sediment grain size, water depth and water-particle
orbits in the waves. Symmetrical ripples are commonly found in shallow waters.
Beaches are a good place to find these ripples. A typical example is shown in
Figure 5.2.
Though wave-formed ripples are traditionally described as symmetrical,
asymmetric wave ripples are common in shallow waters along sandy shores. They
are produced by bottom oscillations generated by passing breaker waves, which
have unequal intensity in opposite directions [2].
Wave-formed ripples indicate an environment with weak currents where water
motion is dominated by wave oscillations [3]. However, ripples are relatively small
elongated ridges that form on bed surfaces perpendicular to current flow. With
continuous current flow in one direction, asymmetrical ripples form. Asymmetrical
ripples contain a steeper slope downstream. With an alternation in current flow
from one direction to the opposite, symmetrical ripples form. Symmetrical ripples
tend to have the same slope on both sides of the crest [4].
Capillary wave. It is a wave travelling along the phase boundary of a fluid
whose dynamics and phase velocity are dominated by the effects of surface tension.
The wavelength of capillary waves on water is typically less than a few centi-
metres, with a phase speed in excess of 0.2–0.3 m/s [5].
This type of wave is known as a gravity wave. In addition, the group velocity is
times the phase velocity as: (i) the wavelength is short and (ii) the phase velocity is
dominated by the surface tension effects. This type of wave is known as a capillary
wave (Figure 5.3).
Swell waves. A swell, in the context of an ocean, sea or lake, is a series of
mechanical waves that propagate along the interface between water and air and
thus are often referred to as surface gravity waves. Occasionally, swells which are
longer than 700 m occur as a result of severe storms (wind, tides and rainfall). A
swell has a period in the order of a fraction of a minute to several minutes. Due to
Traditional gravity waves
Capillary Deepwater waves Swell waves Peaking waves Breaking waves

waves
Nearby shore
1.74 cm 10 m
Figure 5.3 Capillary wave formation
Tide
Wind
Rainfall
Reef/
Swell
sand
Ocean floor
Figure 5.4 Swell wave formation
its long wavelength, this water wave travels very fast. It is also very rounded and
most people when first encountering it may not recognize this as a wave. Some call
it the hills and valleys of the sea. Swells need a large body of water to develop and
are usually, but not always, associated with the ocean. A typical graph of swell is
depicted in Figure 5.4.
A swell is a wave; so, often to prevent confusion, the shorter wavelength
variety is called chop or choppy waves. These waves have periods in the order of a
few seconds. Most waves are wind generated. As chop and swell are waves, they
can interact and pass through each other. Most waves are wind generated. Swells
move larger amounts of water, so it takes longer for the energy to build up and
dissipate. Swells also travel fast and for a long distance. For that reason, there is
often a swell present in many places of the ocean. Chop can whip up pretty quick
and die off just as fast. A variant of chop is a whitecap. A whitecap is when the
wind transfers more energy to a water wave than the wave can transmit, the top
may curl and blow in the wind making the top appear white. Ocean sea spray is the
result of whitecaps.
5.2.2 Shallow waves or shore waves

They are formulated when the wave energy interacts with the bottom or land
(Figure 5.5). However, there is an ocean wave called a breaker. Breakers strike fear
and concern in the heart of any wise seafarer. A breaker is a swell interfering with a
shallow bottom in normally deep water. The result is a localized, out of place, chop
that does not seem to travel on its way. It is an indication of shallow rocks, reefs or
a shallow submerged object.
Breaking waves. They are formed out in the open ocean and can travel vast
distances before breaking on a distant coastline (Figure 5.6). The energy carried by
these waves and the way they break against the shoreline has a dramatic impact on
erosion and how shorelines are shaped over time.
As waves approach the shore, the bottom of the wave meets the ocean floor. As
they drag across the bottom, the front waves slow down, and the wavelength is
reduced. The following waves start to build up behind the slow ones, and as the
wavelengths get shorter, the wave energy gets transferred upwards, increasing
wave height. The friction along the bottom slows down the base of the wave while
the water at the surface continues forward. When the wave steepness (wave height
to wavelength ratio) exceeds a ratio of 1:7, it becomes unstable and breaks at the
surf zone (Figure 5.7). Surf is the wave activity in the area between the shoreline
and the outer limit of breakers. It may refer to a breaking wave in shallow water,
upon the shore or in the area in which waves break.
Therefore, the slope of the sea floor greatly influences how waves break. A
physical graph of breaking waves is shown in Figure 5.8.
5.3 Wave propagation
As explained in Section 5.2, it can be concluded that along the wind direction
towards the shoreline, the form of water waves may change from one to another
depending on the depth of the seabed D. A typical example is shown in Figure 5.5.
At the full depth of water, the deepwater waves may swell and propagate as tran-
sitional waves. Decreasing water depth results in an increase of the wave peak
Shallow-water Transitional Deepwater

waves waves waves
Breaker Peaking waves
D < (1/20)λ Swell
D = ½λ
D > ½λ
Figure 5.5 Types of ocean waves in terms of water depth

Water particles, in Water particles, in Wave top is thrown forwards

deep water, move in shallow water, move in and break
circular motion elliptical motion
Figure 5.6 Breaking waves near the shore
Wave crowding Increase of wave height Surf zone
Wave-back
Decreased depth and moves faster than
water is pushed up the front. The
wave becomes
Water gets shallower
very high and
and waves slowdown
breaks
Figure 5.7 Breaking mechanism
(peaking waves). When approaching the shoreline, the increase of wave peak
continues and the wave form changes to breaking wave form, and then shallow-
water wave.
The deepwater wave (also called ‘surface wave’) is the most common form for
which the most wave energy devices are designed to extract energy. For the water
waves to behave as a deepwater wave, the mean depth of the seabed should be more
than about half the wavelength, l (i.e. D > l=2). The deepwater wave form can be
characterized by the parameters: crest and trough amplitudes, height (H), mean depth
of the water or still-water level (D) and wavelength (l) as illustrated in Figure 5.9.
As depicted in Figures 5.6 and 5.7, in a deepwater wave, the motion of parti-
cles is circular with an amplitude that decreases exponentially with depth and
7m
1m
Public Domain pdphoto.org / The COMET Program
Figure 5.8 Typical physical graph of breaking waves. Source: Earth and Ocean
Sciences (EOSC)
Crest
Wavelength λ
Height H
Still-water level
Mean depth of the

seabed D
Trough
Seabed
Figure 5.9 Parameters of deepwater wave
becomes ignored at D < l=2. The particle motion for shallow-water wave is
changed to be elliptical and the water movement occurs against the sea bottom
producing energy dissipation (Figure 5.10(a) and (b)).
Deepwater waves have their own properties to be considered for the mathe-
matical representation of these waves. They can be summarized as follows:
● Water on the surface remains on the surface.
● Surface waves are sets of continuous sine waves of varying wavelength, phase
and direction.
Elliptical motion Mean surface

of water position
Circular
λ/2π particles in
motion of
shallower water
water D
particles in
deepwater
wave D
(a)
Deepwater
wave energy
movement
Breaking
waves
Z Transition waves
Y
land
X
Shore
Surf
Circular motion of water particles
Elliptical
motion of
shallower
water particles
Shallower water depth, >λ/2,

shortens wavelength
(b)
Figure 5.10 Particle motion in water waves (a) in deepwater and shallow-water
waves and (b) representation of particle motion for different wave
forms in three-dimension (X ¼ direction of wave energy movement,
Y ¼ wave width, Z ¼ wave height)
● The motion of water particles is circular. The surface waves have a shape that
progresses as a moving wave, although the particles themselves do not pro-
gress and have no net motion.
● The amplitude of the surface wave depends mainly on the history of the wind
regimes above the surface and is slightly dependent on the wavelength, velo-
city and duration.
Direction of wave propagation
Wave surface tangent

a F2 ω
H = 2a
F1 F
F2 ω
F ω ω
F1 F2
F2
F
F1 F1
F
Surface particle Between crest At bottom of Between trough

At top of crest
location and trough trough and crest
Figure 5.11 Forces affecting surface water particles at different locations
● The amplitude of water particle motion decreases exponentially with the water
depth below the mean surface position; that is, it reduces to l/e at a depth of
l=2p (e ¼ the base of natural logarithms ¼ 2.72) as shown in Figure 5.10(a).
● The particle motion is negligible at depth of l=2 as its amplitude is less than
5% of that of the surface motion.
5.4 Fundamentals of deepwater waves

For deepwater waves, the water particles in the surface are affected by friction,
surface tension and inertial forces as well as forces of gravity (F1) and circular
motion (F2), which are the dominant forces as the other forces are relatively small.
The amplitude of the wave equals the radius of the surface particle circular motion,
a, and therefore, the height H is twice the amplitude, that is, H ¼ 2a, and w (radian
per second) is the angular velocity of the water particles.
As shown in Figure 5.11, along the direction of the wave motion the moving
shape results from the phase differences in the motion of successive particles of
water. As one particle in the crest drops to a lower position, another particle in a
forward position circles up to continue the crest shape and the forward motion of
the wave.
The surface water particle of mass m is subjected to a resultant force F of two
components: the gravitational force F1 and the centrifugal force F2. The water
surface takes up the position produced by this resultant force, that being perpen-
dicular to the tangent to the surface. For instance, F2 moves a particle at the top of a
crest upwards. The particle is dropping a moment later and the position in the crest
is taken by a neighbouring particle rotating with a delayed phase. At a particle
position between crest and trough nearby the still-water level, the surface orientates
perpendicular to the resultant force F. Similarly, for a water particle at the bottom
of the trough, the centrifugal force is downward and the resultant force is max-
imum; whereas at a position between trough and crest, the water particle is going to
complete a full cycle of its motion [6].
Consider a general position of a surface particle at a height of h above the

mean water level and the wave moving along X-direction !
as shown in! Figure 5.12.
2
The forces!affecting
!
that
!
particle can be expressed as F 1 ¼ maw , F2 ¼ mg and,
therefore, F ¼ F1 þ F2 in a direction perpendicular to the wave surface tangent.
Assuming that the particle is initially at time t ¼ 0, at the average water level,
then,
p
j¼ wt
2
The resultant acceleration g, as in Figure 5.12, can be calculated as follows:
2 2
g2 ¼ g þ aw2 cos j þ aw2 sin j
Hence,
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2aw2
g¼g 1þ cos j practically; aw2 g (5.1)
g
The vertical component of particle velocity equals the rate of change of the
height with respect to the time, that is, dh/dt, where
dh
¼ aw sin j ¼ aw cos wt (5.2)
dt
The negative sign in (5.2) indicates the opposition of rate of height change to
the positive direction of Z-axis.
dh aw2 sin j 1
The slope of surface tangent ¼ tan q ¼ ¼ ¼ aw2 sin j
dx g þ aw cos j g
2
(5.3)
Incorporating (5.2) and (5.3) can get
2
w x
h ¼ a sin wt (5.4)
g
Height above
mean water
level
θ
ωt
aω2 Resultant h Z
θ
acceleration X Wave
direction
g
Figure 5.12 Vector diagram for forces affecting a surface-water particle

But the general equation of wave motion having wavelength l and velocity c can be
written as

2p 2p
h ¼ a sin ðx ctÞ ¼ a sin x wt ¼ a sinðkx wtÞ (5.5)
l l
where k ¼ 2pl (it is defined as the wave number) and c ¼ the phase velocity of the
travelling wave.
From (5.4) and (5.5), it can be seen that
2pg
l¼ (5.6)
w2
The period of motion or the period of the wave, T, is the time taken for one full
wave or one wavelength to pass a fixed position.
sffiffiffiffiffiffiffiffi
2p 2pl
T¼ ¼ (5.7)
w g
The velocity of the particle at the wave crest, v, is given by

rffiffiffiffiffiffiffiffi
2pg
v ¼ aw ¼ a (5.8)
l
The wave surface velocity, c, which is called the phase velocity of the travel-
ling wave (made by the surface motion) in the X-direction is given by
sffiffiffiffiffiffiffiffi
l w wl l gT
c¼ ¼ ¼ ¼g ¼ (5.9)
T k 2p 2pg 2p
Hence, longer waves travel faster than the shorter ones. In addition, from (5.8)
and (5.9), it is to be noted that the particle velocity, v, is a function of the amplitude
a, whereas the phase velocity, c, does not depend on the amplitude. This means that
c does not pertain to v.
5.5 Estimation of wave energy

Consider a single regular deepwater wave with surface particles moving in circular
orbits of amplitude a, at varying phase, in the direction of propagation X
(Figure 5.13). In a vertical column, the amplitude decreases exponentially with the
increase of depth below the mean surface water level and becomes negligible at the
sea bottom. It means that the amplitude of oscillation of particles decreases as
moving into the depth of the sea (Figure 5.14). Therefore, the particle at mean
Y
Z
dth
it wi
Un
Figure 5.13 A wave of unit width and amplitude r moving along X-direction
dy ω Z Y
H = 2a
dz X
dx a
aekz = r
Figure 5.14 Elemental strip of water in a deep-water wave

position of z below the surface (i.e. in the negative direction of Z) moves in a circle
of radius r that is given by [6]
r ¼ aekz (5.10)
where k is the wave number, 2p=l. It is to be noted that the particle motion remains
circular if the seabed depth D > 0:5l.
As shown in Figure 5.14, a vertical column as an elemental strip of water
across the wavefront of dimension dx dy dz (lengthwidthheight) is
considered. The volume of the strip is
dV ¼ dx dy dz
Thus, the volume per unit width of the wave is given by
dV ¼ dx dz (5.11)
and the mass is
dm ¼ r dV ¼ r dx dz (5.12)
where r is the density of seawater.
Assuming EK is the kinetic energy of the total wave motion per unit length in
X-direction, the kinetic energy of particle of width dx is
dEK dx ¼ ½mv2 ¼ ½ðr dz dxÞr2 w2

That is,
dEK ¼ ½rr2 w2 dz (5.13)

where each elemental strip of water is in circular motion at constant angular
velocity w, radius of circular orbit r and angular velocity v ¼ rw.
Consider a moment in time when the element is at its mean position, and all
other elements in the column are moving vertically with the same phase in the
Z-direction. Incorporating (5.10) and (5.13) gets

dEK ¼ ½r a2 e2kz w2 dz
Integrating both sides and multiplying by dx, the total kinetic energy in the
column is
ð z¼0
1 a2 w2
EK dx ¼ ½ra2 w2 e2kz dz dx ¼ r dx
z¼1 4 k
By substituting the value of k ¼ 2p=l and the value of w as in (5.6) can obtain the
kinetic energy per unit length of wave per unit width of wavefront as
1 2pg l 1
EK ¼ ra2 ¼ ra2 g (5.14)
4 l 2p 4
For harmonic motions, the average kinetic and potential energies are equal,
and so the potential energy per unit length of deepwater wave per unit width EP is
given by
1
EP ¼ ra2 g (5.15)
4
Summation of (5.14) and (5.15) gives the total energy per unit length per unit
width of wavefront, which is also the total energy per unit area of surface, as
E ¼ EK þ EP ¼ ½ra2 g (5.16)
Equation (5.16) indicates that the energy within a wave is contained near the
surface and falls off sharply with depth.
The energy per unit width of front wave, per unit wavelength in the direction of
wave motion, El, is
El ¼ El ¼ ½ra2 gl (5.17)
Substituting the value of l given by (5.6) gets
ag 2
El ¼ pr (5.18)
w
and in terms of period of motion T given by (5.7), (5.18) can be rewritten as
r
El ¼ ðagT Þ2 (5.19)
4p
5.5.1 Derivation of power formulae

As written in (5.9), from the phase velocity, c ¼ wk ; and by the value of k and (5.6),
it can be found that
w2 ¼ kg (5.20)
The corresponding group velocity, cg, in deepwater wave is
@w g c
cg ¼ ¼ ¼ (5.21)
@k 2w 2
This velocity is defined as the velocity at which the unseen energy in the group
of waves is carried forward. The total energy (kinetic þ potential) per unit length
per unit width of wavefront is given by (5.16). Accordingly, the power carried in
such waves across a vertical plane, per unit width of wavefront, can be expressed as
1 gT
P ¼ E cg ¼ rga2 ¼ rg 2 TH 2 =32p W=m (5.22)
2 4p
In nature, waves are irregular and can be described by statistical models. If the
wave conditions are measured, over 20 min for example, the mean wave height,
Absorption Mechanical Generation

stage interface stage stage
Wave energy
Electrical Electrical
WECs Turbines
generators energy
Figure 5.15 Stages of wave power conversion
Hm, and the significant wave height, Hs, can be calculated. The significant wave
height is defined as the average of the highest 33% of the waves. Under such
circumstances, the wave power can be stated to be
P ¼ 550Hs2 Tz W=m (5.23)

where Tz ; the mean zero crossing period is the duration of the record divided by n,
where (n þ 1) is the number of upward crossings of the mean water level.
5.6 Wave energy converters

Wave energy should be extracted in a form adequate for transformation to electrical
energy. WECs are devices used for this purpose, that is, to convert the wave power
into electrical power. The conversion passes through three steps in cascade as
shown in Figure 5.15. First, absorption stage at which the wave energy is trans-
formed to energy in working fluids (different working fluids are used depending on
the type of the device) and that is the function of WECs. Second, at mechanical
interface stage, the working fluid energy is transformed to mechanical energy using
a motor or a turbine. Third and finally, which is the electricity generation stage, the
mechanical energy rotates an electric generator to produce electrical energy [7–9].
The efficiency of a WEC can be estimated as the ratio of power output, Pout, to
power input, Pin. Input power per unit width can be calculated using either (5.22) or
(5.23). Therefore, the output power from the device must be divided by the wave
width, w, to get a dimensionless ratio that expresses the efficiency, that is
Pout
h¼ (5.24)
wPin
In practice, estimation of capture wave width (may be called ‘absorption
width’) meets difficulties as the device absorbs power from a wavefront wider than
the device itself. So, it is better to define the device efficiency as the delivered
power from the device to the power incident immediately on the device, Pinc . Thus,
Pout
h¼ (5.25)
Pinc
It will be noted that the efficiency depends greatly on the wavelength, ampli-
tude of the oncoming waves and the characteristics of the device.
Design of WECs is based on a large variety of energy conversion techniques.
However, WECs are generally categorized by location, type and modes of
operation.
5.6.1 Location of WECs

There are three locations for WECs: onshore, nearshore and offshore.
Onshore. It means coastal areas in which the water depth is 10–15 m and the
maximum wave height is 7.8 m. Shoreline devices have the advantage of being
close to the utility network, are easy to maintain and as waves are attenuated as they
travel through shallow water, WECs have a reduced likelihood of being damaged in
extreme conditions. On the other hand, this leads to one of the disadvantages of
shore-mounted devices, as shallow water leads to lower wave power.
Nearshore. It means shallow water areas in which the water depth is 15–25 m
and the maximum wave height is 15.6 m. Nearshore devices are defined as devices
that are in relatively shallow water of depth less than one-quarter wavelength.
Devices in this location are often attached to the seabed, which gives a
suitable stationary base against oscillations produced by an oscillating bodywork.
Like shoreline devices, a disadvantage is that shallow water leads to waves with
reduced power, limiting the energy harvesting potential.
Offshore. It means deepwater areas in which there is no natural limitation for
both water depth and wave height. The water depth is always more than 50 m and
wave height can exceed 30 m. In some literature, the depth is defined by ‘greater
than 40 m’ and in others by ‘exceeding one third of the wavelength’. The advantage
of siting a WEC in this location is the capability of harvesting greater amounts of
energy because of the higher energy content in deepwater waves. On the contrary,
more difficulties for construction and maintenance of devices are present as well as
the likely extreme climatic conditions require a proper design of the device con-
struction with additional cost.
5.6.2 Types of WECs

There is a large variation in designs of WECs. However, there are three pre-
dominant types into which WECs are classified. Type classification is based on one
or more of variations that arise when a wave passes a stationary position: (i) change
of wave height at the surface, (ii) change of water pressure under the surface and
(iii) rotation of small volumes of water near the surface.
Point absorber. It is a device that possesses small dimensions where both
width and length are much less than the incident wavelength. So, wave direction is
not important as the device size is relatively small and it absorbs the energy from
all wave directions because of their small size compared to the wave length. This
type of device is a floating structure that will move up and down on the water
surface or is submerged below the surface relying on pressure differential. Such
devices have large capture width to extract power from the oncoming waves. They
are basically offshore oscillating bodies, either floating or fully submerged.
The point absorbers usually use translational motion, such as surge, sway and
heave, to produce electricity. Indeed, the energy will come from the motion of the
buoyant top relative to the base. If one considers a floating buoy which has a
generator at the bottom of the sea, all translational movements make a source for
power generation. The larger the buoy is, the more energy can be absorbed as the
exposed area to wave increases. However a larger buoy will cause an increase in
diffraction forces. So, to design a point absorber the key parameter is the desired
amount of produced electricity. Hence, it is required to configure the entire system
considering the dimensions, the mass and inertia of the system. An optimization
algorithm is used for a selected site with the given met-ocean data. There are many
examples of point absorbers, the most well known is the ocean power technology’s
(OPT) power buoy as shown in Figure 5.16.
Floating point absorbers are divided into two categories, depending on whether
the bottom end of the structure is fixed to the seabed or to a bottom-fixed structure,
or if it is allowed to be in motion. Currently, the major point absorbers with a non-
fixed bottom end include the power buoy as shown in Figure 5.17. This device
floats on the surface of the water, and is held in place by cables connected to the
seabed. The construction involves a cylindrical structure with one component
relatively immobile as the bottom end, and a second component with movement
driven by wave motion as the top end floating buoy inside a fixed cylinder [10].
The relative motion of the two components that is caused by the rising and
falling of the waves is used to drive electromechanical generators or hydraulic
energy converters. The electric power generated is transmitted to shore over a
submerged transmission line. A 150 kW buoy has a diameter of 11 m and is 44 m
tall, with approximately 10 m of the unit rising above the ocean surface.
Floating buoy
Incident wave
Fully submerged buoy
Sea floor
Figure 5.16 Point absorber device (OPT power buoy)

Buoy
Seawater
Spar
Base
Cable
Seabed
Figure 5.17 Power buoy with non-fixed bottom
Attenuators (line absorbers). This device is a floating device that operates

parallel to the wave direction and literally rides the waves. An example of this type
is the well-known ‘Pelamis wave power’, which is developed by ocean power
delivery. These devices capture energy from the relative motion of the two arms as
the wave passes them. More particularly, cylindrical steel sections are connected to
each other by two degrees of freedom. Each joint contains the complete power
conversion unit. Those joints are going to face both horizontal and vertical motions
that are accommodated by four hydraulic cylinders. They act as pumps, which drive
the fluid through a hydraulic motor, which in turn drives an electrical generator. In
addition, accumulators are used to decouple not only the primary circuit which is
the pump but also the secondary circuit which is the motor. This allows to regulate
the flow of fluid and thus produce a more constant output. These devices have to be
at around 50 m off the coast.
A Pelamis machine which has four 35 m long by 3.5 m diameter floating
cylindrical sections connected by three hinged joints is shown in Figure 5.18. The
Pelamis technology is intended for general deployment and operation offshore with
water depths of 50–70 m and with nominal wave power of 55 kW/m. The four
sections of the device articulate with the movement of the waves, each resisting
motion between this and the next section. In this way, as waves run down the length
of the device, since it is positioned parallel to the direction of wave travel, flexing
at the hinged joints due to wave action pumps pressurized oil to drive hydraulic
pumps and motors built into the joints. Electricity generated in each joint is
transmitted to shore by a subsea cable.
Terminators. Terminator devices have their principal axis parallel to the
wavefront, extend perpendicular to the predominant wave direction and capture or
Wave direction Motion-up
Motion-down
Seabed
Figure 5.18 Pelamis wave attenuator
Lin
ea
bso
rbe
r
or
nat
Point
mi
Ter
absorber
Figure 5.19 Point absorber, line absorber and terminator devices
reflect the power of the wave. These devices are typically installed onshore or
nearshore, and the device structure is typically fixed on the seabed or on the shore.
Placement of point absorber, line absorber and terminator devices with respect to
the water waves is shown in Figure 5.19.
5.6.3 Modes of operation

WECs can also be categorized by their mode of operation such as submerged
pressure differential device, oscillating wave surge converter, oscillating water
column (OWC) and overtopping device.
Submerged pressure differential. The device is a submerged point absorber.
The pressure difference between the crests and troughs of the wave passing above
the device is used to actuate an air-filled cylindrical chamber that contains two
parts: a seabed lower fixed chamber (called ‘basement’) and a moveable upper
cylinder (called ‘floater’). As a crest passes over the device, the water pressure
above the device compresses the air within the cylinder, moving the upper cylinder
down. As a trough passes over, the water pressure on the device reduces and the
upper cylinder rises (Figure 5.20). Consequently, this up and down motion can be
resisted by a linear electrical generator, with the interior air pressure acting as a
spring. Because of the fixed part being attached to the seabed, the device is located
nearshore. A typical example of this device is the Archimedes wave swing.
Oscillating wave surge converter. An oscillating wave surge converter is
generally comprised of a hinged deflector, positioned perpendicular to the wave
direction (a terminator), that moves back and forth exploiting the horizontal particle
velocity of the wave. An example is the Aquamarine Power Oyster, a nearshore
Particle at Particle at
wave crest wave trough
Water
wave
Floater
Seawater
Basement
Seabed
Figure 5.20 Submerged pressure differential device
device, where the top of the deflector is above the water surface and is hinged from
the seabed. A prototype of this device has been constructed as illustrated in
Figure 5.21.
Oscillating water column. An OWC consists of a chamber with an opening to
the sea below the waterline. As waves approach the device, water is forced into the
chamber, applying pressure on the air within the chamber. This air escapes to
atmosphere through a turbine. As the water retreats, air is then drawn in through the
turbine. A low-pressure Wells turbine is often used in this application as it rotates
in the same direction irrespective of the flow direction, removing the need to rectify
the air flow. It has been suggested that one of the advantages of the OWC concept
is its simplicity and robustness [12].
There are examples of OWCs as point absorbers, as well as that being built into
the shoreline, acting as a terminator. An example of a shoreline mounted device is
the Wavegen Limpet. The device is installed on the island of Islay, Western
Scotland, and produces power for the national grid. The design of the Limpet is
shown in Figure 5.22. The OWC concept has also been proposed by Oceanlinx, an
Australian wave energy developer, in a nearshore tethered device [13].
Overtopping device. An overtopping device captures seawater of incident
waves in a reservoir above the sea level, then releases the water back to sea through
turbines. An example of such a device is the Wave Dragon, a schematic diagram of
which is shown in Figure 5.23(a). This device uses a pair of large curved reflectors,
as shown in Figure 5.23(b), to gather waves into the central receiving part, where
they flow up a ramp and over the top into a raised reservoir, from which the water
can return to the sea via a number of low-head turbines.
The various categories of WECs described above can be summarized as illu-
strated by the block diagram shown in Figure 5.24.
Generator
Pelton wheel
Substation
or
lat
Seawater
cil
Os
return
Seawater
piston line
Flow
Seabed
Figure 5.21 Oscillating wave surge converter: Aquamarine Power Oyster [11]
The Wells turbine rotates in the same direction

regardless of the air flow. Thus, generating
irrespective of the upward or downward
movement of the water column Generator
Front wall
Air flow
Reinforced concrete capture chamber
set into the incavated rock face
Back wall
Sea waves
Air is compressed and

decompressed by the OWC to
be forced and drawn back
through the Wells turbine
Seabed
Figure 5.22 OWC: the Limpet
5.7 Mechanical interfaces

Wave energy captured using a WEC should be transformed into an adequate form
that can drive a rotating generator to produce electricity. So, intermediate
mechanical interfaces such as turbines are needed to make this adequate formation.
Three major interfaces can be used: (i) self-rectifying air turbine, (ii) high- or low-
head hydraulic turbine and (iii) high-pressure oil-driven hydraulic motor.
5.7.1 Air turbines

This type of turbine allows a single-directional air flow, and so it is typically used
in all terminator or OWC devices. The self-rectifying air turbines have the
Overtopping Reflectors
Reservoir
Ramp
Reservoir
Turbine outlet
(a) (b)
Figure 5.23 Overtopping WEC: the Wave Dragon. (a) A schematic diagram.
(b) The large curved reflector
Categories of WECs
Location Type Mode of operation
Submerged
Onshore Point absorber pressure
differential
Nearshore Line absorber Oscillating

wave surge
converter
Offshore Terminator
OWC
Overtopping
device
Figure 5.24 WEC categories
capability of controlling the rotational speed to match the current sea state. Wells
turbine, as a typical example, is the most commonly used as an air turbine. It has
(i) the capability to rotate at high speed for a low velocity of air flowing through the
turbine, (ii) an efficiency of 70%–80% and (iii) simplicity and high reliability. On
the other hand, it has a negative torque at low flow rates and aerodynamic losses at
flow rates exceeding the stall-free critical value, i.e. resulting in output power
drop. Kaplan turbines are the modified and enhanced version of Wells as the effi-
ciency is increased by controlling the rotor blade pitch angle [14].
A block diagram of the air turbine is shown in Figure 5.25. The input of
the turbine model is the air velocity, vair, and the output is the mechanical torque,
Tm [15].
Kvair Pair
vair Pshaft Tm
ηPair Pshaft /ωm
φ
Vair/ωmr F(φ) η
ωm
Figure 5.25 Block diagram of air turbine model
The output mechanical torque is given by

Tm ¼ Pshaft =wm (5.26)
where wm is the shaft angular speed and Pshaft is the turbine shaft power
determined by
Pshaft ¼ hPair (5.27)
where h is the turbine efficiency and Pair is the air power. Turbine efficiency is a
function of flow coefficient j as given by Figure 5.26 [15]. The air power is pro-
portional to the air velocity as
Pair ¼ Kvair (5.28)
where K is a constant. The flow coefficient j is determined by the turbine air
velocity and the shaft speed as
j ¼ vair =ðwm rÞ (5.29)
where r is the turbine radius. The output turbine torque has a pulsating nature
because of the variation of air velocity at the turbine input with the variation of
wavelength and wave height. The pulsation of the output torque can be smoothed
by coupling a large flywheel along the shaft between the turbine and the generator.
5.7.2 Hydraulic turbines

The operation of this type of turbine is based on converting a head of water, created
between the reservoir of an overtopping device and the mean sea level, into
mechanical rotational power to drive an electrical generator. Hydraulic turbines are
classified into two categories: low-head (typically 3–4 m) and high-head (typically
tens to hundreds of metres) turbines. The flow can be controlled by adjusting the
inlet guide vanes or by a servomechanism, whereas the blades of the runners are
also adjusted to improve the efficiency over a wide range of flow. This is costly and
commonly not used in the small turbines. On the other hand, high-head hydraulic
turbines (e.g. impulse Pelton type) are employed in some line and point absorber
devices. They can also be considered as alternative to hydraulic motors. In general,
60
Turbine efficiency η (%)

40
20
0.0
0.0 0.1 0.2 0.3 0.4 0.5

Flow coefficient φ
Figure 5.26 Turbine efficiency versus flow coefficient
the efficiency of the two types of hydraulic turbines (low and high head) is very
sensitive to the head-to-rotational speed ratio and may reach 90%. So, the use of
variable speed electrical generator is highly favourable.
The functional block diagram of a hydraulic turbine-generator set as a typical
example is shown in Figure 5.27. More details including the mathematical repre-
sentation of each function can be found in [16].
5.7.3 High-pressure oil hydraulic motors

The oil hydraulic motor acts as a prime mover of the electrical generator by con-
verting its received hydraulic input energy that comes from the motion of an
oscillating body through a hydraulic cylinder into rotational energy. As an example,
a typical application of this oil hydraulic system to power buoy point absorbing
device is illustrated in Figure 5.28. The hydraulic system is equipped with a gas
accumulator system capable of storing energy over a few wave periods with a goal
of smoothing out the irregular power absorbed from the wave. The oil-hydraulic
mechanical interface can be used for point absorber or attenuator devices.
Direct mechanical interface. A floating cylinder (called ‘Bristol cylinder’)
collects the wave’s movement and is mechanically connected to the energy unit by
flexible joints and rods. The rods move slowly with cylinder and the reciprocating
motion is transferred to the axles in converter unit. This converter unit is the heart
of the system and converts the reciprocating motion to a rotating shaft that is
directly connected to a suitable slow speed electric generator for producing elec-
tricity at high efficiency. When transferring converter movements with mechanical
arms and rotation to the generator, the efficiency should be kept as high as possible.
The floating cylinder operates under sea level as shown in Figure 5.29.
The pendulum system can also be installed in the shoreline. It consists of a
parallelepiped concrete box, which is open to the sea at one end, as shown in
Figure 5.30. A pendulum flap is hinged over this opening, so that the actions of the
Interchange power Electrical system
Assigned unit generation

Automatic generation P
(generators, network and e
control Frequency loads)
Angle Turbine-
Speed generator set
inertia
Governor Speed
Speed Governor-controlled
speed control Turbine
governor valves or gates Pm
changer mechanism
Speed-governing system Turbine and
energy system
Figure 5.27 Functional block diagram of hydraulic turbine-generator set.

Pe ¼ electrical power; Pm ¼ mechanical power
LP accumulator HP accumulator
Buoy
Motor
Controller
Hydraulic
cylinder
Figure 5.28 Hydraulic mechanism with point absorber device. HP = high

pressure, LP = low pressure
waves cause it to swing back and forth. This motion is then used to power a
hydraulic pump and an electric generator.
5.8 Electricity generation
The generator delivers power into the grid with constant frequency and rms vol-
tage. Because the turbine rotates with a variable speed, a synchronous machine is
Sea level
Waves Floating
cylinder
Under sea Electric

generator
O
move scillating
ment
conve
rter
Figure 5.29 Direct mechanical interface for wave energy extraction
Hydraulic pump
Pendulum
flap
Wave
Concrete box
Figure 5.30 Onshore pendulum system
not appropriate. Instead, asynchronous machine such as a doubly fed wound rotor
induction generator (DFIG) is used, as shown diagrammatically in Figure 5.31. The
wound rotor is fed by the stator using a converter, and with this arrangement, the
frequency and voltage are kept constant for a large range of turbine speed variation.
More details about DFIGs are explained in Chapter 2, Section 2.4.
5.8.1 Direct drive wave energy

It is seen that the rotary generator is driven by one of the intermediate mechanical
interfaces such as air turbine or hydraulic turbine. This leads to the use of large
Converter
Wells
turbine
Grid
Air
flow
Double fed wound rotor

induction generator
Figure 5.31 DFIG with a prime mover of Wells turbine
Heaving buoy
Permanent magnets
Translator
Stator
Coil
Drag plate
Figure 5.32 Schematic diagram of a linear permanent magnet electric generator
expensive equipment, reduced efficiency, limited operating conditions and reduced

reliability of the WEC. Recently, wave power generation using a linear generator
has received more attention to be applied as it can convert the reciprocating motion
of a wave directly into electrical energy without any intermediate mechanism.
Thus, the linear generator is required to convert the slow and large forces of the
wave into electrical energy.
The linear generator has a stator which is a number of coils fixed to the seabed
as well as a linear rotor (called ‘translator’) which is the moving part. For wave
energy applications, the linear permanent magnet generator (LPMG) is commonly
used (Figure 5.32). When the buoy is lifted by the wave, the buoy sets the translator
in motion. It is the relative motion between the stator and the translator in the
generator, which induces a voltage in the stator windings. But when a generator
moves with varying speed and direction, it results in an induced voltage with
PTO system
WEC Interface
mechanism
Piston
Accumulator
Rotary
electrical
generator
Air chamber Air turbine
Power
PCC electronics
Ocean Accumulator Hydro turbine Grid
wave
Working fluid power

Direct mechanical drive system Mechanical power
Electrical power
Direct electrical drive system

LPMG
Figure 5.33 PTO systems
irregular amplitude and frequency. The peak value of the output power will be
several times higher than the average power production. The generator and the
electrical system have to be dimensioned for these peaks in power. More details
about the principal of operation and modelling of LPMG, integrated with interface
mechanisms, can be found in [17,18].
The power produced cannot be delivered directly to the grid without conversion.
This is done in several steps. First, the voltage of each generator is rectified to DC and
the resulting output DC voltage is filtered through capacitive filters. The filters smooth
out the voltage from the generators and create a stable DC voltage. Outputs of various
generators are connected to a bus, during short periods of time, and the power after the
filter will also be constant. If the system is studied during hour scales, or more, there
will be variations in the produced power and these variations are due to changes in the
sea state. To compensate for voltage variations of the output that occur due to sea state
variations, a DC/AC converter and a tap-changing transformer can be used.
5.9 PTO systems
As shown in Figure 5.33, the function of the power take-off (PTO) systems is to
harvest the ocean wave energy and convert this energy into an adequate electric form
to be delivered to the grid. Consequently, PTO systems can be classified as follows:
● Indirect PTO systems which use mechanical interface mechanism and deliver
power to the grid through a chain of equipment as the following:
– piston and accumulator-hydraulic motor-rotary generator;
– air chamber-air turbine-rotary generator;
– accumulator-hydro turbine-rotary generator;

– direct mechanical interface-rotary generator.
● Direct PTO systems which directly deliver the ocean wave power to the grid
through an LPMG and without any interface mechanism.
It is to be noted that for all PTO classes power electronics is used at the point
of common coupling for power adoption-based power quality [19–21].
5.10 Pros and cons
Advantages of wave energy

1. Renewable: Waves are created by the winds and their energy can be captured.
The wave energy will never run out and can be delivered with all benefits that
renewable energy sources (RESs) would have.
2. Green: Power extraction from ocean waves does not emit harmful by-products
such as gas, waste and pollution like there is with fossil fuels. So, it is con-
tributing to the reduction of CO2 emissions. The wave energy can be taken
directly to electrical generators to produce electricity.
3. Energy potential: A huge amount of power can be extracted from the waves. It
still needs more research and development to exploit this power in addition to
increasing the efficiency of WECs.
4. Predictable: One of the major advantages of wave energy is its predictability as
the waves are consistent throughout the day. Thus, it is easy not only to cal-
culate the amount of power that can be produced but also to properly manage it.
5. Power density: It is greater for wave energy than that of wind energy and thus it
will be more efficient.
6. Reliability: Compared to wind and solar as RESs, the generation of electrical
power from wave energy is more reliable as the waves are hardly interrupted
and are always in motion.
7. Onshore and offshore wave power: For onshore wave energy plants, the cap-
turing of wave energy leads to reducing the power with which the waves hit the
shoreline, thus causing less damage to the shore. On the other hand, offshore
wave energy plants are installed in locations at which the waves have larger
energy potential as well as they could minimize the negative environmental
impacts.
Disadvantages of wave energy
1. Cost: Cost of harnessing wave power is generally very high because of
equipment installation such as WECs and interface mechanisms in the sea area
and in many cases offshore. Thus, the equipment may be exposed to storms,
strong ocean waves and seawater. Therefore, monitoring and maintenance,
which is not easy to implement, are required and of course this will increase
the cost.
2. Location: Wave energy is not a viable power source everywhere. Its location is
restricted to be onshore or offshore. Cities or lands near the ocean will benefit
directly from it, whereas those far away have to find alternate sources.
3. Source of disturbance for private and commercial vessels: Another downside is
that it disturbs commercial and private vessels. Power plants that gather wave
energy have to be placed by the coastline to do their job, and they have to be
near cities and other populated areas to be of much use to anybody. But these
are the places that are major thoroughfares for cargo ships, cruise ships,
recreational vehicles and beach goers. All of these people and vessels will be
disrupted by the installation of a wave energy gathering source. This means
that government officials and private companies that want to invest in wave
energy sources have to take into account and consider the needs of those that
may be disturbed.
4. Wavelength: Wind power is highly dependent on wavelength, that is, wave
speed, wavelength and water density. They require a consistent flow of pow-
erful waves to generate significant amount of wave power. Some areas
experience unreliable wave behaviour and it becomes unpredictable to forecast
accurate wave power and, therefore, cannot be trusted as a reliable energy
source.
5. Wave energy generators may be unpleasant for some who live close to coastal
regions. They look like large machines working in the middle of the ocean and
destroy the beauty of the ocean. They also generate noise pollution but the
noise is often shrouded by the noise of waves, which is much more than that of
wave energy generators.
5.11 Principle of OTEC operation

Oceans of the globe constitute nearly 70% of the earth’s surface and can be con-
sidered as large reservoirs for storing solar energy. The concept of ocean thermal
energy conversion (OTEC) has been explored for years with the goal of utilizing
this energy as a RES. It can be extracted by exploiting the thermal gradient that
exists in the tropical ocean waters. OTEC is a process based on the temperature
difference between the warm surface layer and deep cold layer of ocean water. The
temperature difference is transformed to power to drive a heat engine-generator set
for producing electricity.
OTEC plant operation is dependent on the natural thermal gradient found in
tropical ocean waters, that is, 20 latitude in the Atlantic and Pacific oceans.
Thus, the plant must be located at a site having (i) abundant warm surface water
(upper layer) to provide the heat source for a heat engine and (ii) equally abundant
cold water supply (lower layer) to be used as a heat sink for the engine.
Typical thermal gradient is shown in Figure 5.34. The temperature at the upper
layer (~100 m from the water surface) is about 27 C. It first decreases fast with
increasing depth and then slowly to reach 5 C at a depth of 1,000 m. Therefore, the
thermal potential of 27 5 C ¼ 22 C can be used to operate the OTEC system
Temperature (°C)
0
10 20
~27 (°C)
0.5
Depth × 103 m
1
1.5
2 ~5 °C
Figure 5.34 Typical temperature gradient in the tropics
[22]. Thermodynamically, OTEC plant can be designed to operate in either an

open-cycle or a closed-cycle operation modes [23].
5.11.1 Open-cycle OTEC

Employing Claude cycle, seawater is used as a working fluid in the open-cycle
system, which consists of the following steps: (i) flash evaporation of a fraction of
the warm seawater by reduction of pressure below the saturation value corre-
sponding to its temperature, (ii) expansion of the vapour through a turbine to
generate power, (iii) heat transfer to the cold seawater thermal sink resulting in
condensation of the working fluid and (iv) compression of the non-condensable
gases (air released from the seawater streams at the low operating pressure) to
pressures required to discharge them from the system. These steps are depicted in
Figure 5.35. In case of a surface condenser, the condensate (desalinated water) must
be compressed to pressures required to discharge it from the power-generating
system. Other useful outcomes from the OTEC system can be obtained such as
(i) generate hydrogen for use as a clean fuel source, (ii) generate fertilizer from
biological nutrients that are drawn up from the ocean floor in the cold water intake
and (iii) source of ocean water to be used as drinking water via desalination (taking
out the salt).
The evaporator, turbine and condenser operate in partial vacuum ranging from
3% to 1% atmospheric pressure. This poses a number of practical concerns that
must be addressed. First, the system must be carefully sealed to prevent in-leakage
of atmospheric air that can severely degrade or shut down operation. Second, the
specific volume of the low-pressure steam is very large compared to that of the
pressurized working fluid used in closed-cycle OTEC. This means that components
must have large flow areas to ensure that steam velocities do not attain excessively
high values. Finally, gases such as oxygen, nitrogen and carbon dioxide that are
dissolved in seawater (essentially air) come out of solution in a vacuum. These
gases are uncondensable and must be exhausted from the system. Despite the
complications, the Claude cycle enjoys certain benefits from the selection of water
Warm water
intake
Evaporator Steam
Turbine-generator set
Warm water
outlet
Cold seawater Cold water

return to sea Condenser intake
Vacuum
pump
Dissolved gas Desalinated
water
Figure 5.35 Open-cycle OTEC flow diagram
as the working fluid. Water, unlike ammonia, is non-toxic and environmentally

benign. Moreover, since the evaporator produces desalinated steam, the condenser
can be designed to yield fresh water. In many potential sites in the tropics,
potable water is a highly desired commodity that can be marketed to offset the price
of OTEC-generated electricity.
5.11.2 Closed-cycle OTEC

The fundamental thermodynamic description of a closed-cycle OTEC plant is
similar to that of other thermal plants operating in closed-cycle mode, which fol-
lows a Rankine cycle. The common working fluid is water, as used in coal and
nuclear power stations, but the small temperature differences and near-ambient
conditions for OTEC lead to using ammonia as a working fluid. The cycle com-
prises four processes as shown in Figure 5.36(a) and (b), indicating the
temperature-entropy (T-S) diagram and the corresponding schematic diagram of an
OTEC system, respectively [24–26]. According to these figures and applying the
first law of thermodynamics, the four processes can be described as follows.
Process #1, isobaric heat supply (evaporator) from State 1 to State 2: During
this process, the heat from the high-temperature source is added to the pressurized
working fluid to convert it into high pressure vapour at State 2. The heat supplied to
T
2
1
Vapour
Liquid
4 Liquid + Vapour 3
S
(a)
2 ~21.5 °C Electrical
generator Power line
~27 °C
Turbine to the grid
Qin Wout
24 °C Evaporator 3
Warm water intake
~10.7 °C 5 °C
Condenser
Cold water intake
Circulating
pump 7 °C
1 4
Working fluid
Qout
Win
Pump Pump
Ocean surface warm water ~27 °C
Ocean depths cold water ~5 ºC

(b)
Figure 5.36 Rankine cycle: (a) T-S diagram and (b) OTEC system with typical
values of temperature at each process
the cycle (evaporator) is given by

Qin ¼ mðh2 h1 Þ kJ (5.30)
where m is the mass flow of the cycle and hi is the enthalpy of the working fluid at
the ith state.
Process #2, isentropic expansion (turbine) from State 2 to State 3: an isentropic
process, in which the entropy of working fluid remains constant. The vapour flows
through the turbine where its internal energy is converted into mechanical work to
drive the electrical generator. Thus, the work output of the working fluid during this
process can be expressed as
Wout ¼ ðh2 h3 Þ kJ=kg (5.31)
Hence, the power delivered by the turbine to the electrical generator, Pout, is
given by
Pout ¼ mWout ¼ mðh2 h3 Þ kW (5.32)
Pout ¼ mWout ¼ mðh2 h3 Þ kW (5.33)
Process #3, isobaric heat rejection (condenser) from State 3 to State 4: an
isobaric process is one in which the pressure of working fluid remains constant. The
condenser is a large shell-and-tube-type heat exchanger. This is positioned next to
the turbine in order to receive a large flow rate of low pressure vapour. This vapour
in the condenser goes under a phase change from vapour to liquid. External cooling
water is pumped through thousands of tubes in the condenser to transport the heat
of condensation of the vapour away from the plant. Upon leaving the condenser, the
condensate is at a low temperature and pressure. Therefore, the heat rejected from
the cycle, Qout, is given by
Qout ¼ mðh4 h3 Þ kJ (5.34)
The value of Qout is negative as h4 < h3 . Thus, consistent with the sign convention,
Qout represents an outflow of heat from the condensed vapour.
Process #4, isentropic compression (pump) from State 4 to State 1: During the
isentropic compression process, external work is done on the working fluid by
means of pumping operation. The pump is a device that moves liquid from a low
pressure to high pressure. In the Rankine cycle, the condensate is raised to the
pressure of the evaporator by feed pumps. The high-pressure working fluid entering
the evaporator is called feed-working fluid and is reused around the Rankine cycle
again. The work, Win, and the corresponding power, Ppump, required to drive the
pump are given by
Win ¼ ðh4 h1 Þ kJ=kg (5.35)
Ppump ¼ mðh4 h1 Þ kW (5.36)
The pump work has a negative value as h1 > h4 . This is consistent with the
thermodynamic sign convention, which indicates that work and power must be
supplied to operate the pump.
A measure of the effectiveness of an energy conversion device is its thermal
efficiency, hth. This is defined as the ratio of the cycle network to the heat supplied
from external sources. Thus, using (5.30), (5.32) and (5.36) the ideal Rankine-cycle
thermal efficiency in terms of cycle enthalpies is given as the ratio of the net power
delivered by the Rankine cycle, which is the difference between the turbine power
and the magnitude of the pump power, to the total heat supplied to the cycle, that is,
Pnet Pout Ppump h2 h3 þ h4 h1
hth ¼ ¼ ¼ (5.37)
Qin Qin h2 h1
To determine the overall efficiency and simulate the OTEC system as a power
source connected to a grid, the combined turbo-generator efficiency, htg, should be
considered rather than the turbine efficiency. OTEC simulation is defined by a
given value of the volume flow rate of OTEC deep seawater per unit (latitude–
longitude) area, wcw, and the OTEC net power area density Pnet is calculated from
the following formula [27,28]:
( )
9 ðDT Þ2 9
Pnet ¼ wcw rcp htg (5.38)
80 T 200
where T is the absolute temperature of the OTEC surface seawater intake, and DT is
the temperature difference between surface and deep seawater intakes. The average
seawater density r is taken as 1,025 kg/m3, the seawater specific enthalpy cp as
4,000 J/kg K and the OTEC combined turbo-generator efficiency htg as 0.75. The
numerical coefficients account for a surface-to-deep seawater flow rate ratio of 1.5,
and seawater pumping power losses equal to 30% of the turbo-generator output at
standard conditions ðT ¼ 300 K; DT ¼ 20 KÞ. It is noted that these somewhat
arbitrarily defined standard conditions for the OTEC thermal resource are con-
servative, and correspond to the net power production of 1 MW with about 3 m3/s
of deep cold seawater. These figures would be much better in most of the OTEC
region; that is, less seawater would produce the same amount of net power, or
equivalently, pumping power losses would represent a smaller fraction of the turbo-
generator output.
5.12 OTEC and environment

It is seen that the OTEC technology requires drawing seawater from the surface
layer and the deep layer to feed the evaporator and condenser, respectively, and
returning it to the sea. From the environmental point of view, it is noted the fol-
lowing [29]:
● The carbon dioxide outgassing from the seawater used for OTEC operation of
either open cycle or closed cycle is much less than that released by fossil fuel
plants.
● Protection of heat exchangers (evaporators and condensers) from biofouling is
necessary because of their exposition to surface seawater. Several mechanical
and chemical means can be used for this purpose that leads to reducing the
probable release of working fluid during operations as well as having effective

heat transfer.
● Marine food web is minimally affected because of the absence of the sea
surface temperature anomalies that may result from the sustained flow of cold
seawater.
● The working fluid (ammonia is commonly used) and biocide emissions from
the plant must be too low to avoid their potential hazard to the workers and
persons in surrounding areas.
● The inlet flow velocity of intakes should be limited to minimize entrainment
and impingement of organisms in addition to keeping benign fishing life.
● Similar to the other steam-electric power generation plants, OTEC is asso-
ciated with risks of safety issues such as electrical hazards, use of compressed
gases and so on. On the other hand, because OTEC power plant operates at low
temperature and low pressure, the probable hazards to plant workers and local
population are much less than those of fossil fuel plants.
5.13 Technical limitations and challenges

Thermodynamic fundamentals applied to determine the performance of OTEC
power-generating cycles are the same as those of conventional steam power plants.
However, large quantities of both warm and cold seawater are required for heat
transfer processes. In addition, the pumping process, the pipe-fluid frictional losses
and gravitational energy due to the density difference between the colder water
inside the pipe and the surrounding water column, all consume about 20%–30% of
the power generated by the turbine-generator set [29].
The actual energy conversion for 26 C warm seawater and 4 C cold seawater
is from 3% to 4%, whereas the ideal is about 8%. Thus, the efficiency in both cases
is very low, in particular, compared to the efficiency of conventional power plants.
So, to improve the performance of OTEC system, it is required to increase the size
of cold water pipe and use of large diameter pipes to transport the deep seawater to
the evaporators, which are presently limited and represent a big challenge of sig-
nificant magnitude [30].
Design and installation of OTEC plants present engineering challenges. For
instance, position keeping of the components of OTEC floating plants such as
submarine power cable, mooring systems designed for water depth of more than
1,000 m [31,32].
5.14 Site selection

The OTEC concept is based on the difference in temperature, DT, between the
warm (Tw of about 22–29 C) tropical surface waters and the cold deep ocean
waters (Tc of about 4–5 C at 1,000-m depth), as the source of the thermal energy
required.
Tc at a given depth, below 500 m, does not vary much throughout the regions
of interest for OTEC. It is also a weak function of depth, with a typical gradient of
1 C per 150 m between 500 and 1,000 m. These considerations may lead one to
regard Tc as nearly constant, with a value of 4–5 C at 1,000 m.
A desirable OTEC thermal resource of at least 20 C requires typical values of
Tw of 25 C. Generally, regions between latitudes 20 N and 20 S are adequate.
Some definite exceptions exist due to strong cold currents: (i) along the west coast
of South America, tropical coastal water temperatures remain below 20 C, and are
often 15 C; and (ii) a similar situation prevails to a lesser extent for the west coast
of Southern Africa. Moreover, Tw varies throughout the year, and sometimes
exhibits a significant seasonal drop due to the upwelling of deeper water induced by
the action of the wind. A careful OTEC site selection requires comprehensive
knowledge of local climate features as they may affect Tw seasonally [30].
The following summarizes the availability of the OTEC thermal resource
throughout the world:
● Equatorial waters, defined as lying between 10 N and 10 S, are adequate
except for the west coasts of South America and Southern Africa.
● Tropical waters, defined as extending from the equatorial region boundary to,
20 N and 20 S are adequate, except for the west coasts of South America and
of Southern Africa; moreover, seasonal upwelling phenomena would require
significant temperature enhancement for the west coast of Northern Africa, the
Horn of Africa and off of the Arabian Peninsula.
5.15 Advantages and disadvantages
Oceans surround many of the world’s countries, and these provide an abundant
source of tapping this form of ocean thermal energy. But, in order to benefit from
such green energy, it is crucial to understand the advantages and disadvantages of
OTEC. In addition, this will help in understanding why ocean thermal energy is not
widely used despite most countries having a coastal region.
Advantages
● OTEC technology feeds off an unchanging condition, the solar energy stored
in the world’s oceans. Therefore, it can run virtually continuously, unlike other
RESs such as solar and wind power. On an average day, the world’s oceans
absorb an amount of energy equivalent to 250 million barrels of oil, approxi-
mately 4,000 times the population’s current energy needs. Once the generators
and water pipes are in place, only minimal upkeep is required to keep the flow
of electricity running and no harmful by-products result from the process.
● OTEC can also support numerous spin-off industries. Chilled water that has
already been used can be pumped out of the plant and used in air conditioning,
industrial cooling and chill-soil agriculture (where pipes containing chilled
water are used to chill soil so it can support temperate crops in tropical cli-
mates). Additionally, desalinized water can be produced through OTEC
systems by implementing surface condensers to turn evaporated seawater into

potable (fresh) water. A 2 MW plant, for instance, could produce about
4,300 m3 of potable water.
Disadvantages
● At present, government subsidies are required to make OTEC energy eco-
nomically viable. Electricity can be produced at about $0.07/kWh, as opposed
to subsidized wind power systems that can produce energy for as low as $0.05/
kWh. Moreover, OTEC requires expensive, large-diameter pipes submerged
more than 1 km below the ocean’s surface. Many of the countries within the
viable geographical belt (between the Tropic of Cancer and Tropic of
Capricorn) lack the economic resources to build this infrastructure.
● Because OTEC facilities are stationary surface platforms, they are essentially
considered artificial islands and, therefore, their exact location affects their
legal status under the United Nations Convention on the Law of the Sea treaty
(UNCLOS). According to UNCLOS, coastal nations are given 5-, 20- and
320-km zones of varying legal authority. The amount of political autonomy
among these zones varies greatly. Consequently, jurisdictional conflicts could
arise based on international boundary disputes between nations.
References
[1] Rodrigues L. ‘Wave power conversion systems for electrical energy

production’. RE&PQJ. 2008;1(6):601–7.
[2] Evans O. F. ‘The classification of wave-formed ripple marks’. Journal of
Sedimentary Research. 1941;11(1):37–41.
[3] Monroe J. and Wicander R. ‘The changing Earth: Exploring geology and
evolution’. 7th edn. Oceans, shorelines, and shoreline processes. Stamford,
CT, USA: Cengage Learning; 2015, pp. 385–403.
[4] Marshak S. ‘Earth portrait of a planet’. 5th edn. Sedimentary rocks.
New York: W.W. Norton & Company; 2015, pp. 202–30.
[5] Hung L. and Tsai W. ‘The formation of parasitic capillary ripples on gravity-
capillary waves and the underlying vertical structures’. Journal of Physical
Oceanography. 2009;39:263–89.
[6] Drew B., Plummer A. R. and Sahinkaya N. N. ‘A review of wave energy
converter technology’. Proceedings of the Institution of Mechanical
Engineers, Part A: Journal of Power and Energy, IMechE. 2009;
223(8):887–902.
[7] Chenari B. A., Saadatian S. S. and Ferreira A. D. ‘Wave energy systems: An
overview of different wave energy converters and recommendation for
future improvements’. 8th International Technology, Education and
Development Conference; Valencia, Spain, March 2014, pp. 1–7.
[8] Kempener R. and Neumann F. Wave energy technology brief. International
Renewable Energy Agency (IRENA) Report, June 2014.
[9] Twidell J. and Weir T. Renewable energy resources. 3rd edn. USA: Taylor &
Francis Group, Routledge (imprint of); 2015.
[10] Poullikkas A. ‘Technology prospects of wave power systems’. Electronic
Journal of Energy & Environment. 2014;2(1):47–69.
[11] Aquamarine Power Oyster. Available from http://tinyurl.com/cn9k4k/
[Accessed 2 July 2018].
[12] Duckers L. ‘Wave energy’ in Boyle G. (ed.). Renewable energy. 2nd edn.
Oxford, UK: Oxford University Press; 2004, Chapter 8.
[13] Prudell J. H., Schacher A. and Rhinefrank K. ‘Direct drive ocean wave
energy electric plant design methodology’. Proceedings of 2012 Oceans;
Hampton Roads, VA, USA, Oct 2012. IEEE; 2012, pp. 1–7.
[14] Huang F. and Beattie W. C. ‘Modeling and simulation of the Islay wave
power conversion system’. 1997 IEEE Electric Machines and Drive
Conference Record; Milwaukee, WI, USA, May 1997, pp. MB2/2.1–MB2/
2.4.
[15] Wu F., Ju P., Zhang X., et al. ‘Modeling, control strategy, and power
conditioning for direct-drive wave energy conversion to operate with power
grid’. Proceedings of the IEEE. 2013;101(4):925–41.
[16] IEEE Committee. ‘Dynamic models for system and hydro turbines in power
system studies’. IEEE Transactions on Power Apparatus and Systems.
1973;92(6):1904–15.
[17] Sanada M., Morimoto S. and Inoue Y. ‘Generator design and characteristics
in direct-link wave power generating system considering appearance
probability of waves’. Renewable Energy Research and Applications
(ICRERA), 2012 International Conference on; Nagasaki, Japan, Nov 2012,
pp. 1–6.
[18] Okano K., Sanada M., Morimoto S. and Inoue Y. ‘Output characteristics in
direct-link wave power generating systems considering place of installation’.
Power Electronics and Drive Systems (PEDS), 2013 IEEE 10th
International Conference on; Kitakyushu, Japan, April 2013. IEEE; 2013,
pp. 187–93.
[19] Qing K., Xi X., Zanxiang N., Lipei H. and Kai S. ‘Design of grid-connected
directly driven wave power generation system with optimal control of output
power’. Power Electronics and Applications, 2013 15th European Conference
on; Lille, France, Sep 2013, pp. 1–8.
[20] Müller N., Kouro S., Glarfa J. and Malinowski M. ‘Medium-voltage power
converter interface for wave dragon wave energy conversion system’.
Energy Conversion Congress and Exposition (ECCE); Denver, CO, USA,
Sep 2013. IEEE; 2013, pp. 352–8.
[21] Szabo L., Oprea C., Festilă C. and Dulf É. ‘Study on a wave energy-based
power system’. Proceedings of Electrical machines, 2008. ICEM. 18th
International Conference on; Vilamoura, Portugal, Sep 2008. IEEE; 2008,
pp. 1–6.
[22] Lavi A. ‘Ocean thermal energy conversion: A general introduction’. Journal
of Energy. 1980;5(6):469–80.
[23] Nihous G. C. ‘Ocean thermal energy conversion’ in Crawley G. M. (ed.).

The world scientific handbook of energy. Singapore: World Scientific
Publishing Co.; 2013, pp. 359–72.
[24] Kapooria R. K., Kumar S. and Kasana K. S. ‘An analysis of a thermal power
plant working on a Rankine cycle: A theoretical investigation’. Journal of
Energy in Southern Africa. 2008;19(1): 77–83.
[25] Chan M. ‘Ocean thermal energy conversion’. Penn Sustainability Review,
University of Pennsylvania. 2015;1(7):29–31.
[26] Irianto D., Dibyo S., Salimy D. H. and Pane J. S. ‘Thermodynamic analysis
on Rankine cycle steam for cogeneration systems RGTT200K’. Prosiding
Seminar Nasional Teknologi Energi Nuklir 2016; Batam, Aug 2016,
pp. 865–72.
[27] Rajagopalan K. and Nihous G. C. ‘Estimates of global ocean thermal energy
conversion (OTEC) resources using an ocean general circulation model’. An
International Journal, Renewable Energy. 2013;50:532–40.
[28] Jia Y., Nihous G. C. and Rajagopalan K. ‘An evaluation of the large-scale
implementation of ocean thermal energy conversion (OTEC) using an ocean
general circulation model with low-complexity atmospheric feedback
effects’. Journal of Marine Science and Engineering. 2018;6(1):1–28.
[29] Vega L. A. ‘Ocean thermal energy conversion’. Encyclopedia of sustain-
ability and technology. New York: Springer; 2012, pp. 7296–328.
[30] Lohani B. and Vega L. A. Wave energy conversion and ocean thermal
energy conversion potential in developing member countries. Asian
Development Bank, ADB, Report, 2014.
[31] Tanner D. ‘Ocean thermal energy conversion: Current overview and future
outlook’. An International Journal, Renewable Energy. 1995;6(3):367–73.
[32] Martins M., Sinama F. and Lucas F. ‘Equivalent Gibbs systems for model-
ling an onshore OTEC experimental plant on Reunion Island’. International
Journal of Energy Research. 2013;37:1112–21.
Chapter 6
Biomass energy
6.1 What is biomass?

Biomass is defined as the organic material that comes from living plants and ani-
mals. It has the ability of storing the energy absorbed from the sun (called ‘bio-
energy’) internally. As with most living things, energy is stored in the form of
carbohydrates including celluloses, starches and sugars that are produced as part of
the process of photosynthesis. In the process of photosynthesis, plants convert
radiant energy from the sun into chemical energy in the form of glucose or sugar, as
shown in Figure 6.1, which can be expressed as
Water þ Carbon dioxide þ Sunlight ! Glucose þ Oxygen
or
6H2 O þ 6CO2 þ Radiant energy ! C6 H12 O6 þ 6O2
The process of photosynthesis enables green plants to capture the electro-
magnetic radiation from the sun and transform it into chemical energy, with the
biomass energy obtained by reversing this photosynthesis process. The stored
energy within the living matter is released when the chemical compounds within
the organic materials are broken down due to decomposition. The organic residue
left over from this process is called ‘biomass’. When biomass is burnt, the chemical
energy in biomass is released as heat. Biomass can be burnt directly or converted to
liquid biofuels or biogas that can be burnt as fuels [1].
Therefore, biomass energy can be produced from a total natural process and
source, and consequently it is regarded as a green and climate-friendly form of
energy, making biomass energy a renewable energy resource unlike fossil fuels
(Figure 6.2).
6.2 Types of biomass
Biomass sources can potentially comprise a wide variety of different natural

materials from liquids to solids to gas. Biomass can be made out of all natural
products, for example, agricultural and forestry products, some of the energy
crops growing everywhere including trees, wood, grass, plants and animal wastes
Su
nli
tgh
Trees
Figure 6.1 Photosynthesis process [National Energy Education Project (public

domain)]
and slurries. In addition, humans are an integral part of the systems making life,
and so the waste from human beings also can be a source of biomass (e.g. rub-
bish). All these can be used to produce biomass and bioenergy as shown in
Figure 6.3 [2,3].
Biomass can be used in solid form (the most common example is that of wood)
and it is very advantageous because of having a wide range of conversion and
refining processes through which it can be made into both liquids and gases. These
biomass solids, liquids and gases can then be used to produce electrical power,
heat, new chemicals and even combustible liquid fuels.
Biomass energy 171
Input: Biomass Processing Output: Bioenergy
Wastes Processing Fuel production

technology
Figure 6.2 Biomass as a renewable source of energy
Wastes for biodiesel

production
Marine wastes Industrial wastes
Municipal and
Animal wastes Biomass sources
sewage wastes
Agriculture crops Forestry crops and

and residues residues
Wastes for
fermentation
Figure 6.3 Biomass sources
6.3 Biomass energy
Energy can generally be recovered through several ways such as burning of fossil
fuels, utilization of sunshine to generate solar energy or by making use of flow of
water for hydroelectric generation. All these are ways of creating energy. Even the
Earth’s core can be used to collect energy, e.g. geothermal energy. However, today
biomass is quickly becoming a growing source of renewable energy around the

world. It continues to grow and help reduce the consumption of fossil fuels.
As mentioned above, there are many different types of biomass energy and
biomass fuel sources available with many related terms, such as biomass, bioenergy
and biofuels being used as though they all mean the same. Generally speaking,
biomass is the organic non-fossil material of biological origin that can be used as a
renewable energy source. Bioenergy is the final energy product or useful energy
that is made available from the conversion of biomass, whereas biofuels are liquid
or gaseous fuels of biomass origin.
The conversion process to obtain the final energy product depends on the type
of organic materials included in the biomass such as follows:
1. Solid biomass, such as wood and garbage, can be burnt directly to produce heat
for building, industry and generating electrical energy. Biomass can also be
converted into a gas called biogas or into liquid biofuels such as ethanol and
biodiesel. These fuels can then be burnt to produce energy.
2. Biomass materials, such as paper, food scraps and yard waste, that normally
decompose in landfills can be burnt to generate electricity or converted into
biogas in landfills. Similarly, sewage and animal manure can be processed in
special vessels called digesters to produce biogas. Biogas can be burnt as a fuel.
3. Ethanol can be made by fermentation from crops such as corn and sugarcane.
Biodiesel is produced from vegetable oils and animal fats. These liquid fuels
(ethanol or biodiesel) can then be used as a source of energy in vehicles and/or
as process heat.
Biomass energy production involves converting biological materials or wastes
into substances that can be used as carbon negative or at the very least carbon
neutral fuel [4]. Carbon neutral means that the carbon released during the use of the
fuel is reabsorbed and balanced by the carbon absorbed by new plant growth during
photosynthesis. The biomass crops are then harvested to make the next batch of
fuel, thus maintaining the cycle of carbon within the Earth and its atmosphere
without adding to the problem. Although the environmental impact of burning
biomass is less harmful than those associated with fossil fuel, it can be more
harmful than the impact associated with burning natural gas.
A typical example is shown in Figure 6.4 to illustrate the carbon cycle that
starts by the harvesting of the crops (e.g. corn) to be separated into their component
sugars. Then, the sugars are distilled to produce ethanol, which in turn can be used
as an alternative fuel to power vehicles or engines that produce electricity. The
released carbon dioxide, CO2, is reabsorbed by the original crops to complete the
cycle that is repeated again.
6.4 Biomass properties

The choice of biomass is affected by the form of energy required, and its properties
determine the choice of conversion process. Therefore, the properties of concern
Biomass energy 173
Crops are separated

into their component
sugars
Crops (e.g. corn) Sugars are distilled

for making ethanol
Reabsorbed CO2
by original crops
Released CO2
Ethanol is used as a fuel
Figure 6.4 Carbon cycle from crops to a liquid fuel
that should be investigated to specify the adequacy of the biomass materials and
consequently select the conversion process are based on [5] the following:
● Calorific value to define how much energy content in the material can be
released during burning in the air.
● Alkali metal content, which is important for the thermochemical conversion
processes.
● Ash residue content produced by combustion in air, which affects the cost of
bioenergy cycle.
● Moisture content because of its impact on the conversion process, in particular,
the thermal conversion which requires low moisture content feedstock.
● Proportions of fixed carbon and volatile matter which provide an indication of
how the biomass is to be exploited as an energy source.
● Bulk density which determines the sizing of the materials handling system and
consequently the costs of transport and storage.
6.5 Biomass energy conversion technologies
Key biomass energy conversion technologies include thermal, thermochemical,

biological and biochemical as briefly presented next and depicted in Figure 6.5 [6,7].
Thermal technology. Direct combustion converts forestry, agricultural
and municipal residues into heat and power under environmentally sound condi-
tions. The biomass feedstock after handling and storage goes through different
Biomass feedstock
Conversion
technologies
Thermal Thermochemical Biological Biochemical
Thermal and Anaerobic

Process Combustion Gasification Pyrolysis Fermentation
digestion
catalytic
Steam Diesel fuel

Final Heat Bio-oils and and blending Methane-
Syngas Bioethanol
product Electricity distillates rich biogas
stocks
Figure 6.5 Conversion of biomass feedstock into final bio-products
Biomass Exhaust
feedstock Steam
Turbine
Biomass Electricity
Biomass pre- Biomass
handling processing Boiler
and storage treatment
Generator
Ash
Condenser
Figure 6.6 Thermal technology by direct combustion
stages: preparation and treatment, combustion to produce heat, using the heat
energy to provide steam, and finally driving a turbine-generator set to generate
electricity (Figure 6.6).
Thermochemical technology. It includes the following:
● Gasification: It converts forestry, agricultural and municipal residues into syngas.
● Pyrolysis: It converts forestry and agricultural residues into bio-oils and value-
added products [8,9].
● Thermal and catalytic processing: It converts a variety of new and used
vegetable oils, liquid rosin (obtained from the kraft pulping process), waste
greases and animal fats into renewable diesel fuels and diesel blending stocks.
An example of gasification technology is illustrated by the block diagram of
Figure 6.7. The gasification process is carried out on the collected biomass, and
then by combustion to drive the gas-turbine as a prime mover of a generator to
produce electricity.
Biological technology. Fermentation to convert the starch and cellulose com-
ponents of biomass to bio-ethanol. The liquid fuel, ethanol, as an example, is the
final product of processing the biomass by fermentation and distillation [10,11].
Biomass energy 175
Biomass Air
Air or steam
feedstock
Turbine Generator
Biomass
handling Gasifier Gas cleaning Combustion Electricity
and storage chamber
Heat
exchanger
Hot water Cool water
Stack
Figure 6.7 Thermochemical technology by gasification
Crops Reabsorbed CO2

(corn)
Enzymes
and yeast
Biomass
handling Biomass pre- Liquefy and Fermentation Distillation Ethanol
and storage treatment saccharify
Figure 6.8 Biological conversion technology to produce ethanol
Ethanol is a corn alcohol that can be used to power cars or engines that produce
electricity (Figure 6.8).
Biochemical technology. Anaerobic digestion to convert manures and food
processing and municipal wastes into methane-rich biogas.
6.6 Pros and cons of biomass energy

Using biomass for energy has positive and negative effects [12,13]. Biomass and
biofuels made from biomass are alternative energy sources to fossil fuels: coal,
petroleum and natural gas. Burning either fossil fuels or biomass releases CO2, a
greenhouse gas. However, the plants that are the source of biomass capture a nearly
equivalent amount of CO2 through photosynthesis while they are growing, which
can make biomass a carbon-neutral energy source.
6.6.1 Burning wood

Using wood, wood pellets and charcoal for heating and cooking can replace fossil
fuels and may result in lower CO2 emissions overall. Wood can be harvested from
forests, woodlots that have to be thinned or urban trees that fall down or have to be
cut down [14].
Wood smoke contains harmful pollutants such as carbon monoxide and parti-
culate matter. Modern wood-burning stoves, pellet stoves and fireplace inserts can
reduce the amount of particulates from burning wood. Wood and charcoal are major
cooking and heating fuels in poor countries, but if people harvest the wood faster
than the time required for the trees to grow, it causes deforestation. Planting fast-
growing trees for fuel and using fuel-efficient cooking stoves can help slow defor-
estation and improve the environment.
6.6.2 Burning municipal solid waste or wood waste

Burning municipal solid waste or garbage to produce energy in waste-to-energy
plants means that less waste is buried in landfills. However, burning garbage pro-
duces air pollution and releases the chemicals and substances from the waste into
the air. Some of these chemicals can be hazardous to the people and the environ-
ment if the chemicals are not properly controlled.
The US Environmental Protection Agency (EPA) applies strict environmental
rules to waste-to-energy plants and mandates that waste-to-energy plants use air
pollution control devices such as scrubbers, fabric filters and electrostatic pre-
cipitators to capture air pollutants [15].
Scrubbers clean emissions from waste-to-energy facilities by spraying a liquid
into the combustion gases to neutralize the acids present in the stream of emissions.
Fabric filters and electrostatic precipitators also remove particles from the com-
bustion gases. The particles called ‘fly ash’ are then mixed with the ash that is
removed from the bottom of the waste-to-energy furnace.
A waste-to-energy furnace operates at high temperatures (1,800 F to 2,000 F),
which breaks down the chemicals in municipal solid waste into simpler, less harmful
compounds.
6.6.3 Disposing ash from waste-to-energy plants

Ash can contain high concentrations of various metals that were present in the
original waste. Textile dyes, printing inks and ceramics, for example, may contain
lead and cadmium.
Separating waste before burning can solve part of the problem. As batteries are
the largest source of lead and cadmium in municipal waste, they should not be
included in regular trash. Florescent light bulbs should also not be put in regular
trash because they contain small amounts of mercury.
The EPA tests ash from waste-to-energy plants to make sure that it is not hazar-
dous. The test looks for chemicals and metals that could contaminate ground water.
Some municipal solid waste landfills use ash that is considered safe as a cover layer for
their landfills, and some other ash is used to make concrete blocks and bricks.
6.6.4 Collecting landfill gas or biogas

Biogas forms as a result of biological processes in sewage treatment plants, waste
landfills and livestock manure management systems. Biogas is composed mainly of
methane (a greenhouse gas) and CO2. Many facilities that produce biogas capture it
Biomass energy 177
and burn the methane for heat or to generate electricity. This electricity is con-
sidered renewable and, in many states, contributes to meeting state renewable
portfolio standards. This electricity may replace electricity generation from fossil
fuels and can result in a net reduction in CO2 emissions. Burning methane produces
CO2, but because methane is a stronger greenhouse gas than CO2, the overall
greenhouse effect is lower.
6.6.5 Liquid biofuels: ethanol and biodiesel

Biofuels are transportation fuels such as ethanol and biodiesel. Ethanol is promoted
to be used as a transportation fuel to help reduce oil imports and CO2 emissions. As
a result, nearly all gasoline now sold in some countries, for example, the United
States, contain some ethanol [16].
Biofuels may be carbon-neutral because the plants that are used to make bio-
fuels (such as corn and sugarcane for ethanol and soya beans and palm oil trees for
biodiesel) absorb CO2 as they grow and may offset the CO2 emissions when bio-
fuels are produced and burnt.
Growing plants for biofuels is controversial because the land, fertilizers and
energy for growing biofuel crops could be used instead to grow food crops. In some
parts of the world, large areas of natural vegetation and forests have been cut down
to grow sugar cane for ethanol and soya beans and palm oil trees for biodiesel. The
US government supports efforts to develop alternative sources of biomass that do
not compete with food crops and that use less fertilizer and pesticides than corn and
sugar cane. The US government also supports methods to produce ethanol that
require less energy than conventional fermentation. Ethanol can also be made from
waste paper, and biodiesel can be made from waste grease and oils and even algae.
Ethanol and gasoline–ethanol blends burn cleaner and have higher octane
ratings than pure gasoline, but they have higher evaporative emissions from fuel
tanks and dispensing equipment. These evaporative emissions contribute to the
formation of harmful, ground-level ozone and smog. Gasoline requires extra pro-
cessing to reduce evaporative emissions before it is blended with ethanol. Biodiesel
combustion produces fewer sulphur oxides, less particulate matter, less carbon
monoxide, and fewer unburnt and other hydrocarbons, but it does produce more
nitrogen oxide than petroleum diesel.
It can be seen from the above that no particular approach is dealt with and each
one has its own positive and negative aspects. While making a decision about
which approach to adopt, it is necessary to weigh various alternatives and adopt one
that offers the most positive outcome with respect to net environmental benefits.
References
[1] Mirza U. K., Ahmad N. and Majeed T. ‘An overview of biomass energy
utilization in Pakistan’. Renewable and Sustainable Energy Reviews.
2008;12:1988–96.
[2] Field C. B., Campbell J. E. and Lobell D. B. ‘Biomass energy: The scale of
the potential resource’. Trends in Ecology and Evolution. 2007;23(2):65–72.
[3] Doelle H. W. Biomass and organic waste conversion to food, feed, fuel,
fertilizer, energy, and commodity products [online]. Encyclopedia of Life
Support Systems (EOLSS). 2003. Available from http://www.eolss.net/Eolss-
sampleAllChapter.aspx [Accessed 8 Feb 2020].
[4] Tilman D., Hill J. and Lehman C. ‘Carbon-negative biofuels from low-input
high-diversity grassland biomass’. Science Magazine. 2006;314(5805):
1598–1600.
[5] McKendry P. ‘Energy production from biomass (Part 1): Overview of bio-
mass’. Bioresource Technology. 2002;83:37–46.
[6] McKendry P. ‘Energy production from biomass (Part 2): Conversion tech-
nologies’. Bioresource Technology. 2002;83:47–54.
[7] McKendry P. ‘Energy production from biomass (Part 3): Gasification tech-
nologies’. Bioresource Technology. 2002;83:55–63.
[8] Shuit S. H., Tan K. T., Lee K. T. and Kamaruddin A. H. ‘Oil palm biomass
as a sustainable energy source: A Malaysian case study’. Elsevier, Energy.
2009;34:1225–35.
[9] Mekhilef S., Saidur R., Safari A. and Mustafa W. E. S. B. ‘Biomass energy
in Malaysia: Current state and prospects’. Renewable and Sustainable
Energy Reviews. 2011;15:3360–70.
[10] Abbasi T. and Abbasi S. A. ‘Biomass energy and the environmental impacts
associated with its production and utilization’. Renewable and Sustainable
Energy Reviews. 2010;14:919–37.
[11] Farrell A. E., Plevin R. J., Turner B. T., Jones A. D., Hare M. O. and
Kammen D. M. ‘Ethanol can contribute to energy and environmental goals’.
Science Magazine. 2006;311(5760):506–8.
[12] Herbert G. M. J. and Krishnan A. U. ‘Quantifying environmental perfor-
mance of biomass energy’. Renewable and Sustainable Energy Reviews.
2016;59:292–308.
[13] Reilly R. and Paltsev S. Biomass energy and competition for land. Science
and Policy of Global Change Report no. 145, 2007.
[14] Dornburg V. and Faaij A. P. C. ‘Efficiency and economy of wood-fired
energy systems in relation to scale regarding heat and power generation
using combustion and gasification technologies’. Biomass and Bioenergy.
2001;21:91–108.
[15] Goldemberg J. ‘Ethanol for a sustainable energy future’. Science Magazine.
2007;315(5813):808–10.
[16] Heller M. C., Keoleian G. A., Mann M. K. and Volk T. A. ‘Life scale energy
and environmental benefits of generating electricity from willow biomass’.
Renewable Energy. 2004;29:1023–42.
Part II
Energy storage systems
Chapter 7
Electrical energy storage
7.1 Introduction
The traditional power system comprised a chain of three segments in cascade:
generation, transmission and distribution systems. Generation of electricity is
derived from a source of energy to produce electrical power, for example, fossil
fuels for thermal plants, water for hydraulic plants, wind, solar irradiation, etc.
Traditional generation of electrical power is highly centralized and, often, a long
distance away from its end users. The transmission system (second segment) is
required to transmit the electrical power at a high-voltage level to the end users
through the distribution system at medium voltage (third segment) in a unidirec-
tional power flow (Figure 7.1). It is crucial to recognize that the centralized pro-
duction of electricity leads to a complex power system configuration, high losses,
low electric power reliability and increased risk of instability. To solve these pro-
blems, distributed energy resources (DERs) can offer an alternative solution.
Residential
Generation Transmission Distribution Commercial

system system system
Unidirectional power flow Industrial
Load area
Figure 7.1 Schematic diagram for traditional power system

EES
EES EES Solar PV system

Electrical
substation
EES
Generation
system Compressed air energy storage
Transmission
system
Bi-directional power flow
EES
Wind turbines farm

Distribution
system EES EES
Industrial
Commercial Residential
Microgrid
Figure 7.2 Future perspectives of power systems
DERs are small, modular, energy generation and storage technologies that
provide electric energy installed on site and of a size meeting local need.
Furthermore, DER systems may be either connected to the local electric grid or
isolated from the grid in stand-alone applications. DER technologies include
renewable energy sources (e.g. wind turbines and solar photovoltaics), micro-
turbines, reciprocating engines, combustion turbines, cogeneration and electrical
energy storage (EES) systems. Utilities can use DER technologies to delay, reduce
or even eliminate the need to obtain additional central power generation, trans-
mission and distribution equipment and infrastructure. Meanwhile, DER systems
can provide voltage support and enhance local reliability.
Therefore, the perspective of the electrical power grid in future can be seen as
an integration of conventional generation and DERs and consequently the power
flow can be bi-directional as shown in Figure 7.2.
7.2 EES and renewable energy
Because of the rapid growth and development of renewable energy, sustainable

energy systems will have to be based not strictly on traditional resources but make
greater use of renewable energy. Furthermore, distributed electrical energy pro-
duction from renewable energy sources leads to a cleaner supply with fewer
environmental hazards for consumers [1].
However, the unpredictable character of renewable resources requires more
attention from planners to keep the system operating in an optimal manner. In
Electrical energy storage 183
addition, the power supplied by these resources is fluctuating as it depends on

unpredictable primary sources of energy, for example, wind speed for power gen-
erated by wind turbines and solar irradiation for solar photovoltaic systems. For the
reasons of unpredictability and fluctuating renewable energy, use of energy storage
systems is imperative to avert the problems, such as power imbalance between
supply and demand and the difficulty of voltage and frequency regulation that may
appear. This will help increase the penetration rate of renewable energy to the
power networks [2].
7.3 EES concept

Electrical energy is non-storable in its original form. EES refers to a process of
converting electrical energy from a power network into a storable form that can be
converted back to electrical energy when needed. Such a process enables electrical
energy to be captured at any time, for example, low demand, low generation cost or
surplus from intermittent energy sources, and to be used later of high demand, high
generation cost or in the presence of generation capacity shortage. The process of
energy capture is regarded as ‘charging,’ whereas the process of releasing the
energy to be used is regarded as ‘discharging’. The energy is stored using different
technologies and the storage medium (commonly known as carrier) is decided
accordingly.
7.3.1 Elements of EES

Energy storage systems can be configured by three major components as shown in
Figure 7.3 [3]:
● Energy storage device: It is designated to contain the carrier by which the
energy is held until needed, for example, water reservoirs in pumped hydro
storage or chemical/electrolyte used in battery, flywheel, etc.
● Power conversion system: It is the power electronics used to adjust the power
characteristics of the storage (e.g. the voltage and current) based on the load
requirements and ensures proper and safe charge/discharge of storage device.
● Facilities for housing the equipment, wiring, control and system protection
from physical damage.
7.4 The need for EES

Major reasons for installing energy storage can be summarized as follows:
● Renewable integration: EES can store extra renewable energy to be used at the
desirable time, eliminating power curtailment and oversized construction of
hybrid power system capacity. For instance, for wind power generation, EES
can contribute in relieving the fluctuations and provide voltage control support
during low-voltage ride through, resulting in smooth power output.
Charging
losses
Power grid EES
Charging
Input power
Discharging Power conversion system
Output power
Discharging losses
Energy
storage
Storage device
losses
Facilities
Figure 7.3 Elements of EES systems and energy losses
● Transmission and distribution (T&D) upgrade deferral: Storage provides

buffering capability to enable high penetration of variable renewables and
asset deferral for T&D systems (load management and ancillary services).
● Power quality: EES contributes in providing electrical service to customers
without any oscillations or disruptions to the voltage waveform such as swells/
sags, spikes or harmonics.
● Improved power reliability: Bridging power can be supplied by the EES for
consumers to ride through a power disruption.
● Energy systems stability: Energy storage helps in power system planning and
matching demand with supply to maintain grid stability.
7.5 Critical challenges

There exist some critical challenges in making energy storage cost-competitive.
These are the following:
● Critical challenges for energy storage are high system cost and cycle life with
– existing storage solutions too expensive;
– deep discharge and longer cycle life;
– safe and reliable chemistry;
– scalable technology to cover all markets.
● To make storage cost-competitive, advances across all major areas are required:
– storage, power electronics and power conversion systems;
– power quality and integration;
– engineered safety of large systems;

– codes and standards;
– optimal use of storage resources across the entire electricity infrastructure.
7.6 Functions of EES

An electric utility normally encompasses many electricity-generating plants.
A characteristic of electricity is that it is consumed at the same time as it is
generated. Thus, a proper amount of electricity must always be generated at any
time to meet the varying demand, thus keeping the balance between power
generation and consumption. Imbalance between supply and demand may result
in undesirable quality (voltage and frequency) of power supply and even lead to
instability.
Environmental concerns are leading to difficulties in installing new transmis-
sion lines to feed remote consumers. Depending on both network topology and
power flow, there is the probability of congestion happening in a specific line
causing a failure and consequently power interruption.
To continue supplying the load with enough energy and reconcile the differences
between supply and demand during both normal operation and contingencies, power
plants are built and operated with a variety of functions, as described in the following [4]:
● Base-load plants, such as large fossil fuel thermal power stations, provide
centralized generation with a bulk capacity to meet the base (constant) demand
for electricity. Such stations are normally run at full rating because technical
constraints do not allow rapid load following.
● Load-following plants, also known as ‘cycling’ or ‘intermediate load plants’,
have the capability to change the power output quickly and thus can follow the
changing balance between supply and demand. Hydroelectric and natural gas
fired plants are typically used in this category.
● Peaking units, designated to meet the peak demand, generally run less than a
few hundred hours per year.
Taking a typical daily load curve, Figure 7.4 as an illustrative example, the
base and intermediate loads are supplied, respectively, by base-load units and load
following plants with more flexible but less cost-effective generation. The base-
load and load following units are complemented by the peaking units to supply the
load during the peak load conditions, #1 and #2.
To meet unforeseen increase in demand, loss of conventional units and trans-
mission lines, additional generating units, known as ‘operating reserves units’, are
also made available by the utilities. Furthermore, some generating units, called
spinning reserves, are required for (i) frequency regulation, i.e. providing the
ability to respond to small, random fluctuations around normal load; (ii) providing
the capability to respond to a greater or less than predicted change in demand due to
load forecasting errors; and (iii) contingencies so as to be able to respond to, for
example, unplanned generator or transmission line outage.
100
Peak #1
Peak
#2
Demand (%)
65 Intermediate load
Base load
30
0 6 12 18 24
Hours (h)
Figure 7.4 Typical daily load curve [4]
Excess energy available in a form that is difficult to store can be converted to a

form in which it is more convenient or economic to store and use later. In case of
excess electrical energy available from the electric grid or another source, parti-
cularly a renewable energy source, energy can be stored in different forms of
energy such as electrochemical, electromagnetic, kinetic or potential energy, using
a suitable energy storage device and make it available when needed. Some energy
conversion technologies for energy storage provide short-term energy storage,
whereas others can store for longer periods. Various technologies available for EES
and the role played by them in the operation of the power system corresponds to
the available generation capacity versus the varying load demand over a period,
say 1 day.
To increase the reliability and resilience of the electric grid, EES technologies
and devices can play a prominent role in a variety of the functions mentioned above
as described in the following [5,6]:
Electric energy time shift: Because of the dependency of generation cost on
the type of generating plant, base, load following and peaking, the price of
electricity is higher at peak demand periods and lower at off-peak periods.
Electric energy time shift is often referred to as arbitrage. Unlike arbitrage in the
financial sense, energy arbitrage involves the purchase and sale of electricity at
different times to benefit from a price discrepancy in the cost of generation. For
example, electricity generated from wind at night or solar power during the day
can be purchased at low cost during these off-peak hours and sold during peak
hours. Storage devices, such as pumped hydro plants, compressed air energy
storage facilities and large battery installations, used for electric energy time shift
EES
100 100 Discharging
EES Discharging
Demand (%)
Demand (%)
EES charging
EES charging Intermediate load Intermediate load
65 65
Base load Base load

30 30
0 6 12 18 24 0 6 12 18 24
Hours (h) Hours (h)
(a) (b)
Figure 7.5 Load profile for large-scale EES system: (a) for load following and
levelling; (b) for peak clipping
can store large amounts of electricity to optimize the gain from electricity price
differentials between off-peak and peak demand times. This can also offset the
disadvantages of intermittent renewable energy sources by shifting this energy to
times when it is needed most.
Performance targets for electric energy time shift, a medium-duration energy
management application, focus on system capital cost, operations and maintenance
cost, discharge duration, efficiency and response time. The environmental impact
of storage devices used for electric energy time shift is also an important factor to
consider.
Load following and levelling: Energy storage technologies with very quick
response are ideal for load following, a medium-duration energy management
application (Figure 7.5(a)), and thus can isolate the rest of the grid from rapid and
substantial changes in net supply in comparison to demand. Many of these technol-
ogies can operate at partial output/input capacity without compromising the system
performance. In addition, they can quickly respond to increasing or decreasing load,
respectively, by discharging or charging rapidly.
The need to quickly respond to fluctuating demands requires the storage
device to have the ability to communicate and respond to the grid through the
system operator. Performance targets for the storage technologies used for load
following emphasize the importance of system capital cost, operations and
maintenance cost, and discharge duration. An energy storage technology, which
can meet these targets for the load following application, is well positioned to be
adopted by the electric power industry to assist with abrupt fluctuations in elec-
tricity demand.
Peak clipping: It is more economical and desirable to build enough generating
capacity to meet average demand rather than peak demands that may only last for a
short time each year. Utility planners wish to reduce the load during peak demand
periods as it defers the need for additional generation capacity. Therefore, large-
scale EES can be used to implement peak clipping by providing the electrical
energy required during peak periods (Figure 7.5(b)).
Frequency regulation: Sustained electricity interruptions can be costly and fast

response time can minimize both momentary and sustained interruptions. Some
storage technologies can vary output, from zero output to full output, within sec-
onds. Use of an appropriate energy storage technology offers the potential for
frequency regulation, a short-duration power management application that is far
faster than by a gas or steam turbine and, thereby, significantly reduce the impact of
power interruption of high-tech industrial and commercial facilities.
For optimum efficiency and response time, it is desirable that the energy sto-
rage technologies used for frequency regulation is able to communicate with the
grid in a quick and efficient manner. The performance targets of energy storage
technologies used for frequency regulation emphasize the importance of system
cost, system lifetime, discharge duration, response time and roundtrip efficiency
metrics.
T&D deferral and substitution: The continued increase in demand for elec-
tricity requires additional T&D infrastructure to transport electricity from the point
of generation to where it is consumed. It requires building new transmission lines,
substations and distribution infrastructure, a costly and time-consuming process.
Also, as transmission lines are designed for high reliability at peak conditions, they
normally operate at a capacity lower than the design capacity and thus low
utilization.
The ability to store the power generated during times of low demand and use
during the period of high demand can reduce the stress on power plants as well as
on the T&D infrastructure. In addition, EES closer to the point of use lowers the
congestion of the transmission system during peak demand. The target of the sto-
rage in this case, a long-duration energy management application, can be achieved
using relatively small amount of storage and help to postpone or eliminate the need
to build new T&D lines.
In addition to emphasizing the importance of system cost, discharge duration,
capacity, reliability and system lifetime in the performance targets for energy sto-
rage technologies applied to T&D upgrade deferral and substitution, safety is also
an important metric for storage systems used for this application. The storage
locations throughout the grid must be safe and have cybersecurity protection con-
sistent with the current utility practice. The targets mentioned above must be met to
justify investment in energy storage technologies compared to investing in
upgrades to the generation, transmission and distribution infrastructure.
Increased integration of renewables in the grid: Increasing the integration of
renewable energy into the grid can reduce reliance on fossil fuels and thus emis-
sions from electricity generation. However, when connected to the electric grid on a
large scale, the intermittent nature of these alternative sources introduces genera-
tion variability that raises other operational and integration issues that can be
categorized into the following two general areas.
First, output power fluctuations (e.g. ramp-up/ramp-down) of short duration
may make system frequency control difficult. It may also lead to deterioration of
the system operation under wide frequency deviation. Frequency control in con-
ventional systems is mostly managed by the capability to control the output of
conventional, hydro and thermal, generating units. This necessitates a partial

capacity operation of these generators with the output margin being used to adjust
frequency, even if it yields inefficient operation. Increased penetration of renewable
generation will need a further increase of the output margin, and, consequently, may
even further lower the efficiency of conventional generation. Proper use of EES can
mitigate the output fluctuation of renewable generation sources and reduce the
margins of conventional generators that can be operated at a higher efficiency. A
coordinated and judicious operation of renewable generation sources and energy
storage technologies can smooth the power generated from these sources. It can thus
support the increased penetration of renewables-generated electricity and, thereby,
ease the grid operation where large amounts of wind and solar generation are
deployed. To optimize the effectiveness of these operations, it is necessary to pro-
vide the storage technologies the ability to communicate and respond to the opera-
tional requirements of the grid through the system operator.
Second, ESS can be used for mitigation of medium-duration output power
fluctuation. For example, in the operation with a function of electric energy time
shift, ESS can store large amounts of electricity at off-peak periods (low electricity
price). Then, the energy can be discharged when it is needed to match energy
production with demand.
Performance targets for storage technologies used for renewables grid inte-
gration address roundtrip efficiency, system lifetime, capacity and response time
metrics.
7.7 Classification of EES technologies
Different technologies are used for storing the electrical energy. A widely used
approach for classifying these technologies is according to the form of stored energy.
It includes mechanical, chemical, electrochemical, electrical and thermal energy
storage. The different classes of EES technologies are shown in Figure 7.6. More
description about these matured technologies is presented in the forthcoming chap-
ters of this part. There are other technologies which still are in the demonstration and
development stage such as cryogenic and synthetic natural gas energy storage [7].
EES technologies
Mechanical Chemical Electrochemical Electrical Thermal

– Pumped hydro – Hydrogen – Batteries – Electrostatic – Sensible heat
– Compressed air (fuel cell/synthetic – Conventional (supercapacitor) storage
– Flywheel natural gas) – Flow – Magnetic
(superconducting
magnetic energy)
Figure 7.6 Classification of EES technologies

7.8 Characteristics of EES technologies

Each EES technology has its own characteristics (technical and economic), which
may differ from the others. Therefore, it is critical to generally identify the main
items of these characteristics to help the planners to properly select the storage
system that meets the application needs. These items are the following:
● Power availability: It is determined by both the average and peak value of
charge/discharge power (Pav and Pmax, respectively).
● Storage capacity: It is defined as the total energy stored, Est (Wh), which
exceeds the actual retrieved stored energy, Eact (Wh), as the discharge is often
incomplete. It is to be noted that the usable energy is limited by the depth of
discharge (DoD) and so storage capacity represents the limit of discharge depth.
● DoD: Energy storage should have a quick energy release action on demand.
The power discharge can be a limiting factor and called ‘DoD’ or ‘transmission
rate’ by which the time required to extract the stored energy is determined.
● Discharge time: It is the duration of maximum power discharge, tdis, and is
given by
Est
tdis ¼ (7.1)
Pmax
● It is a characteristic of storage system adequacy for some applications.
● Efficiency: On a single operating point, the efficiency, h, can be calculated as
the ratio between released and stored energy, that is,
Eact
h¼ (7.2)
Est
It is a simplified definition as the systems have charging, no-load and dis-
charging losses, which must be considered. In addition, in conditions of quick
charge/discharge, the efficiency deteriorates as Eact is much less than Est.
● Durability: It is also named as ‘cycling capacity’ and expressed as the max-
imum number of cycles where ‘on’ cycle corresponds to one charge and one
discharge. So, durability refers to the times the storage unit can release the
designed energy level after each recharge.
● Autonomy: It is the maximum time the storage unit can continuously release
energy. It is calculated as the ratio of retrieved stored energy, Eact, to the
maximum discharge power, Pdis, that is,
Eact
Aut ¼ (7.3)
Pdis
● Self-discharge: It is the portion of the energy that was initially stored and has
dissipated over a given non-use time.
● Energy density: It represents the maximum amount of stored energy per unit
mass or volume of the storage unit. It is an important indicator, especially for
portable applications.
● Operational constraints: They are commonly related to safety and operating

conditions such as temperature, pressure and so on, which affect the choice of a
storage technology in terms of energy needs.
● Reliability: For energy storage systems, reliability is an important character-
istic as it is a guarantee of on-demand service.
● Costs: As shown in Section 7.5, one of the main challenges is the cost of EES
systems, which determines how much the storage system is competitive. Of
course, investment of using a storage system is highly useful and interesting
when the total gains exceed the total expenses. Like any other investment, the
costs mainly include fixed and running costs. Fixed costs represent the cost of
purchase, installation and delivery of system assets, whereas operation and
maintenance (O&M) costs form the running costs. It is, therefore, crucial to
study the cost of EES technologies [8]. However, as reported in the literature,
two main approaches can be applied to estimate the cost as described in the
following:
(i) Total capital cost (TCC) to evaluate the related costs of the system ele-
ments shown in Figure 7.3. TCC includes the following:
– PCS costs, CPCS, expressed in per unit of power capacity ($/kW) and
represent the cost of purchase, installation and delivery.
– Energy storage costs, Cst, in per unit of stored or delivered energy
($/kWh) and represent the cost of purchase, installation and delivery
of reservoirs.
– Balance of power costs, CBOP, including costs of project engineer-
ing, grid connection interface, integration facilities, land and all
other required services. CBOP can be expressed in per unit of power
($/kW) or energy ($/kWh).
Accordingly, TCC can be represented by the following formula [3]:
TCC ¼ CPCS þ CBOP þ Cst h (7.4)
where h is the charging/discharging time.

(ii) Life cycle cost, LCC, which is a more effective indicator from the own-
ership point of view. LCC can be calculated by adding to TCC the fixed
and variable expenses of O&M, replacement, disposal and recycling. It is
represented annually in per unit of power ($/kW/year). Consequently,
TCC/year should be calculated by its present value of money as [9]
follows:
TCC=year ¼ TCC CRF (7.5)
and
ið1 þ iÞT
CRF ¼ (7.6)
ð1 þ iÞT 1
where
CRF is the capital recovery factor;
i is the interest rate;
T is the lifetime.
The annual O&M costs (CO&M/year) can be expressed by adding the annualized costs
of fixed O&M (CFO&M/year) and variable O&M (CVOM/year) multiplied by yearly
operating hours (h) as in the following.
CO&M=year ¼ CFO&M=year þ CVO&M=year h (7.7)
Thus,
LCC ¼ TCC=year þ CO&M=year þ CR þ CDR (7.8)
where CR is the annual replacement cost, and CDR is the annual disposal and
recycling cost.
7.9 Fields of energy storage applications
The principle of energy storage can be applied to both the ‘bulk power’ transmission
grid and the distribution system applications. Some of the technologies are available
for both applications. Currently, distribution applications are receiving more attention
due to the concept of micro-grid and distributed generation, and other power quality
issues. In addition, the need to become less dependent on fossil fuels and imported oil
is also driving research into energy storage devices for the transportation industry.
There are many things in common between these energy storage applications for
the transportation and backup power systems industries as they utilize similar
technologies. The concept of energy storage in the transmission grid is not new
either. Large pumped hydroelectric plants have been operating off this principle
for many decades. Over the past two decades many other energy storage tech-
nologies have matured to the point that they have become viable options for
selected applications in large-scale power systems. Some have already seen small-
to-medium scale applications. The viable storage technologies for different
applications are shown in Figure 7.7 [10].
It is to be noted that, for a specific application, different available EES systems
can be used commercially, but no single storage system meets all the requirements
that satisfy the best of all characteristics (high density, high efficiency, long life-
time, low cost, etc.). Each storage system has a suitable range; for instance, bat-
teries, compressed air, flywheel and pumped hydroelectric are suitable for load
following or load levelling and energy management applications. In contrast,
supercapacitors and flywheel are suitable for power quality. So, more than one
solution can be found. In this case, technical and economical comparison should be
investigated to achieve the best solution [11,12].
EES systems
Electrical distribution applications Large power grid connection

low and medium power
Batteries
Compressed air
Flywheel
Transportation Backup/emergency
Pumped hydroelectric
applications power applications
Superconducting magnetic
Supercapacitor
Batteries Batteries
Flywheel Compressed air
Supercapacitor Supercapacitor
Flywheel
Thermal energy
Fuel cell
Figure 7.7 Applications of different storage technologies
References
[1] Ibrahim H., Ilinca A. and Perron J. ‘Energy storage systems – Characteristics
and comparison’. Renewable & Sustainable Energy Reviews. 2008;12:
1221–50.
[2] Beaudin M., Zareipour H., Schellenberglabe A. and Rosehart W. ‘Energy
storage for mitigating the variability of renewable electricity sources: An
updated review’. Energy for Sustainable Development. 2010;14:302–14.
[3] Zakeri B. and Syri S. ‘Electrical energy storage systems: A comparative life
cycle cost analysis’. Renewable and Sustainable Energy Reviews. 2015;
42:569–96.
[4] Denholm P., Ela E., Kirby B. and Milligan M. The role of energy storage
with renewable electricity generation. National Renewable Energy
Laboratory report NREL/TP-6A2-47187. January 2010. Available at http://
www.osti.gov/bridge [Accessed 2 August 2020].
[5] International Electrotechanical Commission, IEC. Electrical energy storage.
White Paper, 2011.
[6] Lichtner S., Brindle R. and Pack L. Electric power industry needs for grid-
scale storage applications. Sandia National Laboratories Report, 2010.
[7] Aneke M. and Wang M. ‘Energy storage technologies and real-life appli-
cations – A state of the art review’. Applied Energy. 2016;179:350–77.
[8] Chen H., Cong T. N., Yang W., Tan C., Li Y. and Ding Y. ‘Progress in
electrical energy storage system: A critical review’. Progress in Natural
Science. 2009;19:291–312.
[9] Brealey R. A., Myres S. C. and Allen F. Principles of corporate finance.
New York, NY: McGraw-Hill/Irwin; 2011.
[10] Kondoh J., Ishii I., Yamaguchi H., et al. ‘Electrical energy storage systems
for energy networks’. Energy Conversion & Management. 2000;41(17):
1863–74.
[11] Schoenung S. M. and Burns C. ‘Utility energy storage applications studies’.
IEEE Transactions on Energy Conversion. 1996;11(3):658–65.
[12] Smith S. C., Sen P. K. and Kroposki B. ‘Advancement of energy storage
devices and applications in electrical power systems’. Proceedings of the
2008 IEEE Power and Energy Society General Meeting – Conversion and
Delivery of Electrical Energy in the 21st Century; Pittsburgh, Pennsylvania,
USA; 2008, pp. 1–8.
Chapter 8
Mechanical energy storage systems
Among the energy storage system (EES) types based on the form of energy stored
(Chapter 7, Section 7.7), mechanical energy storage (MES) systems are one of
these technologies. They include pumped hydroelectric storage (PHES), com-
pressed air energy storage (CAES) and flywheels (FWs).
PHES technology is suitable for energy management applications that move
the power over longer time scales and require continuous discharge ratings of
several hours and more. CAES is adequate for short- and long-duration energy
management applications, whereas FWs can be used for power quality applications
that require rapid response (less than a second), transient stability and frequency
regulation. Principles and operation of each technology are described in the forth-
coming sections.
8.1 PHES system

8.1.1 Physical principles
The principle of PHES is to store electrical energy by utilizing the potential energy
(PE) of water at a specific head, established between two reservoirs, one being
highly elevated than the other.
At periods of low demand and high availability of electrical energy, water is
pumped from the lower reservoir and stored in the upper reservoir. On demand,
water is released back into the lower reservoir through the same piping system to
generate electricity. The potential energy is converted into kinetic energy (KE) and
enthalpy (h). If m1 is a part of total mass m, which is moving with velocity v to
produce the kinetic energy, then the remaining mass, (m m1), multiplied by
pressure, p, over water density, r, gives the enthalpy. Therefore, the energy balance
is verified by the relation
PE ¼ KE þ h
1 (8.1)
¼ m1 v2 þ ðm m1 Þp=r
2
The generated output power, P, in watts (W) can be given by [1]
P ¼QH rh (8.2)
where Q is the fluid flow in cubic metres per second (m3/s), H is the hydraulic head
height in metres (m), r is the fluid density in kilograms per cubic metre (1,000 kg/m3
for water), g is the gravitational acceleration (9.81 m/s2) and h is the facility efficiency.
It is to be noted from (8.2) that the changeable variables are Q, H and h. The
flow is inversely proportional to the head; that is, if the head (H) is maximized, the
water utilization can be minimized and vice versa.
The cycle of operation is depicted in Figure 8.1. It is seen that the water dis-
charged from the turbine is stored in the lower-level water tank. The corresponding
components and their functions are given in Table 8.1 [2].
Recently, the total number of components has been decreased to reduce the
capital cost. Pump, turbine, motor and generator can be arranged in different ways,
which increases the overall efficiency and makes the system more compact. The
system which is composed of four units, i.e. motor, generator, pump and turbine,
makes the power house big in size. So, reversible motor/generator together with
pump and turbine configuration having three units can be used instead.
Additionally, reversible motor/generator and reversible pump/turbine systems that
have two compact components are preferred, as the investment cost is highly
reduced. A schematic diagram of a typical PHES system is shown in Figure 8.2.
Another aspect to increase the operation range of PHES is using asynchronous
motor-generators. The so-called doubly fed induction machines (DFIMs) increase the
flexibility, particularly during pumping mode. While the efficient pumping for syn-
chronous motor-generators is in the range of 90%–100% of the nominal power (PN),
DFIM can operate in pumping mode between 65% and 100% of PN.
Kaplan or Francis turbines are commonly used to maximize efficiency [3].
They are reversible and capable of handling both the pumping and the generating
Gate Water at high level

valve
Static head
Pen
stoc
Control line System

k
controller
Electric
power
Motor operated
by surplus
electric power Turbine
Water at low level
M Pump
Outlet water Generator
Figure 8.1 PHES operation cycle

Mechanical energy storage systems 197
Table 8.1 Functions of the main components of PHES system
Component name Component function

A hydroelectric pumped storage system Both reservoirs are open tanks that feature an
consists of two reservoirs (high and low), inlet/outlet to a pipe, located on the
pumps, turbines, motors, generators, bottom. The elevation difference between
penstocks, flow control valves, system the two reservoirs is called the ‘head’
controller and auxiliary equipment
Penstocks These are large diameter steel pipes,
connecting the reservoirs to the pump and
turbine
Pump When the system must store energy, pump
shaft is rotated with the force coming from
an electric motor
Turbine The centrifugal turbine converts stored
energy to mechanical energy, rotating the
shaft connected to the generator
Generator The generator converts mechanical energy
from the turbine output shaft to electrical
energy and feds into the electric grid
Motor The motor converts electrical energy (AC
power from the grid) to mechanical energy
to drive the pump when the system must
store energy
Flow control valve The flow control valve modulates the water
flow in the turbine and pump mode. It
serves as an emergency valve during
operation. It is completely open
during operation
System controller The system controller decides the operating
mode of the system, based on the power
plant output and grid demand
Auxiliary equipment Auxiliary equipment such as transformers
and AC/DC panels are used for operation
and protection of the system
process. Similarly, synchronous machine can be operated as motor during pumping

and as a generator to generate electricity.
Developments in the machinery do not only decrease the number of units but
also increase the round-trip efficiencies of the system. Global efficiency of a
pumped-storage system is between 75% and 80%. Typical key performance data is
given in Table 8.2.
8.1.2 Modelling of PHES dynamics

Dynamic models for PHES systems are necessary for studying the dynamic inter-
actions between these systems and the grid incorporating renewable energy
sources.
Water Upper
tunnel reservoir
Control
centre
Penstock
erate
Gen
Pump
Power house
Lower
reservoir Reversible
pump turbine
Figure 8.2 Developed PHES scheme
Table 8.2 Typical performance data
Power range (in general 10 MW–30 GW Life duration >80 years

no limitation)
Energy range Up to some Reaction time Some seconds to
100 GWh few minutes
Discharge time Some minutes Cycle efficiency 70%–85%
to 10 h
Cycle life Technically unlimited Energy (power) 0.5–3 Wh/kg
density
The PHES system shown in Figure 8.2 includes upper and lower reservoirs,
reversible pump-turbine unit and reversible synchronous machine. The water flow
direction and synchronous machine rotation direction in the generating mode are
opposite to those in the pumping mode. So, the synchronous machine rotation
direction should be reversed by changing the phase sequence at the machine
terminals to get it operating as a motor rather than generator.
Pumped-turbine system Electrical system

Head Power
Tunnel and Electrical system:
penstock water Pump-turbine Generator/
dynamics – Grid
dynamics motor – Loads
Flow ω
Field
Distributer Pilot valve and

Governor
Governor valve and gate gate Automatic
-controlled speed
servomotor servomotor generation
gates control
Dashpot
Governor
speed
changer
position
Speed governor system
Figure 8.3 Functional block diagram of a PHES system (solid arrows for
generating mode and doted arrows for pumping mode). The speed
governor system only operates in the generating mode
The functional block diagram, which represents the main elements of a typical
PHES system, is depicted in Figure 8.3. It is to be noted that the speed governor
system only operates to control the valve opening during the generating mode.
The valve opening is normally fixed at the maximum efficiency point in the
pumping mode.
The water dynamics for the pumped-turbine system penstock during generat-
ing mode can be represented by a simplified relation as follows [4]:
1
Q¼ ðHT H Hloss Þ (8.3)
Tw s
where Q is the output water flow from the system, HT is the total available static
head, H is the dynamic head established by pump-turbine unit, Hloss is the frictional
head loss, Tw is water starting time and s is the Laplace operator. The output
mechanical power, Pm, can be expressed as
Pm ¼ H ðQ Qnl Þ (8.4)
where Qnl is the no load flow.

The water flow is defined as positive when it flows downwards to drive the
pump turbine, otherwise, it is negative. The head also is positive in the direction
from upper to lower reservoirs (Figure 8.4(a)).
According to (8.3) and (8.4), the block diagram of a pump-turbine unit in
generating mode is as shown in Figure 8.4(b).
Upper
reservoir
HT H
HT + Q + Pm
H ∑ 1 ∑ X
H, Q, v − Tw s
− −
Hloss Qnl
(a) (b)
ω Equation (8.5) H 1 Q 1/ηp Pm

X
Tw s
(c)
Figure 8.4 Water dynamics. (a) Penstock. (b) Block diagram of a pump-turbine
unit in generation mode. (c) Block diagram of a pump-turbine unit in
pumping mode
On the other hand, in the pumping mode, the water dynamics depend on the
pump head-flow characteristic curve, which can be approximated by a quadratic
polynomial relation as given in the following:
H ¼ a0 w2 þ a1 wjQj þ a2 Q2 (8.5)
where a0 ; a1 and a2 are coefficients for curve fitting and w is the pump operating
speed. Consequently, the block diagram of a pump-turbine unit in pumping mode is
shown in Figure 8.4(c).
The water starting time, Tw (s), can be calculated by the relation [5]
Lv
Tw ¼ (8.6)
gHT
where L is the length of the penstock (m), v is water velocity (m/s) and g is the
acceleration of gravity (m/s2). Water velocity can be expressed in terms of power
generated by the system as
P ¼ 0:733vHT Ah (8.7)
where P is the generated electrical power (W), A is the average penstock area (m2)
and h is the product of turbine efficiency and generator efficiency. Hence,
0:139PL
Tw ¼ (8.8)
HT2 Ah
8.1.3 Applications
PHES can be used for energy management applications that include moving power
over longer timescales and generally require continuous discharge ratings of several
hours or more. So, PHES technology plays an important role for load levelling,
reducing the gap between off-peak and peak periods and power grid stabilization.
Besides balancing the peak and off-peak periods, PHES provides ancillary services
such as primary frequency and voltage control to the power grid. To fulfil the power
system control, PHES can switch within seconds for different operation modes [6].
Because of their flexibility, large-scale storage possibilities and grid operations
benefits, PHES systems will enable utilities to efficiently balance the grid and to develop
their renewable energy portfolios. In fact, the installation of intermittent renewable
generation has added a new degree of uncertainty to the dispatch of interconnected
power system. Pumped storage is therefore set to play a key role in enabling renewables’
grid integration while helping countries meet their ambitious targets of cutting green-
house gas emissions and of building additional clean renewable energy capacity.
8.1.4 Types of PHES

In the literature, PHES is classified according to the structure and operation type.
Structurally, pumped-hydro storage systems are classified as three types: pure
pumped-storage (also named as off-stream PHES or closed-loop systems), pumped-
back pumped storage and hybrid pumped storage plants [7].
1. Pure pumped storage plants shift water between two reservoirs, one of which is
located off-stream and other is river, lake or sea.
2. Back pumped storage system approach (mixed pumped storage) is a combi-
nation of pumped storage and conventional hydroelectric plants that use nat-
ural stream flow.
3. Hybrid pumped storage systems seem like more conceptual when compared to
the other two types. There are many examples of hybrid pumped storage power
plant, which are associated with wind power plant (WPP) or photovoltaic (PV)
solar farms. PHES plants are very dependent on geology and topography such
as WPP and PV farms. Physically combining PHES and WPP units in the same
location can be challenging due to their dependency on the natural conditions.
More often, installed capacity of the wind farm, which is the pump capacity of
PHES, is directly related with the energy need for pumping. That is why com-
bining those two plants in the same location restricts the total installed capacity.
According to its reservoir capacity and operation policy, PHES can be classi-
fied as daily, weekly and seasonally storage power plants.
At the daily operating plants, electricity is generated during peak hours and water
is stored at out off-peak hours. In the weekly cycle of water, some portion of the water
used in the generation of electricity during peak hours of weekdays is pumped back to
the upper reservoir within that day. Except during peak hours of weekends, water is
stored in the upper reservoir, which becomes empty at the end of the weekdays. On the
other hand, seasonally operating PHESs store water while the river flow and energy are
in excess, and to increase its firm energy use reserved water while the river flow is low.
8.1.5 Advantages and disadvantages of PHES

Main advantages of PHES. Pumped storage plants have the biggest share among other
storage technologies so far. Their advantages can be summarized as in the following:
● Fast response: Like other hydropower plants, PHES is faster in start-up and it
takes short time to reach its full generating capacity when compared to the
other types of power plants. Thus, it is possible to balance the varying elec-
trical energy demand for consumers and some unplanned outages of other
power plants in the grid system.
● Load balancing: Voltage and frequency are critical. Sudden changes in these
parameters damage the equipment in the consumer premises. Broken equipment
mean loss of value for domestic consumers and loss of production in the industry.
● Black-start ability: Other power plants need an external power during the start-
up of the system. However, like hydropower plants, PHES does not require
external energy to initiate generation. Thus, network restoration is possible if a
power blackout occurs.
● Energy storage: Grid energy storage balances electricity supply and demand,
which ensures stability and reliability in supply.
● High global efficiency: Round-trip efficiency can vary significantly for dif-
ferent energy storage technologies, depending on the number of cycles and
duration of usage. Pumped-hydro storage systems are approximately 80%
efficient and PHES has the highest global cycle efficiency when compared to
other storage technologies.
● Decrease in peak-hour electricity prices: Pumped storage hydro plant provides
the possibility of levelling the price of electricity in the market by being used
in connection with daily peak shaving and load levelling as well as weekly and
seasonal variations. PHES can lower electricity costs as it can store electricity
bought at low off-peak prices and they can use it during peak periods in place
of expensive power.
● Decrease water wastage: Pure pumped storage plants use the same water
several times and prevent water consumption. Seasonally operating PHES
facilities store excess water during the overflow season of the river, preventing
flood and storing water before reaching to the sea.
● Low operation and maintenance cost: PHES operation and maintenance costs
are low and these costs are directly related to the electricity prices. As PHES is
used in peak hours, high prices are lowered significantly when pumped storage
is in operation.
Disadvantages of PHES. There are several drawbacks of PHES facilities. The

negative sides of pumped hydro storage are cost of the overall system and envir-
onmental issues, which are caused by the construction of the power plant, operation
of pumps and turbines as explained in the following [8].
● High capital cost: Although the operation and maintenance costs are very low,
there is a high upfront capital investment, which can be recouped only over a
long time. Construction of reservoirs, dams and in some cases underground
structures is expensive and resource-consuming.
● Very dependent on location: The two basic requirements for a PHES facility
are head and water. Pumped storage systems require not only large volumes of
water but also considerable amount of land with specific type of conditions.
Finding suitable place for a PHES is not an easy task; that is why building the
storage and generation facility may need some improvements in the geology,
topography and so on, which increase the costs.
● Environmental impacts: Environmental impacts are also serious concerns for
PHES projects as well as hydropower. In general, hydropower receives criticism
due to flooding large areas, destroying terrestrial wildlife habitats and sig-
nificantly changing the landscape. Blocking natural water flows and/or change in
the path of the water disrupt the aquatic ecosystem and affect sediment carrying
capacity of the river resulting in scouring downstream. Pumping may also
increase the water temperature and stir up sediments at the bottom of the reser-
voirs and deteriorate the water quality. PHES operation may also trap and
kill fish.
8.2 CAES system

Another type of commercialized MES technology with high technical maturity and
large capacity is CAES. It can supply an output power of more than 100 MW per
single unit in addition to storing bulk energy in the form of high-pressure com-
pressed air. The energy can be stored during the periods of low electrical energy
demand and released to generate electricity at the peak periods to fill the gap
between the grid power and the high demand. CAES can be structured in different
scales (from small to large) and different storage durations (from short to long).
Moreover, CAES can efficiently operate at partial load with moderate response
time compared to other EES technologies.
Different types of CAES technologies have been commercialized (e.g. con-
ventional CAES systems), whereas others are more complicated, such as large-
scale adiabatic CAES (A-CAES) and liquid air energy storage systems. At the
discharge cycle, A-CAES is operated by integrating a thermal energy storage sys-
tem for converting the energy stored in the compressed air into electrical power
without involving a combustion process, which in turn results in a zero emission of
CO2 [9,10]. On the other hand, small-scale EES systems incorporating CAES have
been also developed. It is also possible to use a reversible generator/motor unit to
reduce the system cost and improve system efficiency.
8.2.1 Principles of CAES operation

The structure of CAES system includes an integration of different interacting
components, devices and processes such as the following:
● A motor and a generator or reversible motor/generator unit with clutches on
both ends (to engage/disengage it to/from the compressor train, the expander
train, or both).
● Multi-stage air compressors with intercoolers to reduce the power require-
ments needed during the compression cycle, and with an aftercooler to reduce
the storage volume requirements.
● An expander train consisting of high- and low-pressure turbo-expanders with
combustors.
● Auxiliary equipment (fuel storage and handling, cooling system, mechanical
systems, electrical systems and heat exchangers).
● Underground or aboveground compressed air storage, including piping and
fittings.
A schematic diagram of a conventional cycle compressed-air storage system is
shown in Figure 8.5 and the functional block diagram is depicted in Figure 8.6. The
operation of CAES can be described as follows:
The charging cycle (compression cycle)

● Surplus electricity in the storage system or electricity from the grid during
periods of low demand can be used to run compressors (low pressure (LP) and
Utility grid
Power conditioning system

(PCS)
Air inlet
Compressors Turbines
Combustor
LP HP HP LP
Motor Generator
Intercooler Fuel Exhaust
Coolant
Aftercooler Exhaust Recuperator
Compressed air carven
Charging cycle Discharging cycle
Figure 8.5 Schematic diagram of a conventional CAES system

Reversible
LP and HP compressors generator/motor LP and HP turbines
Air flow Shaft Shaft 6
1
5
Cooler Compressed air reservoir Combustor
Exhaust
2 3
Fuel
Recuperator
Figure 8.6 Functional block diagram of a CAES system with reversible motor/
generator model. Solid arrows represent the air flow direction
whereas the exhaust is indicated by dotted arrows. State points are
shown by numbered black circles
high pressure (HP) compressors) driven by a motor to inject the air into a
cavern (normally, installed underground). The high-pressure compressed air is
stored in the cavern at surrounding formation temperature.
● To improve the compression efficiency, coolers can be used to reduce the
operating temperature of the injected air.
The discharging cycle (expansion cycle)
● The stored high-pressure compressed air is released and heated and conse-
quently expanded through gas turbines (HP and LP turbines) and sometimes
steam turbines.
● The combustor implements the combustion process with the mixed com-
pressed air and fuel to rotate the turbines.
● The turbines are connected to an electrical generator to generate electricity.
● A recuperation device can be used to recycle the waste heat of the system
exhaust before it is released into the atmosphere, thus resulting in reduction of
both CO2 emissions and fuel consumption.
The recuperation cycle
A recuperation cycle plant, the conventional CAES thermal cycle with an addi-
tional component (the recuperator), is shown in Figure 8.5. A recuperator recovers
the low-pressure turbine waste heat to preheat the stored air before it goes into the
high-pressure combustor. This reduces the fuel consumption of the plant (as com-
pared to the conventional plant) by about 25%. This configuration is used in the
Alabama McIntosh plant that was designed for primary operation as a source of
peak power and as a load-management storage plant. As the recuperator reduces the
plant heat rate during generation by about 25%, it reduces the cost of the plant’s
peak power supply.
8.2.2 CAES system modelling

Modelling of the system is based on identifying each component of behaviour and
the following assumptions [11]:
● Ideal air with a constant heat capacity.
● Fixed maximum and minimum pressure in the tank.
● Constant heat transfer exchange coefficient for the tank.
● Accumulation (mass and energy) only occurs in the tank.
To identify the behaviour of each component, CAES circuit can be described
by the stages of the major transformation of the air in the system. At these stages,
represented by the state points (1–2, 2–3, 3–4, 4–5, 5–6) as shown in Figure 8.6, the
thermodynamic variables of the air (temperature, pressure, specific volume, spe-
cific enthalpy, specific entropy and specific internal energy) are instantly calculated
by the relationships of ideal gas law, and energy and mass balances [10–13].
At point #1, the air flows through the compressors and then to the cooler
providing the compressor isentropic output at point #2. Assuming an isentropic
transformation, from point #1 to point #2, the compressor can be modelled at point
#3 by means of its isentropic efficiency that is defined by
h2 ðtÞ h1 ðtÞ
his;comp ¼ (8.9)
h3 ðtÞ h1 ðtÞ
In the compressed air reservoir, at point #4, the rate of mass change, dmdtðtÞ, and
the rate of increase in internal energy of the air, dm dt , that is, energy balance equation
u
can be written as
9
dmðtÞ >
=
¼ m_ in ðtÞ m_ out ðtÞ þ N
dt (8.10)
dmu >
¼m _ in ðtÞh3 ðtÞ m_ out ðtÞh4 ðtÞ þ Q_ tnk þ W_ ;
dt
where h stands for the specific enthalpy of unit mass of the fluid, m _ in ðtÞ and m
_ out ðtÞ
are the inlet and outlet mass flow rates, respectively, N represents the net produc-
tion rate of mass, m, by the chemical reactions (equals zero when there is no che-
mical reaction inside the reservoir), Q_ tnk is the rate of change of heat transfer and
W_ is the rate of change in energy transfer because of work done (equals zero for the
energy balance of compressed air storage).
Q_ tnk , m_ in ðtÞ and the volumetric efficiency, hV ;comp , are, respectively, given by
9
Q_ tnk ðtÞ ¼ Krsv Arsv ðT4 ðtÞ Text ðtÞÞ >
>
>
V_ >
>
=
m_ in ¼ hV ;comp
V1 ðt Þ (8.11)
0:714 >
>
>
>
hV ;comp ¼ 1 0:05 bcomp ðtÞ 1 >
;
where Krsv is the overall heat exchange coefficient between the reservoir and the
environment, Arsv is the external area of the reservoir, hV ;comp is the compressor
volumetric efficiency and bcomp is the compression ratio.
At points #5 and #6, the power required by the compressor and that produced
by the turbine, respectively, can be calculated by the following relations:
9
1
Pcomp ðtÞ ¼ m_ in ðtÞ ðh3 ðtÞ h1 ðtÞÞ =
hm;comp (8.12)
;
Pturb ðtÞ ¼ hm;turb m_ out ðtÞ ðh6 ðtÞ h5 ðtÞÞ
where hm;comp and hm;turb are the mechanical efficiency of compressor and turbine,
respectively.
Thus, the overall efficiency of the CAES system (hCAES) and the coverage ratio
(CRCAES) can be expressed as
Eturb
hCAES ¼ (8.13)
Ecomp
Eturb
CRCAES ¼ (8.14)
EBldg
where Eturb, Ecomp and EBldg are, respectively, the turbine energy production, the
compressor energy consumption and the total building demand over a specific
period.
8.2.3 Advantages and disadvantages of CAES

Application of a CAES system provides the grid with ancillary services such as
peak shaving, spinning reserve, reactive power (VAr) support and arbitrage. By
exploiting CAES, the energy delivered from a variety of sources (e.g. wind, solar
and the grid itself) can be temporarily stored to be recovered when needed.
Advantages. For specific applications, CAES plants are designed to provide
economic benefit to owners and/or operators of power generation facilities. The
benefits of using a CAES plant to support power generation include the following:
● Increased use of generation facilities during off-peak hours (i.e. during the
storage plant charging cycle) and times of rapidly decreasing demand as the
compressor absorbs excess generating capacity. This is particularly useful
when base nuclear, hydro or fossil capacity is available at very low prices
during off-peak time periods.
● Provide ramping, intermediate and peaking power during the day.
● Store daytime solar energy and night-time wind energy to be delivered during
the higher priced time.
● Provide frequency regulation as CAES can provide much better frequency
control than a base-load power plant.
● Provide VAr support (e.g. by operating the CAES plant to supply reactive
power in the synchronous condenser mode).
● Provide off-peak to on-peak arbitrage.

● CAES technology is the only one that can provide significant energy storage
(in the thousands of megawatt hours) at relatively low cost. The plant has
practically unlimited flexibility for providing significant load management at
the utility or regional levels.
● CAES plants are capable of black start.
● CAES plants have fast start-up time. If a CAES plant is operated as a hot
spinning reserve, it can reach the maximum capacity within a few minutes.
● CAES plants have a ramp rate of about 30% of maximum load per minute.
● The nominal heat rate of a CAES plant at maximum load is about three
times lower than the heat rate of a comparable combustion turbine plant
using the same turbine expander. CAES plants also excel at part load. Their
heat rate at 20% of maximum load is 80% of the nominal heat rate at
maximum load. This is very good and unique, since all other oil, gas and
coal power plants have poor efficiency at 20% load, making them uneco-
nomical for operation at part load for normal duty. This characteristic of
CAES plants makes them very useful (and efficient) for ramping, part load
and regulation duty [14].
Disadvantages. A few of the potential reasons are listed as follows:
● The underground geology is likely perceived as a risk issue by utilities.
● Site selection is somewhat limited because of the need of installing under-
ground tanks (air compressed reservoirs) that may require certain geological
formations.
● Like any energy conversion technology, certain losses are inevitable.
● The requirement for additional heating in the expansion process is the most
significant disadvantage. By some estimates, 1 kWh worth of natural gas
will be needed for every 3 kWh generated from a CAES system. This is
particularly problematic if fossil fuels are used for the heat generation. As
natural gas prices increase, the economics of CAES, marginal at present,
could fail [14].
8.3 FW energy storage
As an energy storage device, flywheel energy storage (FES) stores electrical energy
in the form of kinetic energy in a rotating mass. It operates like any other storage
technology in the charging and discharging mode. FES systems can be categorized
into two classes: (i) conventional low-speed systems (up to 10,000 rpm) in which
steel rotors are used to increase inertia with mechanical bearings that result in large
and heavy system and relatively large standby losses (between 0.5% and 1% of
rated power) and (ii) high-speed composite systems with speeds up to 100,000 rpm
in which the rotor is made of composite materials with magnetic or super-
conducting bearings resulting in a compact system but because of the high speed,
air drag forces cause losses of about 0.1% of the rated power [15].
However, the cost and maintenance of low-speed FWs is less than that of high-
speed FWs. On the other hand, high-speed FWs have the advantage of low losses,
low mass and high-energy storage capacity.
8.3.1 FES architecture and operation

An FES system, Figure 8.7, includes the following main components:
● An FW (rotor), which stores the kinetic energy in the form of a rotating mass.
It is made of carbon-fibre composite in high-speed FWs to have higher tensile
strength than steal in low-speed FWs. Hence, the overall configuration as well
as the specific energy and specific power of high-speed FWs can be optimized.
● Mechanical or magnetic bearings to support the rotor. The type of bearings is
the major element that the FW speed and performance are dependent upon.
Mechanical bearings require frequent maintenance and produce friction, which
can hinder rotation at high rates, and so, they are commonly used in the low-
Axes of rotation
Upper
magnetic
bearing
a
b
Motor/
generator c
Vacuum-
sealed
containment
Rotor
Shaft
Lower
Vacuum magnetic
pump bearing
Figure 8.7 Typical flywheel main components

Utility grid Power flow during discharging mode

Power flow during charging mode
Flywheel
DC Flywheel
Grid converter
link converter
Three-phase
Motor/
transformer Back-to-back converter generator
Figure 8.8 Schematic diagram of a flywheel connected to a grid
speed FWs. On the other hand, magnetic or superconducting bearings elim-

inate the friction between the rotor and its supports and have long life and low
maintenance, and thus, they are preferred for high-speed FW systems [16].
● Vacuum enclosure for housing the system components at low pressure to help
reduce the self-discharge losses.
● Electrical motor/generator coupled to the shaft of the FW. It acts as a motor
during the charging mode and as a generator during the discharging mode.
● Power conditioning system, which includes two converters (back-to-back
converter) connected by a common DC link, one for motor/generator and the
other for the grid connection as shown in Figure 8.8 [17].
FES system, also named as ‘flywheel battery’ as it looks like a mechanical
battery, uses the inertia of a spinning mass to kinetically store energy or generate
power. Kinetic energy is quantified by one half the product of mass and speed
squared. So, the stored energy for low-speed FWs is directly proportional to the
mass, whereas for high-speed FW it is proportional to speed squared.
FES system operates by accelerating the FW (rotor) to a high speed and storing
the energy in the system as rotational energy; that is, it works in a charging mode
and the electrical machine acts as a motor. In the discharging mode, the energy is
extracted from the system as the electrical machine works as a generator, and
hence, the rotational speed of the FW is reduced according to energy conservation
principle. This means that electricity is used to accelerate the rotor in the charging
mode and decelerate the rotor in the discharging mode.
8.3.2 Mathematical relations

FW is a rotating, cylindrical object of considerable mass. It stores the kinetic
energy or increases the moment of inertia of a given system. Its design is based on
the storage capacity and the maximum rotating speed. The total energy stored in the
FW as a rotating mass, E, can be calculated by
E ¼ ½Iw2 (8.15)
where I is the moment of inertia and w is the rotational speed. For simplicity, the
FW is approximated to be a thin-walled cylinder. Thus, the moment of inertia can
be expressed in terms of its mass, m, and radius, r, as
I ¼ mr2 (8.16)
Substituting (8.16) into (8.15) gives
E ¼ ½mr2 w2 (8.17)
The useful stored energy, intrinsically, depends on the ratio between the
minimum and maximum rotating speed, wmin, wmax, respectively, as [15]
w2max w2min
E ¼ Emax (8.18)
w2min
The FW is mainly affected by two forces: the circumferential and radial
stresses. Therefore, the tensile strength, s, of a specific material gives a measure
of how much the FW can withstand these stresses during rotation. It can be
calculated by
s ¼ rr2 w2 (8.19)
where r is the density of material.
From (8.19), the allowable rotational speed can be determined. In addition,
substituting (8.19) into (8.17) gives the energy stored by the FW as
ms
E¼½ (8.20)
r
Hence, increasing the tensile strength and decreasing the material density will
lead to an FW with a high ratio of energy to weight. This in turn necessitates a
proper selection of the FW material according to the desired speed.
For FW dynamic analysis, the general form of equation of motion can be
written as [18]

½M Z€ þ ½C Z_ þ ½K fZ g ¼ ff g (8.21)
where ½M ¼ symmetric mass matrix, ½C ¼ symmetric damping matrix,
½K ¼ symmetric stiffness matrix, fZ g ¼ generalized coordinate vector and
ff g ¼ external force vector.
Several types of electrical machines can be used with FWs such as induction
machine (IM), DFIM, permanent magnet synchronous machine (PMSM),
brushless DC (BLDC) machine, and switched reluctance machine (SRM).
Advantages and disadvantages of each type are summarized in Table 8.3 [15].
More details about the dynamics and equivalent circuits for the first three types
are given in Chapter 2, Section 2.4, whereas the latter two types are explained in
Chapter 12, Sections 12.4.3 and 12.4.4, respectively. The power conditioning
systems used for interfacing are explained in Chapter 13.
Table 8.3 Pros and cons of types of motor/generator unit
Machine type Advantages Disadvantages

*
IM ● High reliability ● Speed limitations
● High robustness
● Low cost
DFIM ● Very large inertia ● Low power-conversion rating
● More losses in power electronics
PMSM ● Low rotor losses ● Idle losses due to eddy currents in the stator
● High energy density ● High cost
● High overall efficiency ● Low ruggedness
BLDC ● High energy density ● More power electronics are required, e.g.
● High efficiency electronic switching controller
● High cost
SRM ● Low idle losses ● High torque, flux and current ripples
● Wide speed range ● Difficulty of torque control at low speed
● Good robustness
*The best choice for power smoothing in wind energy conversion system and power conditioning
applications [15].
8.3.3 FES characteristics

FES systems can be mainly specified by the following characteristics:
● High specific energy (Wh/kg).
● Round-trip efficiency reaches 90%.
● Large maximum power output.
● Capacities ranging from small to large kilowatt hours are available.
● Time charging response of a system may be less than 15 min.
● Little or no maintenance.
● Very long lifetime as well as FWs have hundreds of thousands cycles of use.
8.3.4 Applications
Different fields of applications can use FES systems such as transportation, unin-
terruptible power supplies, aircraft launching systems, spacecraft energy storage,
pulse power source and grid energy storage.
FES systems can be used for grid energy storage as they can perform the
following functions [19]:
● Short-term spinning reserve for momentary frequency regulation of the grid
and voltage sag correction [20,21].
● Balancing sudden changes between generation and demand.
● Storing the energy during off-peak hours rather than traditional sources, for
example, natural gas turbines with no carbon emissions and faster response.
Table 8.4 Comparison between FWs and batteries
FWs Batteries
● Temperature changes have no effect and ● Operation is sensitive to the temperature
FWs can operate at a wider range changes
● Not subjected to failures of chemical ● Subjected and affected by the failures and
reactions incorrect chemical reactions
● Less potentially damaging to the ● May cause environment damage
environment
● Easy to know exactly the amount of ● Not easy to know the exact amount of
stored energy by measuring the stored energy
rotational speed
● Almost infinite working lifespan ● Operates for a finite period
● Minimal maintenance throughout ● Frequent maintenance is required
the service life is required as the
containment is completely sealed. For
instance, magnetic bearing FWs in a
vacuum enclosure do not need bearing
maintenance, whereas with mechanical
bearing, FWs have a limited lifespan
because of wear
● In some applications, especially in those ● Batteries can be arranged in a wide variety
having constrained space, FWs may not of configurations
be applicable
● Supporting the intermittent renewable energy sources, for example, wind

turbines [22].
FES operation is like the batteries in the same applications but may differ
economically. A comparison between them is given in Table 8.4.
References
[1] Barnes F. S. and Levine J. G. Large energy storage systems. Boca Raton, FL,
USA: CRC Press, Taylor & Francis Group; 2011, pp. 51–110.
[2] Çetinkaya S. Capacity determination of pumped storage projects using
market electricity prices. MSc Thesis, Middle East Technical University
(METU); 2014.
[3] Luo X., Wang J., Dooner M. and Clarka J. ‘Overview of current develop-
ment in electrical energy storage technologies and the application potential
in power system operation’. Applied Energy. 2015;137:511–36.
[4] Liang J. and Harley R. G. (eds.). ‘Pumped storage hydro-plant models for
system transient and long-term dynamic studies’. IEEE PES General
Meeting; RI, USA, July 2010, pp. 1–8.
[5] IEEE Committee. ‘Dynamic models for steam and hydro turbines in power
system studies’. IEEE PES Winter Meeting; New York, USA, Jan–Feb 1973,
pp. 1904–15.
[6] Denholm P., Ela E. and Kirby B. The role of energy storage with renewable
electricity generation. National Renewable Energy Laboratory (NREL)
Report, 2010.
[7] Ming Z., Junjie F. and Song X. ‘Development of China’s pumped storage
plant and related policy analysis’. Energy Policy. 2013;61:104–13.
[8] Baxter R. Energy storage—A nontechnical guide. Nashville, TN, USA,
PennWell Corporation; 2007.
[9] Luo X., Wang J., Dooner M., Clarka J. and Krupke C. ‘Overview of current
development in compressed air energy storage technology’. Int. Conf. on
Sustainability in Energy and Buildings, SEB-14; Warwick, UK. Energy
Procedia. 2014;62:603–11. Available from www.sciencedirect.com
[Accessed 6 Feb 2020].
[10] Luo X., Wang J., Krupke C., et al. ‘Modeling study, efficiency analysis and
optimization of large-scale adiabatic compressed air energy storage sys-
tems’. Applied Energy. 2016;162:589–600.
[11] Simpore S., Garde F., David M. and Marc O. ‘Design and dynamic simula-
tion of a compressed air energy storage system (CAES) coupled with a
building an electric grid and a photovoltaic power plant’. Proceedings of the
12th REHVA World Congress; Heiselberg, Per Kvols, Denmark, May 2016,
pp. 1–10.
[12] Haung Y., Chen H. S., Zhang X. J., et al. ‘Techno-economic modelling of
large-scale compressed air energy storage systems’. Energy Procedia.
2017;105:4034–9.
[13] Shamshirgaran S. R., Ameri M. and Ghadamian H. ‘Design of compressed
air energy storage (CAES) power plant using the genetic algorithm’. 4th
International Conference on Sustainable Energy and environment (SEE
2011); Bangkok, Thailand, Feb 2012, pp. 401–8.
[14] McCalley J. D. Compressed air energy storage. Educational Chapter. USA:
Trishna Das; 2012.
[15] Daoud M. I., Abdel-Khalik A. S., Massoud A., Ahmed S. and Abbasy N. H.
‘On the development of flywheel storage systems for power system appli-
cations: A survey’. Electrical Machines (ICEM), 2012 XXth International
Conference on; Marseille, Sep 2012, pp. 2119–25.
[16] Yamauchi Y., Uchiyama N., Suzuki E., Kubota M., Fujii M. and Ohsaki H.
‘Development of 50 kWh-class superconducting flywheel energy storage
system’. International Symposium on Power Electronics, Electrical Drives,
Automation and Motion, SPEEDAM 2006; Taormina, Italy, May 2006,
pp. 484–6.
[17] Chen Z., Zou X., Duan S. and Wei H. ‘Power conditioning system of fly-
wheel energy storage’. 8th International Conference on Power Electronics
ECCE Asia; The Shilla Jeju, Korea, May 2011, pp. 2763–8.
[18] Owusu-Ansah P., Yefa H. and Misbawu A. ‘Rotor dynamic modeling and
analysis of a flywheel rotor’. Proceedings of the 2015 International
Conference on Electronics and Automation Control (ESAC 2015);
Zhengzhou, China, Aug 2015, pp. 35–9.
[19] Theubou T. A., Wamkeue T. R. and Kamwa I. ‘Modeling and simulation of
a flywheel energy storage system for microgrids power plant applications’.
EIC Climate Change Technology Conference (CCTC 2015); Montreal, QC,
Canada, May 2015, pp. 1–12.
[20] Takahashi R. and Tamura J. ‘Frequency stabilization of small power system
with wind farm by using flywheel energy storage system’. Diagnostic for
Electric Machines, Power Electronics and Drives, 2007, SDEMPED 2007,
IEEE International Symposium on; Cracow, Poland, Sep 2007, pp. 393–8.
[21] Samineni S., Johnson B. K., Hess H. L. and Law J. D. ‘Modeling and ana-
lysis of a flywheel energy storage system for voltage sag correction’. IEEE
Transactions on Industry and Applications. 2006;42(1):42–52.
[22] Takahashi R. and Tamura J. ‘Frequency control of isolated power system
with wind farm by using flywheel energy storage system’. Proceedings of
the 2008 International Conference on Electrical Machines, ICEM’08;
Vilamoura, Portugal, Sep 2008, pp. 1–6.
Chapter 9
Chemical energy storage systems: fuel cells
and power-to-gas
9.1 Introduction
A fuel cell (FC) is a static device having energy conversion function. Chemical
energy of a fuel is supplied as an input to the FC, which converts it directly into
electrical energy. Energy conversion results from a chemical reaction of positively
charged hydrogen ions with oxygen or another oxidizing agent. FCs using hydro-
gen (H2) may be called ‘hydrogen FCs’. They are different from batteries in that a
continuous source of fuel and oxygen (or air) are required to sustain the chemical
reaction, whereas the chemicals in batteries react with each other to generate an
electromotive force (EMF). FCs can produce electricity continuously if the proper
inputs are supplied. They can be used, as an alternative energy source with low
emission of pollutant gases, for primary and backup power for commercial,
industrial and residential buildings as well as in remote or inaccessible areas.
Different types of FCs such as polymer electrolyte membrane fuel cell
(PEMFC), solid oxide fuel cell (SOFC) and molten carbonate fuel cell (MCFC)
exist. They have a wide variety of potential applications, particularly in distributed
generation (DG) applications, for example, transportation, stationary power for
buildings, auxiliary power and cogeneration applications.
Compared to conventional power plants, FC systems have many advantages.
These include high efficiency, zero or low emission of pollutant gases and flexible
modular structure. In addition, FCs can normally be located at any site at the dis-
tribution level in a power system resulting in grid reinforcement, deferring or
eliminating the need for system upgrades, and improving system integrity, relia-
bility and efficiency. However, the most common disadvantage of all types of FCs
currently is the cost.
The major processes of a hydrogen FC energy system are depicted in
Figure 9.1. They start with cleaning the fuel and converting it into a hydrogen-rich
gas using the fuel processor. Power-to-gas (PtG) energy system can be used as a
source of fuel to provide the FC with natural gas or hydrogen as discussed in
Section 9.9. FC stacks normally convert the energy of hydrogen into electric energy
in DC form. Then, power conditioner unit converts the electric power from DC into
regulated DC or AC for consumer use. An energy storage system may also be
included as a part of the FC system for energy management and mitigating load
Heat and water

Clean Energy
exhaust storage
system
Fuel Hydrogen-rich
Fuel processor Fuel cell DC Power AC
gas stacks conditioner
Fuel cell system Air (oxygen)
Figure 9.1 Processes of an FC energy system
transient impacts, where bi-directional energy flow is available between FC system

and the storage system. FC stack by-products include heat, water and clean exhaust,
which can be used for space heating or to produce additional electricity.
To produce the desired power and voltage for a specific application, the cells
are bundled together in series and parallel combinations forming what is called ‘FC
stack’. Also, the output power and voltage can be controlled through proper control
of the interfacing power electronic devices.
9.2 Principles of FC operation

The physical structure of all types of FCs is mainly composed of two electrodes
(anode and cathode) separated by a thin layer of electrolyte that only allows posi-
tively charged hydrogen ions (or protons) to move between the two sides of the FC
(Figure 9.2). Electrodes are usually made flat to achieve the maximum possible
contact between the electrode, electrolyte and the gas. They contain catalysts that
cause the fuel to undergo oxidation reactions generating positively charged
hydrogen ions and electrons. Both the electrolyte from one side and the gas from
the other side can penetrate each electrode [1].
To understand how the reaction between hydrogen and oxygen produces DC
electricity, reactions that occur at each individual electrode are considered. An acid
electrolyte is assumed to be used as it is the most common type.
At the anode, the hydrogen gas ionizes, releasing electrons and creating Hþ
ions (or protons) according to
2H2 ! 4Hþ þ 4e (9.1)
The hydrogen ions are drawn through the electrolyte after the reaction. Meanwhile,
electrons are drawn from the anode to the cathode through an external circuit
producing DC current.
Chemical energy storage systems: fuel cells and power-to-gas 219
Electric load
Electric load
Air in Fuel in
e– e–
e–
–
O2 e H+ Hydrogen
H2
Unused H+
gases Excess
out H O fuel Oxygen
2
Cathode Anode
Cathode Electrolyte Anode Electrolyte
Figure 9.2 Physical structure of an FC
At the cathode, hydrogen ions, electrons and oxygen react to form water as the
reaction is expressed by
O2 þ 4e þ 4Hþ ! 2H2 O (9.2)

þ
Continuing the reactions (9.1) and (9.2), H ions are produced at the anode and
must pass through the electrolyte, whereas electrons must pass through an external
electrical circuit to the cathode (Figure 9.3). It is to be noted that the cathode is the
electrically positive terminal of the FC as the electrons are flowing from the anode
to the cathode, that is, from to þ. This holds true for electrolysis, cells, valves,
forward biased diodes and FCs where the cathode is always the electrode into
which electrons flow. Similarly, the anode is always the electrode from which
electrons flow and is the electrically negative terminal.
FCs are classified by the type of electrolyte used and by the difference in start-
up time (from 1 s for PEMFC to 10 min for SOFC). Additional parameters, such as
temperature and efficiency, are given in Section 9.8. For all types, the most
important design features in an FC are the following:
● The anode catalyst breaks down the fuel into electrons and ions.
● The cathode catalyst turns the ions into the waste chemicals such as water or
carbon dioxide.
● Reaction rate, an important factor for FC design. It should not proceed slowly.
To deal with this problem, use of catalysts and raising the temperature can be
applied to any chemical reaction. For FCs, an additional way, which is
increasing the electrode area, is applied because the reaction rate is propor-
tional to the area of the electrode. Increasing the electrode area is a vital issue
to improve the FC performance that is determined in terms of the current per
cm2. For this reason, the electrode area can be increased not only by increasing
Electrical load
4e– 4e–
Positive connection
Oxygen Hydrogen
+
O2 4H 2H2 → 4H+ + 4e–
Negative connection
Cathode Anode
Electrolyte
Water Heat
Figure 9.3 Chemical reactions in an FC with acid electrolyte
Positive terminal
Electrical load
Negative terminal
Figure 9.4 An FC stack, four parallel sets and two series FCs per set (arrows
indicate the electron flow)
both width and length but also by using highly porous electrode, which has the
effect of greatly increasing the effective surface area.
● Individual FCs produce relatively small electric potential, 0.6–0.7 V. So, FCs
are stacked or placed in series and parallel combinations to create enough
voltage and current to meet the specific requirements of an application
(Figure 9.4). FCs produce not only electricity but also water, heat and,
depending on the fuel source, very small amounts of nitrogen dioxide and other
emissions. The energy efficiency of an FC is generally between 40% and 60%,

or up to 85% efficient in cogeneration if waste heat is captured for use [2].
9.3 Reversible open-circuit voltage of FCs

The basic principles of operation of an FC are explained in Section 9.2. It is seen
that hydrogen and oxygen provide input energy, whereas energy in the form of
electricity, heat and water is the output (Figure 9.5).
The energy of chemical inputs (H2 and O2) and outputs (H2O) is not simply
defined, whereas the electrical energy output can be easily calculated as a multi-
plication of the voltage, current and time.
The concept ‘Gibbs free energy’ [1] is used to calculate the chemical energies
in FCs. It is defined as the ‘energy available to do external work, neglecting any
work done by changes in pressure and/or volume’. Applying this definition to the
FC, it is seen that ‘external work’ involves moving electrons around an external
circuit, and any work done by a change in volume between input and output is not
involved by the FC. Energy is all the extracted work including that due to some
parameter changes (e.g. pressure, gas concentration and temperature).
To calculate FC efficiency and open-circuit voltage, the point of zero Gibbs free
energy as well as the change of energy must be specified. When working with
chemical reactions, the point of zero Gibbs free energy is defined as pure elements,
in the normal state, at standard temperature and pressure (25 C, 100 kPa). Also,
when adopting this convention, the term ‘Gibbs free energy of formation’, Gf,
instead of ‘Gibbs free energy’ is commonly used. Therefore, Gf of the input is equal
to zero for an ordinary hydrogen FC operating at standard temperature and pressure.
On the other hand, the change in energy in an FC is the change in Gibbs free
energy of formation, DGf, which gives the energy released. It can be given by the
difference between the Gibbs free energy of the products and that of reactants, that is,
DGf ¼ Gf of products Gf of reactants (9.3)
It may be convenient to express the quantities in (9.3) in their ‘per mole’ form.
Thus, (9.3) can be rewritten, replacing the upper-case letters by lower-case letters
under – to denote quantities/mole, as
Dgf ¼ g f of products g f of reactants (9.4)
Inputs Outputs
Electricity
Hydrogen
Fuel cell Heat
Oxygen Water
Figure 9.5 Inputs and outputs of an FC

As an illustrative example, the basic reaction of the hydrogen/oxygen FC is

2H2 þ O2 ! 2H2 O; which is equivalent to H2 þ ½O2 ! H2 O
It is seen that the reactants are one mole of H2 and half a mole of O2, whereas
the product is one mole of H2O. Applying (9.4) gives
Dgf ¼ ðg f ÞH2 O ðg f ÞH2 ½ðg f ÞO2 (9.5)
The values of Dg f in (9.5) at different conditions (temperature and state) are

given in Table 9.1 indicating that the Gibbs free energy of formation is not con-
stant. More detail about calculation of Dg f is given in [1].
The negative values of Dgf indicate that the energy is released. If the process is
reversible, that is, it has no losses, Gibbs free energy is fully converted into elec-
trical energy.
From the basics of chemistry, a mole of any substance always has the same
number of entities, for example, molecules, which is 6.022 1023 and is called
‘Avogadro’s number’, Na.
Thus, a mole of electrons is 6.022 1023 electrons. Accordingly, its charge is
Na.e, where e is the charge on one electron, equal to 1.602 1019 coulombs. (Na.e)
is called the ‘Faraday constant, F’ or the charge on one mole of electrons. Thus,
F ¼ Na :e ¼ 96485 C:
Referring to Section 9.2, it can be found that for the hydrogen FC, two elec-
trons pass around the external circuit for each water molecule produced and each
molecule of hydrogen used. So, for one mole of hydrogen used, 2Na electrons pass
around the external circuit, then the charge that flows is 2Na.e ¼ 2F coulombs.
Assuming the open-circuit voltage of the FC is denoted by VOC, the electrical
work done moving the charge 2F around the circuit is
Electrical work done ¼ charge voltage ¼ 2F:V OC J (9.6)
As the system is reversible, the electrical work done in (9.6) will be equal to
the Gibbs free energy released, Dg f . Hence,
Dgf
Dgf ¼ 2F:VOC or VOC ¼ (9.7)
2F
For instance, from Table 9.1, Dg f ¼ 210.3 kJ for an FC operating at 400 C, so
210:3 103
VOC ¼ ¼ 1:09 V
2 96;485
It is to be noted that this value is higher than the actual as the system is
assumed to be reversible as well as pure hydrogen and oxygen are at standard
pressure (100 kPa). In practice, because some voltage drop occurs and reactions are
irreversible, the voltage, VOC, would be lower even when no current is drawn.
Table 9.1 Values of Dgf for the reaction H2 þ ½O2 ! H2O at different temperatures
Water state Liquid Liquid Gas Gas Gas Gas Gas Gas Gas
Temperature ( C) 25 80 80 100 200 400 600 800 1,000
Dgf (kJ/mol) 237.2 228.2 226.1 225.2 220.4 210.3 199.6 188.6 177.4
The open-circuit voltage given in (9.7) can specifically be applied to hydrogen

FCs as the number ‘2’ in this equation represents the number of electrons per
molecule of that fuel. Thus, to generalize (9.7) to be applied to any other fuel, the
number ‘2’ can be replaced by ‘n’ to obtain the general formula as
Dg f
VOC ¼ (9.8)
nF
9.4 FC’s efficiency

As explained in Section 9.3, for FCs, Gibbs free energy is converted into electrical
energy. If the process of chemical reaction is a reversible process and Gibbs free
energy is assumed to be totally converted into electrical energy, the efficiency of an
FC would be 100% and may be considered as the efficiency limit as the efficiency,
h, is defined as
Electrical output energy
hFC ¼ (9.9)
Gibbs free energy
This is not an accurate definition for the following reasons:
● Gibbs free energy changes not only with temperature change as seen in
Table 9.1 but also with pressure and other factors as explained in the next
sections. Consequently, efficiency limit is always 100% regardless of the
conditions used.
● The definition of energy given in (9.9) depends on the energy produced by
hydrogen as an input to the FC. In fact, there is a preceding process, which is fuel
burning to release its energy. Therefore, it would be more useful to calculate the
efficiency in terms of the heat produced by burning the fuel. This heat differs
from one fuel to another and is evaluated by its calorific value or the change in
‘enthalpy of information’, Dh f with a negative sign when the energy is released.
Accordingly, the efficiency of an FC can be updated and rewritten as
Electrical output energy

hFC ¼ (9.10)
Dh f
It is noted that for hydrogen FCs, Dh f has two different values that can be used
depending on the form of water (gas or liquid) produced by burning hydrogen. In
case of chemical reaction: H2 þ ½O2 ! H2O, the product water may be in the form
of either steam or condensed back to water. For steam, Dh f ¼ 241.83 kJ/mol. and
is called the ‘lower heating value’, LHV, whereas it equals to 285,84 kJ/mol. for
water and is called the ‘higher heating value’, HHV. Thus, for any value of effi-
ciency, the value (LHV or HHV) at which the efficiency is based on should be stated.
LHV that gives higher efficiency is commonly used, particularly, when HHV is not
known [1].
As given by (9.10), FC efficiency reaches its maximum value, efficiency limit,

when the electrical output energy is maximum, that is, equal to Gibbs free energy
(no waste of heat). So,
Dg f
hFCðmaxÞ ¼ 100% (9.11)
Dh f
For instance, at FC temperature of 80 C, Dg f ¼ 228.2 kJ/mol. as given in
Table 9.1. Using (9.7), the corresponding reversible open-circuit voltage, VOC, is
obtained as
228:2 103
VOC ¼ ¼ 1:18 V
2 96;485
Applying (9.11), the efficiency limit will be
Dgf 228:2
hFCðmaxÞ ¼ 100 ¼ 100 ¼ 79:8% referred to HHV
Dh f 285:84
Similarly, at FC temperature of 400 C, Dgf ¼ 210.3 kJ/mol., then

VOC ¼ 1.09 V and the efficiency limit, hFCðmaxÞ ¼ 73.6% referred to HHV.
From these values, it might be suggested that lower temperatures are better, but
the voltage losses are less at higher temperatures. In practice, FC voltages are
usually higher at higher temperatures and the waste heat is more useful as well.
The actual FC efficiency can be estimated in terms of the operating voltage,
VFC. If all the energy from the hydrogen fuel, Dh f , is converted into electrical
energy, the efficiency would be 100% and (9.7) can be adapted to give the corre-
sponding reversible open-circuit voltage as

Dh f 1:48 V referred to HHV
VOC ¼ ¼
2F 1:25 V referred to LHV
Therefore, the actual efficiency of the cell, hFC ; can be obtained as the ratio of
VFC to VOC value (1.48 V related to HHV or 1.25 V related to LHV), that is,

VFC ðVFC =1:48Þ 100% using HHV
hFC ¼ 100% ¼ (9.12)
VOC ðVFC =1:25Þ 100% using LHV
In practice, it is found that not all the fuel fed to the FC reacts as some of it
passes through the electrolyte unreacted by diffusion (it is known as ‘fuel cross-
over’). A fuel utilization coefficient, mf , is used when calculating the efficiency of
the cell by rewriting (9.12) as
VFC
hFC ¼ mf 100% (9.13)
VOC
where mf ¼ 0.95 for a more accurate estimation.

9.5 Factors affecting VOC

As given by (9.8), VOC depends on Gibbs free energy, Dg f . It is noted that Dgf
changes in a reversible electrochemical reaction with temperature as shown in
Table 9.1. Not only the change of Dgf is due to temperature but also can result from
variation of reactant partial pressure and concentration or system pressure as dis-
cussed in the following.
Consider the reversible reaction in a general form as
bX þ cY ) dM (9.14)
where b and c are moles of reactants X and Y, respectively, and d is the moles of
product M. Each of X, Y and M has an associated activity, a. Thus, ax and ay are
the activities of reactants and am is the activity of the product. In the ideal gases,
the activity a is defined as
P
a¼ (9.15)
Po
where P is the pressure or partial pressure when the gases are part of a mixture and
Po is standard pressure, 100 kPa. In case of water produced in FCs, the activity
depends on whether it is in a steam or liquid form. For steam, the activity is
PH2 O
aH2 O ¼
PoH2 O
where PoH2 O is the vapour pressure of the steam at the temperature concerned and
can be obtained from steam tables. If the water produced is in a liquid form,
aH2 O ¼ 1.
Therefore, Gibbs free energy change for (9.14) should be modified by the
activities of the reactants and product to be as [3]

abX :acY
Dgf ¼ Dgf o RT ln (9.16)
adM
where Dgf o is the change in molar Gibbs free energy of formation at standard
pressure, R is the molar gas constant, 8.314 J/(K mol.) and T is the temperature in
kelvin.
For hydrogen FC reaction: H2 þ ½O2 ! H2O, (9.16) becomes
!
o
aH2 :a0:5
O2
Dgf ¼ Dgf RT ln (9.17)
aH2 O
Typical values of Dg f o are given in Table 9.1. From (9.17), with increasing
activity of reactants, Dg f will be more negative and more energy is released,
whereas Dg f increases and becomes less negative when the activity of product
increases.
Accordingly, (9.7) can be modified by substituting (9.17) to obtain VOC as

!
Dg f o RT aH2 :a0:5
O2
VOC ¼ þ ln
2F 2F aH2 O
Or
!
o RT aH2 :a0:5
O2
VOC ¼ VOC þ ln (9.18)
2F aH2 O
o
where VOC is the open-circuit voltage at standard pressure.
The second term on the RHS of (9.18) determines how the voltage is affected
by both temperature and activities of both reactants and product which are function
of the pressure. The voltage calculated from (9.18) is known as ‘Nernst voltage’,
which is the reversible open-circuit voltage of a cell at given temperature and
pressure, and the equation is called ‘Nernst equation’.
As an illustrative example, the reaction: H2 þ ½O2 ! H2O (steam) is con-
sidered at high temperature and assuming that the steam behaves as an ideal gas,
using (9.15), the activities will be as follows:
PH2 P O2
For reactants: aH2 ¼ Po and aO2 ¼ Po
P
For products: aH2 O ¼ PH2oO
Substituting in (9.18) gives
2 3
PH2 PO2 0:5
o RT 6 Po P o
7
VOC ¼ VOC þ ln4 P H2 O 5
2F
Po
Substituting Po ¼ 1 when all pressures are given in bar, obtains

!
o RT PH2 :P0:5
O2
VOC ¼ VOC þ ln (9.19)
2F PH2 O
The pressures in (9.19) are in most cases partial pressures, for example, the
oxygen is mostly a part of air. Assuming the system pressure is P, the relations in
the following can be written.
PH2 ¼ aP; PO2 ¼ bP and PH2 O ¼ gP (9.20)
where a, b and g are constants depending on the molar masses and concentration of
H2, O2 and H2O, respectively.

0:5
RT ab
VOC ¼ VOCo
þ ln :P0:5
2F g
0:5 (9.21)
o RT ab RT
¼ VOC þ ln þ lnðPÞ
2F g 4F
Therefore, two forms of Nernst equation are presented by (9.19) and (9.21). It
is seen that the second term on the RHS of (9.21) determines the effect of con-
centration, for example, if both a and b are decreased and g increases, the second
term is getting smaller, and so the cell voltage will fall. Similarly, the third term
shows the effect of system pressure. If the system pressure changed from P1 to P2,
the change in cell voltage will be

RT RT P2
DVOC ¼ ½lnðP2 Þ lnðP1 Þ ¼ ln (9.22)
4F 4F P1
which shows that the FC voltage is increased by system pressure.
On the other hand, to discuss the effect of reactants partial pressure, for
example, hydrogen that can be supplied pure or as part of a mixture, (9.19) can be
rewritten by isolating the term of hydrogen pressure as
!
0:5
o RT P O RT
VOC ¼ VOC þ ln 2
þ lnðPH2 Þ
2F PH2 O 2F
With a change in the hydrogen pressure from P1H2 to P2H2 at constant PO2 and
PH2 O , the FC voltage will change by

RT RT P2H2
DVOC ¼ ½lnðP2H2 Þ lnðP1H2 Þ ¼ ln
2F 2F P1H2
Similarly, for oxygen partial pressure change from P1O2 to P2O2 , at constant
PH2 and PH2 O , (9.19) can be rewritten as

o RT PH2 RT
VOC ¼ VOC þ ln þ lnðPO2 Þ
2F PH2 O 4F
and the FC voltage change will be

RT RT P2O2
DVOC ¼ ½lnðP2O2 Þ lnðP1O2 Þ ¼ ln
4F 4F P1O2
Calculations of reversible open-circuit voltage, VOC, of an FC are explained in
Sections 9.3 and 9.5 and given by (9.8). VOC may fall at higher temperatures as
illustrated in Table 9.1 or when gas pressure decreases as in (9.22). It is practically
found that losses due to irreversibility may take place causing voltage drops.
Because of these drops, the voltage at the FC terminals (operating voltage) is lower
o
than the voltage developed inside the cell; VOC . Irreversibility losses can be
classified as activation losses, ohmic losses and concentration losses. The voltage
drops produced by these losses are functions of load current and FC temperature
and/or pressure.
9.5.1 Activation losses

Activation losses are the most important irreversibility, particularly, in low and
medium temperature FCs. These losses result from the slowness of chemical
reactions on the surface of the electrodes that may cause a loss of a proportion of
the generated voltage. The voltage drop is mainly at the cathode for hydrogen FCs
whereas the FCs using fuel other than hydrogen may have a voltage drop at both
electrodes (cathode and anode). At higher temperatures, the impact of activation
losses becomes less important.
The voltage drop produced by the activation losses, DVact, is considered as the
sum of the voltage drop at both electrodes. It can be calculated as a function of the
current density of the FC, i, using an empirical formula deduced from experiments
and is known as ‘Tafel equation’, that is [1,3],

i i
DVact ¼ Aa ln þ Ac ln for i > ioa and i > ioc (9.23)
ioa ioc
where parameters subscripted by ‘a’ and those by ‘c’ are referred to anode and
cathode, respectively, and
A ≙ constant with values higher for slow chemical reactions;
io ≙ constant and represents the current density at which the voltage drop
begins to move from zero. It is higher for faster reactions and called ‘exchange
current density’.
The form of (9.23) can be rewritten as

i
DVact ¼ A ln for i > b (9.24)
b
where
Aa =A
A ¼ Aa þ Ac and b ¼ ioa þ iAocc =A
So, it can be found that io has the most variation of voltage rather than A. It is
also noted that for hydrogen FC, the terms referred to the anode in either (9.23) or
(9.24) can be neglected compared to that of the cathode because ioc is much smaller
than ioa. The constant A can be calculated as
RT
A ¼ Ac ¼
2zF
where
z ≙ charge transfer coefficient. Its value depends on the reaction and the
material of the electrode and lies in the range from 0.1 to 0.5.
Therefore, to effectively reduce the activation voltage drop, it is required to
increase the value of io which can be done by (i) raising the cell temperature,
(ii) using more effective catalysts, (iii) increasing the surface area of the electrode
by increasing its roughness and (iv) increasing both reactant concentration and
pressure.
On the other hand, the current density of an FC, i, in (9.24), may be modified
when considering the energy loss resulting from (i) the small amount of fuel not
passing through the electrolyte (fuel crossover), and (ii) the electron conduction
through the electrolyte which causes a small internal current and produces no
current from the cell. The modification includes a value of internal current density,
iint, being added to i to express this impact on the cell voltage. Accordingly, (9.24)
becomes

i þ iint
DVact ¼ A ln for i > b (9.25)
b
However, it is found that the effect of the internal current and diffusion of fuel
through the electrolyte in an FC is usually not very important except in case of low
temperature FCs as it has a noticeable effect on VOC.
9.5.2 Ohmic losses

In FCs, positive hydrogen ions flow in the electrolyte and electrons go through the
electrodes. So, the resistance to the flow of ions in the electrolyte and the electrical
resistance of the electrodes will cause what is called ‘ohmic losses’, resulting in a
voltage drop which is equal to the product of cell current and the equivalent
resistance. To replace the current (mA) by the cell current density (mA/cm2) to
obtain a relation that is consistent with the others for voltage drop, the resistance
(W) should be replaced by the area specific resistance (kW cm2). Thus, the ohmic
voltage drop, DVohm, can be calculated by
DVohm ¼ i:RASR (9.26)
where
i ≙ the current density (mA/cm2) and RASR ¼ the area specific resistance, ASR
(kW cm2).
Therefore, to reduce this ohmic voltage drop, it is essential to select electrodes
with highest conductivity and appropriate materials for cell interconnects and make
the electrolyte as thin as possible.
9.5.3 Concentration losses

Because of a current being drawn from an FC, the gas pressure (or partial pressure)
will be reduced causing a voltage drop. For example, if oxygen at the cathode is
extracted from air, its concentration may change, which in turn causes a reduction
in pressure, depending on the rate of air circulation around the cathode and
how quickly the oxygen can be replenished. Also, at the anode a reduction of
hydrogen pressure may result due to fluid resistance during hydrogen flow
through tubes.
An empirical formula that is the most common approach used for calculating
the voltage drop produced by concentration losses, DVconc, is given by [1]
DVconc ¼ m expðniÞ (9.27)
where
m ≙ constant, typically about 3 105 V;
n ≙ constant, about 8 103 cm2/mA;
i ≙ FC current density.
9.6 FC equivalent circuit
Due to the activation, ohmic and concentration, voltage drops inside FCs, and the
voltage at the FC terminals is lower than the reversible open-circuit voltage, VOC,
which is the voltage developed inside the FC (it can be denoted by E as well).
These voltage drops as given by (9.25), (9.26) and (9.27) are functions of load
current and FC temperature and/or pressure. The ohmic voltage drop is a linear
function of FC load current but the ohmic resistance, RASR, is a function of FC
temperature. The activation and concentration voltage drops are nonlinear func-
tions of load current, pressure and temperature inside the FC. Specifically, for the
two types of FCs, PEMFC and SOFC, which have a great potential for use in
distributed power generation applications, the equivalent circuit can be deduced as
The output voltage, V, at FC terminals can be written as
V ¼ E DVact DVohm DVconc
i þ iint (9.28)
¼ E A ln i:RASR þ m expðniÞ
b
where DVact , DVohm and DVconc are given by (9.25), (9.26) and (9.27), respectively.
In addition to the factors, load current, temperature and pressure, that affect the
voltage, V, as included in (9.28), there is another factor, which is the electric storage
capability of FCs and called ‘capacitance of double-layer charge effect’. It is com-
mon to most FCs and affects their dynamic response as it makes them behave like a
large capacitor in the order of a few Farads. In an FC, the two electrodes (anode and
cathode) are separated by a thin layer of electrolyte forming two opposite polarity
∆Vact ∆Vohm ∆Vconc

E
Activation Ohmic Concentration

– + i
– V +
Load
Figure 9.6 FC simple equivalent circuit

charged layers across the boundary between the electrodes and electrolyte; anode–
electrolyte layer and electrolyte–cathode layer. These two layers (known as electro-
chemical double layers) can store electrical energy and behave much like a capacitor.
The example shown in Figure 9.7 illustrates the situation that might arise at the
electrode of an acid electrolyte FC. Electrons are collected at the surface of
the porous electrode and the positive hydrogen ions will be attracted to the surface of
the electrolyte forming the effect of a capacitor with capacitance, C, given by
A
C¼e (9.29)
l
where
e ≙ electrolyte’s electrical permittivity;
A ≙ effective surface area between the electrolyte and electrode;
l ≙ distance between the layers.
A is large because of the porous structure of electrodes, whereas l is very small
(in the order of nanometers) resulting in a high value of capacitance.
Therefore, to represent the capacitance of double-layer charge effect the
equivalent circuit of Figure 9.6 should be modified to be as shown in Figure 9.8.
Resistances Rohm, Ract and Rconc model the ohmic, activation and concentration
losses, respectively. The voltage across the capacitance, VC, can be obtained by

dV C
VC ¼ i C ðRact þ Rconc Þ (9.30)
dt
and the output voltage, V is
V ¼ E VC Vohm (9.31)
–
– – +
–
+
+
– +
+ + + +
– – – – +
–
– – –
– +
+ +
+
– +
Electrolyte
+ + +
– – –
– +
Electrode
– – –
– –
+
+ + + +
– +
+ + +
– – –
–
– – +
– +
+
+ +
+ +
– – –
–
+
Figure 9.7 Double-layer charge effect at the electrode surface

Ract Rconc
E Rohm
C
– +
+ – i
VC
– +
V
Figure 9.8 FC equivalent circuit with double-layer charge effect
It is found that the time constant, t, of the circuit in Figure 9.8 only depends on
Ract and Rconc, where
t ¼ ðRact þ Rconc ÞC (9.32)
The changes in these two resistances are only due to the electrochemical
reactions inside FCs, whereas the ohmic resistance is the resistance due to the
ohmic voltage drop; that is, a change in current gives an immediate change in the
voltage drop across Rohm. Thus, the output voltage given in (9.31) is a dynamic
quantity as VC is a function of the time constant, t, given by (9.32).
9.7 PEMFC characteristics

PEMFC is one of the most common types used as a cogeneration unit. To predict its
dynamic response, a computer model can be developed based on electrochemical
and thermodynamic characteristics [4–8].
9.7.1 Calculation of resistances

As shown in Figure 9.8, the different voltage drops due to irreversibility losses are
represented by resistances Ract, Rohm and Rconc. Considering the PEMFC, the
resistances can be calculated as follows.
Activation resistance: Activation losses result in a voltage drop as explained in
Section 9.5.1 and can be calculated by (9.23). In case of PEMFC, from (9.23) and
(9.24) the following relation can be obtained:

I R
Vact ¼ Ac ln ¼ T: ½lnðI Þ lnðIo Þ ¼ T :½a þ B lnðI Þ
Ioc 2zF (9.33)
¼ DVact1 þ DVact2
RT
where Ac ¼ 2zF , I ≙ cell current and Ioc ≙ constant and represents the cathode current
at which the voltage drop begins to move from zero. DVact1 is the activation voltage
drop component affected only by FC internal temperature and DVact2 is the compo-
nent affected by both current and temperature. From (9.33), it can be found that
DVact2 ¼ T :B lnðI Þ (9.34)
R
where B ¼ 2zF .
The equivalent resistance of activation is defined as: DVact2 =I. Thus,

T :B lnðI Þ
Ract ¼ (9.35)
I
Ohmic resistance: It composes of polymer membrane resistance, conducting
resistance between the membrane and electrodes and electrodes resistances. The ohmic
resistance is current and temperature dependent and can be expressed as a sum of three
components: constant, current-dependent and temperature-dependent components.
Rohm ¼ Rohm0 þ kRI I þ kRT T (9.36)
where
Rohm0 ≙ the constant component;
kRI and kRT ≙ empirical constants for calculating Rohm in W/A and W/K,
respectively.
Concentration resistance: It can be obtained by calculating the ratio of con-
centration voltage drop, given by (9.27), and the FC current. Thus,
m expðniÞ
Rconc ¼ (9.37)
I
Resistances Ract, Rohm and Rconc are defined by (9.35), (9.36) and (9.37).
Accordingly, the equivalent circuit can be redrawn as shown in Figure 9.9. It is like
that of Figure 9.8, but by adding a voltage source, DVact1 ; with polarity opposite to
E to satisfy (9.33).
9.7.2 PEMFC computer modelling

It is to be noted that the output voltage is a function of temperature and load
current, whereas the cell temperature and the voltage across the capacitance of
double-layer charge effect are a function of time during a transient state. So, the
output voltage is a dynamic quantity.
A computer model based on the block diagram shown in Figure 9.10 can be
developed. The inputs are anode pressure Pa, cathode pressure Pc, initial FC tem-
perature Tinitial and room temperature Troom. The overall fuel and oxidant delay
effect are employed in addition to the other effects (irreversibility and double-layer
charge effects) to determine the internal potential of the FC. This effect is repre-
sented by DVd, which is subtracted from the right side of (9.19). It is a function of
Ract Rconc
E ∆Vact1 Rohm
+ l C
– +
I
+ V –
C
– +
V
Figure 9.9 PEMFC equivalent circuit with double-layer charge effect

Thermodynamic
calculations
Calculation of to get fuel cell
open circuit temperature T
voltage E
by (9.19)
Pa
T
Pc +
–
Pinitial
Calculations of irreversibility Vout
Proom Fuel and oxidant
losses and voltage drops:
delay, ∆Vd by
+ ∆Vact, ∆Vohm, ∆Vconc
(9.38), (9.39)
by (9.25)–(9.27)
–
Double-layer charge effect given Electric
by (9.33) load
Inputs Outputs Feedback
Figure 9.10 Block diagram for PENFC dynamic computer model
time and determines the influence of the FC and oxidant delay on the cell output
voltage during load transients. It can be given in Laplace domain as [3]

1 te s
DVd ¼ le I ðsÞ 1 ¼ le I ðsÞ (9.38)
te s þ 1 te s þ 1
where le is a constant (W) and te is the overall flow delay. In time domain, it
becomes
DVd ¼ le ½iðtÞ iðtÞ expðt=te Þ (9.39)
where is the convolution operator and t=te is the time lag.
By simulation and based on the block diagram of Figure 9.10, a typical
example for a PEMFC stack with specification given in Table 9.2 shows the steady-
state voltage–current and power–current (V–I and P–I) characteristics as depicted
in Figure 9.11.
It is shown that the P–I characteristic of the FC at given temperature and
pressure has only one maximum power point (MPP) at which the efficiency is
maximum. The output power and voltage can be controlled through proper control
and MPP tracking algorithms associated with power electronic interface devices to
improve FC efficiency [9–11].
9.8 Types of FCs
FCs have been classified into different types usually distinguished by the electro-
lyte that is used [12]. Some key figures can be referred to for identifying each type
Table 9.2 Specification of PEMFC stack
Maximum power (Pmax) 613.5 W

Voltage at Pmax (Vmp) 36.1 V
Current at Pmax (Imp) 17 A
Short-circuit current (ISC) 20 A
Open-circuit voltage (VOC) 54 V
55
600 Voltage–current chs
Power–current chs MPP
50
500
45
Fuel cell power (W)
Stack voltage (V)

400
40
300
35
200 30
100 25
0 20
0 2 4 6 8 10 12 14 16 18 20
Fuel cell current (A)
Figure 9.11 V–I and P–I characteristics for a 613.5 W PEMFC stack
such as operating temperature, produced power, cell efficiency and cost per kilo-
watt. These figures for some types of FCs are given in Table 9.3. Cell efficiency
means the efficiency of the FC as an individual element, whereas system efficiency
expresses the efficiency of the overall system including all the components shown
in Figure 9.1.
9.9 Power-to-gas
The concept of PtG technology is about converting electrical energy into chemical
energy (gaseous energy) in the form of hydrogen and/or methane as energy carriers.
Therefore, linking the electricity and gas networks permits effective dealing with
excess and lack of electrical power generated, especially by intermittent renewable
sources (e.g. wind and solar) [13].
9.9.1 Principle of operation

To address the problem of electrical energy conversion into gaseous energy, the
concept of PtG technology can be used as a solution. The core of this concept is
water electrolysis to yield hydrogen and oxygen, then hydrogen conversion with an
oxidation source to methane. The produced hydrogen and methane can be injected
into a gas distribution grid or gas storages to be utilized in all other natural gas users
Table 9.3 Comparison of some FC types
FC type Electrolyte Power (W) Operating Efficiency (%)

temperature ( C)
Cell System
Direct methanol FC Polymer membrane (ionomer) 100 W–1 kW 90–120 20–30 10–25
Reformed methanol FC Polymer membrane (ionomer) 5 W–100 kW 250–300 50–60 25–40
Proton exchange membrane FC Polymer membrane (ionomer) 1 W–500 kW 50–100 50–70 30–50
Phosphoric acid FC Molten phosphoric acid (H3PO4) <10 MW 150–200 55 40 and
cogen. 90
Solid acid FC Hþ-conducting oxyanion salt 10 W–1 kW 200–300 55–60 40–45
(solid acid)
Molten carbonate FC Molten alkaline carbonate 100 MW 600–650 55 45–55
Tabular solid oxide FC (TSOFC) O2-conducting ceramic oxide <100 MW 850–1,100 60–65 55–60
Planar solid oxide FC O2-conducting ceramic oxide <100 MW 500–1,100 60–65 55–60
Wind
Electric power grid

Solar
O2 Fuel cell Chemical and
Vehicles industrial H2
Water Electrolyser
H2
Hydrogen Hydrogen (H2) grid

storage
Gas power plants
Biomass, CHP
waste, H2
atmosphere Heating
CO2 CO2
tank Methanation Methane (CH4) grid
CO
Industrial flue gas
Figure 9.12 PtG concept
such as FC vehicles, chemical industries, combined heat and power (CHP) and
heating as shown in Figure 9.12. Therefore, PtG systems include the main com-
ponents as in the following [14].
Water electrolyser: The surplus electrical energy is delivered from the elec-
trical power grid as input to the electrolysis device to split the water into its com-
ponents, hydrogen and oxygen through the chemical reaction expressed by
H2 O ! H2 þ ½O2
Three electrolysis technologies are commonly used in PtG process: alkaline
(AEL), polymer electrolyte membrane (PEM) and solid oxide electrolysis (SOEC).
Hydrogen storage: Because of the fluctuation of power supplied by renewable
sources, temporary hydrogen storage is of utmost importance to be employed to
overcome the intermittent operation of the electrolyser. The storage capacity and its
operating parameters depend on the amount of electrical power required to supply
the demand. Different methods may be applied to establish hydrogen storage
compressed gas tanks as follows:
1. Cryogenic compressed liquid hydrogen tanks, which have continuous gas
leakage and expensive insulation. However, this type of tanks is a poor choice
for PtG systems.
2. Physical storages, which are normally underground caverns and the cheapest option
of all storage methods. In addition, this type is suitable and favourable for high
capacity, long-term hydrogen storage, that is, not viable for small PtG systems.
3. High-pressure cylindrical gas tanks, which are the best option for temporary
hydrogen storage.
4. Metallic hybrid tanks, which are an alternative method of the third method.
Methanation: It is a process to synthetically produce natural gas, that is,
methane (CH4) as the name is referred to. Carbon dioxide (CO2) or carbon mono-
oxide (CO) is required for methanation and can be obtained from either biogenic
carbon source (e.g. biogas) or industrial processes [15]. Thermodynamically, the
reaction can be expressed as
CO2 þ 4H2 ! CH4 þ 2H2 O
Or
CO þ 3H2 ! CH4 þ H2 O
Methanation can be done in methanation reactors either biological or catalytic
by use of catalyst, e.g. nickel. The gases produced by PtG technology (H2 and CH4)
can be combined to form a combined natural gas (CNG), which can in turn be
integrated into natural gas infrastructure. Figure 9.13 shows the sequence of pro-
cesses of a typical PtG system [16–18].
9.9.2 Applications
The natural gas produced by PtG system can be utilized in stationary power appli-
cations such as CHP plants and small industries by respectively using gas turbines
(GTs) and microturbines, which are described in the next two subsections. However,
CHP covers a wide capacity range from 25 kW microturbines to 250 MW GTs [19].
9.9.2.1 Gas turbines

GTs are designed to produce high-quality thermal output for most CHP applica-
tions and district heating systems as they are a cost-effective alternative for com-
mercial and industrial end users and reliable power generation device. GT size is
available to supply the demand of many end users up to about 250 MW. GTs
operate on either gaseous fuel (natural gas and synthetic gas) or stored liquid fuel
(oils) for backup.
O2 Biological Industrial
carbon source flue gas
(CO2) (CO)
Utility grid Electric current Electrolysis cells to split
Renewable energy sources the water (H2O) as CO2 CO
(wind, solar, etc.) H2O → H2 + ½O2
H2 Methanation
H2 CO2 + 4H2 → CH4 + H2O
Electrical power to grid
H2 storage Or
Water, H2O CO + 3H2 → CH4 + H2O
Chemical industries CH4

Transport sector
H2
Gas to power plants Grid injection (NG)
NG
Figure 9.13 Typical flow diagram of PtG processes (inputs flow is denoted by
black arrows and interior or output flows by grey arrows)
The technology of GTs is mainly based on Brayton thermodynamic cycle,

which includes compression of atmospheric air, feeding and ignition of fuel and
expansion of the heated combustion gases through the gas producer and power
turbine. The provided power from the turbine as a prime mover can drive the
compressor and rotate the electrical generator. Accordingly, the main components
of a GT are shown in Figure 9.14. Simple-cycle industrial GTs electrical efficiency
may vary within the range of 25% to 40%. Development of large GTs using
combined cycles yields an increase in GT electrical efficiency to become
approximately 60% and more economic installations.
Further improvement of electrical efficiency can be achieved by heat recovery.
It is to be noted that there is no heat recovery in the exhaust gas of the simple cycle.
Hot exhaust gas can directly be used in an additional process including a heat
recovery steam generator (HRSG) in a combined cycle as depicted in Figure 9.15.
The exhaust heat can generate steam to drive a steam turbine which in turn rotates
an electrical generator.
Fuel
Combustor
Compressor
Gas producer Generator

Power
turbine
turbine
Figure 9.14 GT components
Gas turbine Fuel
Combustor
Simple cycle water/

heat recovery
Feed water
HRSG Steam
turbine (combined cycle)
Figure 9.15 Heat recovery-GT system

Exhaust Recuperator
Fuel
Air inlet
Combustor
Compressor
Power Generator
turbine
Figure 9.16 Microturbine components
9.9.2.2 Microturbines
Microturbine is a class of small GTs and operates with the same theory as that of
large GTs but a difference in the design due to use of recuperator to recover part of
the exhaust heat for preheating the combustion air. The air is drawn through a
compressor and mixed with fuel (natural gas, gasoline, diesel and alcohol), then
ignited to power the turbine and rotating an electrical generator (e.g. synchronous
generator and permanent-magnet generator) at a speed up to 120,000 rpm as shown
in Figure 9.16.
References
[1] Larminie J. and Dicks A. Fuel cell systems explained. 2nd edn. West Sussex,
England: John Wiley & Sons Ltd.; 2003.
[2] Wikipedia. Fuel cell [online]. Available from https://en.wikipedia.org/wiki/
Fuel_cell [Accessed 24 Aug 20018].
[3] Nehrir M. H. and Wang C. Modeling and control of fuel cells: Distributed
generation applications. Hoboken, NJ, USA: IEEE-Wiley Press; 2009.
[4] Wang C., Nehrir M. H. and Shaw S. R. ‘Dynamic models and model vali-
dation for PEM fuel cells using electrical circuits’. IEEE Transaction on
Energy Conversion. 2005;20(2):442–51.
[5] Nehrir M. H., Wang C. and Shaw S. R. ‘Fuel cells: Promising devices for
distributed generation – Understanding their modeling and need for control’.
IEEE Power Engineering Magazine. 2006;4(1):47–53.
[6] Maher A. R. and Al-Baghdadi S. ‘Modelling of proton exchange membrane
fuel cell performance based on semi-empirical equations’. Renewable
Energy. 2005;30:1587–99.
[7] Ramos-Paja C. A., Giral R., Martinez-Salamero L., Romano J., Romero A.
and Spagnuolo G. ‘A PEM fuel-cell model featuring oxygen-excess-ratio
estimation and power-electronics interaction’. IEEE Transactions on
Industrial Electronics. 2010;57(6):1914–24.
[8] Lee J. H., Lalk T. R. and Appleby A. J. ‘Modeling electrochemical perfor-
mance in large scale proton exchange membrane fuel cell stacks’. Journal of
Power Sources. 1998;70:258–68.
[9] Karami N., Outbib R. and Moubayed N. ‘Maximum power point tracking
with reactant flow optimization of proton exchange membrane fuel cell’.
Journal of Fuel Cell Science and Technology. 2013;10:1–14.
[10] Ramos-Paja C. A., Spagnuolo G., Petrone G. and Giral R. ‘Fuel cell MPPT
for fuel consumption optimization’. IEEE International Symposium on
Circuits and Systems (ISCAS); Paris, France, May/June 2010, pp. 2199–202.
[11] Karami N., Outbib R. and Moubayed N. ‘Fuel flow control of a PEM fuel
cell with MPPT’. IEEE International Symposium on Intelligent Control
(ISIC); Dubrovnik, Croatia, Oct 2012, pp. 289–94.
[12] Chambers A. Renewable energy in nontechnical language. Tulsa,
Oklahoma: PennWell Corporation; 2004.
[13] Götz M., Lefebvre J., Mörs F., et al. ‘Renewable power-to-gas: A techno-
logical and economic review’. Renewable Energy. 2016;85:1371–90.
[14] Liu W., Wen F. and Xue Y. ‘Power-to-gas technology in energy systems:
Current status and prospects of potential operation strategies’. Journal of
Modern Power Systems and Clean Energy. 2017;5(3):439–50.
[15] Mostert C., Ostrander B., Bringezue S. and Kneiske T. M. ‘Comparing
energy storage technologies regarding their material and carbon footprint’.
Journal of Energies. 2018;11(12):3386–411.
[16] Walker S. B., Lanen D. V., Fowler M. and Mukheriee U. ‘Economic analysis
with respect to power-to-gas energy storage with consideration of various
market mechanisms’. International Journal of Hydrogen Energy. 2016;
41(19):7754–65.
[17] Schulze P., Holstein J., Noort A. V. and Knijp J. ‘Power-to-gas in a dec-
arbonized European energy system based on renewable energy sources’.
White paper; European POWERTOGAS. Available from http://www.afhy
pac.org/documents/European%20Power%20to%20Gas_White%20Paper.
pdf [Accessed 8 Feb 2020].
[18] Jentsch M., Trost T. and Sterner M. ‘Optimal use of power-to-gas energy
storage systems in an 85% renewable energy scenario’. Energy Procedia.
2014;46:254–61.
[19] Onsite Sycom Energy Corporation. ‘Review of combined heat and power
technologies’. Office of Industrial Technologies, US Department of Energy;
October 1999, pp. 1–31.
Chapter 10
Electrochemical energy storage systems
Electrochemical energy storage (EcES) systems are technologically mature for

practical use. The electricity is stored as chemical energy, which can be delivered
in the form of electrical energy using electrochemical reactions. They include all
types of battery energy storage (BES): (i) conventional secondary (rechargeable)
BES such as lead-acid (PbA), nickel cadmium (NiCd), nickel metal hydrate
(NiMH), lithium ion (Li-ion), metal air (Me-air), sodium sulphur (NaS) and sodium
nickel chloride (NaNiCl) batteries and (ii) flow batteries (FBs) such as vanadium
redox (VR) and hybrid FBs.
Batteries are built in different sizes with capacity ranging from <100 W to
several megawatts and round efficiency from 60% up to 80% depending on the
operational cycle and the type of electrochemistry. BES systems can be used in
most stationary and mobile applications such as electric vehicles, marine and
submarine missions, aerospace operation and so on. In addition, for grid-scale
application, deep cycle-type batteries are the most commonly used with efficiency
in the range of 70%–80%. A brief description of each type of batteries is illustrated
in the forthcoming sections [1,2].
10.1 Conventional secondary BES technologies
10.1.1 PbA batteries

These types of batteries are the oldest and most widely used rechargeable electro-
chemical devices. It consists of electrodes of lead metal (Pb) with high surface area
as negative electrode (i.e. cathode) and lead oxide (PbO2) as anode, and both are
immersed in an electrolytic solution of sulphuric acid (H2SO4).
A double sulphate reaction takes place inside the battery. In the discharging
state, oxidation and reduction reactions at the anode and cathode, respectively,
result in formation of lead sulphate (PbSO4), and the electrolyte loses its dissolved
sulphuric acid and becomes primarily water. Thus, the electrochemical reactions
are
At anode: Pb þ SO24 $ PbSO4 þ 2e

2 þ
At cathode: PbO2 þ SO4 þ 4H þ 2e $ PbSO4 þ 2H2 O
Overall chemical reaction : Pb þ PbO2 þ 2H2 SO4 $ 2PbSO4 þ 2H2 O
The reactions are reversible in the charging state. The two states, discharging
and charging, are shown in Figure 10.1.
PbA batteries have the advantages of low cost and high reliability and effi-
ciency. They are commonly used for power quality and some spinning reserve
applications.
However, because of their short cycle life and low energy density their appli-
cation to energy management is limited.
10.1.2 Nickel-based batteries

Depending on the type of metallic cathode, different types of nickel (Ni) batteries
exist. The anode of all types is made of the active material ‘nickel oxyhydroxide’
and the electrolyte is an aqueous solution of potassium hydroxide (KOH). Whereas
a metal such as cadmium (Cd), metal hydride (MH), iron (Fe), zinc (Zn) or
hydrogen (H2) is used as cathode, the electrochemical discharging reactions of
nickel-based batteries are
At anode: Cd þ 2OH $ CdðOHÞ2 þ 2e
At cathode: NiOðOHÞ þ H2 O þ e $ OH þ NiðOHÞ2
Overall chemical reaction: 2NiOðOHÞ þ Cd þ 2H2 O $ 2NiðOHÞ2 þ CdðOHÞ2
NiCd and NiMH batteries are common, commercialized and widely used in
applications of uninterruptible power supply (UPS), telecommunication and elec-
tronic devices. NiFe and NiZn batteries have limited and specific applications
Load
Discharging state Charging state

2e− 2e− 2e− 2e−
(PbSO4) Lead sulphate
(PbSO4) Lead sulphate

(PbO2) Lead dioxide
(Pb) Lead
Water and
sulphuric acid
Sulphuric acid
Figure 10.1 Discharging and charging of a typical lead acid battery

Electrochemical energy storage systems 245
because of their high maintenance, low specific energy, short cycle life and high
cost as well as the efficiency is about 75%. NiH2 battery has long cycle life, large
ratings and high overcharging/discharging tolerance, but on the other hand, it has
low volumetric energy density and high self-discharge.
However, NiCd batteries have generally the following advantages and
disadvantages.
Advantages
● Self-discharge rate at 10% per month.
● Efficiency of 70%–90%.
● Energy density of 40–60 Wh/kg.
● Large number of life cycles (200–2,500).
● Nominal cell voltage of 1.2 V.
● Capability of operation at temperatures of 20 to þ40 C.
● Reliable and do not require much maintenance.
● Depth of discharge (DoD), a measure of how much charge is drained from a
battery in one cycle, has been developed to be only 10% and accordingly the
life cycles reach 50,000.
Disadvantages
● The toxic heavy metal ‘cadmium’ causes environmental hazard and it has been
prohibited for consumer use.
● Relatively high cost as the manufacturing process is expensive.
● Suffering from ‘memory effect’ as the full charge takes place only after a
series of full discharges.
NiMH batteries have a nominal capacity much higher than that of equivalent NiCd
and PbA batteries. Their design has been developed to have the advantages of NiCd
batteries with the memory effect mitigated, operating at a wide temperature range,
the environmental impact eliminated, increased specific energy (60–120 Wh/kg),
increased life cycles (S3,000) and very low internal resistance. So, this type of
batteries is very useful for hybrid electric vehicle applications and sensitive
electronic devices. On the other hand, high self-discharge rate (5%–30% per
month), DoD time dependence, life-cycle degradation and overcharging sensitivity
are the limitations of NiMH batteries.
10.1.3 Li-ion batteries

In this kind of battery, the anode is made of layered graphitic carbon and the
cathode is a lithiated metal oxide (LiCoO2, LiMO2 and LiNiO2). The electrodes
(anode and cathode) are separated by a foil to make electrical isolation immersed in
an electrolyte of dissolved lithium salts in organic carbonates (LiPF6). The nominal
voltage of each cell of the battery is 3.7 V, i.e. three times the cell voltage of nickel
batteries.
The operation of Li-ion battery is based on reversible extraction of ions on
the porous electrodes. During charging process, the lithium atoms in the cathode
become ions and travel through the electrolyte towards the anode to combine with
external electrons and are deposited between the carbon layers as lithium atoms.
The process is reversed during the discharging phase as shown in Figure 10.2. The
overall chemical reactions is given as
LiC6 þ CoO2 $ C6 þ LiCoO2
Li-ion batteries are employed in both portable and grid-scale stationary
applications. They have high energy density (~200 Wh/kg), long cycle life (3,000
cycles at 80% DoD), high efficiency (>90%), rapid response time (in milli-
seconds), fast charge/discharge capabilities and low self-discharge rate (<8% per
month). So, Li-ion batteries are also applied in electric vehicles because of their
reliable chemical and thermal characteristics. For large-scale applications, the main
hurdle is the high cost as special packaging and internal overcharging protection
circuits are required.
10.1.4 Me-air batteries

An Me-air electrochemical cell consists of the anode made from pure metals such
as calcium (Ca), magnesium (Mg), Fe, aluminium (Al), lithium (Li) and Zn and a
Discharge
Charge
Foil (separator) Metal
O2 O2 O2
Li+
Li+ Li+
O2 O2 O2
Li+
Li+ Li+
Li+
Li+
O2 O2 O2
Li+ Li+
Li+ O2 O2 O2
Li+ Li+ Li+

Li+
Li+ O2 O2 O2
Li+ Li+
Li+ O2 O2 O2
Layered graphitic Electrolyte Lithiated metal oxide

carbon electrode (LiPF6) electrode
Figure 10.2 Basic components of Li-ion batteries

cathode connected to an inexhaustible air supply provided by open-cell structure.

The oxygen in the air is only used for the electrochemical reaction. Among the
various Me-air batteries, the Li-air is most attractive as its theoretical energy density
excluding oxygen, which is not stored in the battery, is 11.14 kWh/kg (100 times
more than that of other battery types as well as more than that of petrol (10.15 kWh/kg)).
On the other hand, Li-air battery has a high safety risk because of the fire that can result
from the high reactivity of lithium with air and humidity.
At present, only a rechargeable Zn-air battery with a theoretical specific energy
excluding oxygen of 1.35 kWh/kg is technically feasible. It has some properties of
fuel cells and conventional batteries: the zinc is the fuel; the reaction rate can be
controlled by varying air flow and oxidized zinc/electrolyte paste can be replaced
with fresh paste.
As shown in Figure 10.3, in discharging phase, the electrons are liberated
through the electrochemical reaction of the zinc anode (Zn ! Zn2þ þ 2e). In the
presence of the electrolyte, an oxygen reduction reaction takes place at the air
cathode resulting in the formation of hydroxide ions (O2 þ2H2O þ 4e ! 4OH).
These hydroxide ions migrate from cathode to anode and form the insoluble zinc
oxide (ZnO). During charging phase, the reversible redox reaction takes place at the
electrode–electrolyte interface, which stores the electrical energy in the Zn-air
battery. The electrochemical reactions inside the battery are given as
At anode: Zn þ 4OH ! ZnðOHÞ24 þ 2e
In fluid: ZnðOHÞ24 ! ZnO þ H2 O þ 2OH
At cathode: O2 þ H2 O þ 4e ! 4OH
Overall chemical reaction : 2Zn þ O2 ! 2ZnO
10.1.5 NaS batteries

NaS battery has the capability of providing high density (151–170 Wh/m3), is more
than 85% efficient, has long cycle life (2,500 cycles at 90% DoD) and a long
discharge period (~6 h). Furthermore, NaS can provide the response that makes it
very effective in the application of power quality regulation of the utility grid.
These properties are suitable for the utilities to use it as a commercial energy
storage for distribution grid support, wind integration and high-value grid services.
The structure of NaS comprises liquid (molten) sulphur at the positive elec-
trode and liquid (molten) sodium at the negative electrode as active materials
00
separated by a solid beta alumina ceramic (b -Al2 O3 ) electrolyte as shown in
Figure 10.4.
The electrolyte allows only the positive sodium ions (Naþ) to flow through it
and combines with sulphur (S) to form sodium polysulphides (Na2S4) as the elec-
trochemical reactions are
At cathode: 2Zn $ 2Naþ þ 2e
At anode: 4S þ 2Naþ þ 2e $ Na2 S4
Overall chemical reaction : 2Za þ 4S2 ! 2ZnO
Charge
e− e−
Discharge
Load
e− e−
Air
O2
OH−
Zinc metal
(Zn)
Porous
carbon
Catalysts Gas diffusion layer

Electrolyte
Catalytic active
Separator layer
Figure 10.3 Zn-air battery representation
In discharging phase, the sodium is oxidized at the beta alumina ceramic

interface forming Naþ ions, which go through the electrolyte and electrons flow in
the external circuit of the battery producing 2.0 V. The charging phase is the
reverse where the Naþ is released from sodium polysulphides back through the
electrolyte to recombine as elemental sodium.
The operation temperature of NaS battery is kept at 300–350 C, and thus a
heat source is required, which uses the own stored energy of the battery resulting in
partial reduction of its performance.
10.1.6 NaNiCl batteries

NaNiCl batteries are also known as ZEBRA batteries, which are high temperature
(~300 C) devices and can operate across a broad temperature range (40 to þ70 C).
Beta 2V
alumina
tube
Molten
Na
S
Na2S4
Na+
Discharge
Sulphur
Figure 10.4 Main components of NaS battery
They are useful for such applications as time shifting, UPS, peak shaving, smart grid
management, renewable generation optimization, frequency regulation and so on.
NaNiCl2 secondary battery as an energy storage system is based on electro-
chemical charge/discharge reactions consisting mainly of a positive electrode
(cathode) made of Ni and sodium chloride (NaCl) and a negative electrode (anode)
that is typically made of sodium (Na). The electrodes are separated by a beta-
alumina ceramic wall that is conductive for sodium ions but an isolator for elec-
trons. A schematic diagram of the main components is depicted in Figure 10.5.
Charge
2NaCl + Ni NiCl2 + Ni
Discharge
Capillary gap
Na
Wick
Na+
Liquid electrolyte Ceramic electrolyte

(NaAlCl) β"- Al2O2
Figure 10.5 Main components of NaNiCl battery
This beta-alumina ceramic acts as an electrolyte and enables the conduction of

sodium ions between the anode and the cathode of the cells. The battery tempera-
ture is maintained between 270 and 350 C to keep the electrodes in a molten state;
that is, independent heaters are part of the battery system.
The key performance data can be summarized as power range of several MW,
energy range from 4 kWh up to MWh, discharge time from 2 h to several hours,
cycle life of 4,500 cycles, life duration less than 15 years, reaction time of
some milliseconds, efficiency in the range of 85% to 95%, energy density about
100–120 Wh/kg, etc.
10.2 Flow batteries

Unlike the conventional secondary batteries where the energy is charged/
discharged in the active materials of the electrodes, in the FB, which is also a
rechargeable battery, the energy is stored in liquid active electrolytes of two
separate external tanks. The electrolytes are pumped through the electrochemical
cells stack that converts the chemical energy directly into electrical energy and
vice versa.
FBs are fully reversible by applying the DC voltage to the electrodes during
charging process. They have the following attractive features and characteristics:
● The power rating is defined by the size and design of the electrochemical cells
in the stack whereas the energy storage capacity depends on the volume of
tanks and electrolyte concentration.
● Operation near to ambient temperature.
● Long discharge times.
● Capability of high overloading for short times.
● Response time in milliseconds.
● Low self-discharge in sealed storage tanks.
● Long charge/discharge life cycles.
● High round trip efficiency.
Therefore, with these features, FBs can be fitted to a wide range of stationary
applications. Based on the electroactive components of electrolyte, FBs are clas-
sified into redox flow batteries (RFBs) and hybrid flow batteries (HFBs).
10.2.1 Redox flow batteries

The RFB has two liquid electrolyte dissolutions containing dissolved metal ions as
active masses, which are pumped to the opposite sides of the electrochemical cell.
The electrolytes at the negative and positive electrodes are called anolyte and
catholyte, respectively.
One of the most mature FB technologies is the vanadium redox flow battery
(VRFB) in which a common electrolyte, for example, vanadium sulphate dissolved
in sulphuric acid is used to store the energy in charged ions in different valance
states. During charging/discharging process, the Hþ ions will be exchanged
between the vanadium redox couples of bivalent/trivalent (V2þ/V3þ), as oxidizing
agent, in the anolyte tank and tetravalent/pentavalent (V4þ/V5þ), as reducing agent,
in the catholyte tank. Anolyte and catholyte flow through porous electrodes,
separated by a membrane, which allows the ions (Hþ) to pass through it for the
electron transfer process as shown in Figure 10.6 [3]. During the exchange of
charge, a current will flow over the electrodes that can be used by a battery-
powered device. The electrodes are continually supplied with the dissolved active
masses from the tanks; once they are converted, the resulting product is removed to
the tank. The electrochemical reactions are
At anode: V2þ $ V3þ þ e

At cathode: V4þ $ V5 þ e
Overall chemical reaction : V2þ þ V5 $ V3þ þ V4þ
VRFBs have the advantage of using ions of the same metal on both sides. They
are commercialized in wide range of stationary applications with power ratings
Power source/
load
e− e−
e
an
de
de
br
tro
tro
em
ec
Charge Charge
ec
El
El
Tank Tank
V2+/V3+ V2+ V3+ V4+ V5+ V4+/V5+
Anolyte Catholyte
Discharge Discharge
Ions
Pump Pump
Figure 10.6 Schematic diagram of VRFB structure
from kilowatts to megawatts for power quality improvement, UPS, load levelling
and for integration and support of intermittent renewable energy resources. Its
standard open-circuit voltage per cell is ~1.26 V.
10.2.2 Hybrid flow batteries

HFEs combine the features of VRFBs and RFBs as one of the active materials is
internally stored within the electrochemical cell, whereas the other is liquid elec-
trolyte stored externally in a tank. Therefore, their capacity depends on the size of
the electrochemical cell. Zinc bromine (ZnBr) can be used as anolyte yielding an
HFB of a cell voltage (1.85 V), energy density (35–75 Wh/L), operating tempera-
ture (20–50 C), life cycles of 10–20 years, large discharge duration (8–10 h) and
large DoD capability [4].
10.3 Battery modelling

Because of the high non-linearity of batteries operation, accurate models tend to be
more complicated mathematical models. The impact of several factors such as
charge/discharge rates, battery temperature, battery type and battery age on battery
operation should be involved [5]. Therefore, electrochemical models are designed

to consider the chemical, thermodynamic and physical qualities of the batteries and
are more precise and complex. For instance, estimation of battery state of charge
(SoC) and DoD must be as accurate as possible. It exerts additional control over
charging/discharging processes, which can be employed to increase battery life.
SoC is defined as the ratio of the available capacity of a battery to the battery-rated
capacity; that is, how much charge is left in the battery with respect to its nominal
capacity. While physically, it is the ratio of the remaining active material to the
total active material inside the battery that can be converted into electrical energy
from the chemical energy. It can be mathematically estimated at time t by
Ðt
Ahnom 0 I ðtÞdt
SoC ðtÞ ¼ (10.1)
Ahnom
where SoC(t) ¼ 100% at t ¼ 0, I(t) is the battery current (positive and negative
during discharge and charge, respectively) and Ahnom is the nominal battery capa-
city in ampere-hour. It is easy to compute the integral discharge/charge current if
this current is constant. Actually, the current and thus its integral varies because of
the loss variation in battery capacity. In this case, (10.1) will change so that the
relation is evaluated in small time intervals (Dt) using (10.2), and the SoC at kth
time interval is obtained by (10.3) [6].
Ik Dt
DSoC k ¼ (10.2)
Ahnom
SoC k ¼ SoC k1 þ DSoC k (10.3)
To get exactly the value of DoD, as a measure of charge drained from a battery
in one cycle, it equals to the change in the SoC in one discharge cycle, that is,
DoD ¼ DSoC (10.4)
As an example, a DoD of 60% means that the battery is discharged in its first cycle
from an SoC of 100% until SoC of 40% remains. As battery ageing takes place, the SoC
of a fully charged cell after certain number of cycles decreases to 97%. The next dis-
charge cycle with a DoD of 60% will then occur between an SoC of 97% and 37% [7].
Modelling of electrochemical batteries in terms of equivalent circuit is another
way to model the batteries. It is necessary and easier, especially for simulation of BES
in power system applications. It enables the electrical engineers to simulate the battery
behaviour in different operating conditions, static and dynamic, as well as to compute
useful parameters (e.g. short-circuit currents) for analysing a grid incorporating such
batteries and renewable resources. Battery characteristics are predetermined and
constant throughout operation in a static battery model, whereas battery voltage,
current, charge and temperature are varied and dynamically affecting the battery
output in the dynamic model.
The ideal equivalent circuit model is characterized by a voltage source (Em)
only and does not represent internal battery characteristics. Adding an impedance
Im R0 I C1
P R2 R0 I
P
+
R1 Im Ip
Ip
VPN V
Zm(s, θ, SoC) Rp
Zp(s, θ, SoC)
Em
Em(θ, SoC) Ep(θ, SoC)
−
N
N Main branch Parasitic
Main branch Parasitic branch
branch
(a) (b)
Figure 10.7 Equivalent circuit: (a) battery and (b) LA battery
in series with the voltage source to represent the battery internal impedance forms a
simplified model, which can be improved by considering the values of both
impedance and voltage source dependent on the battery SoC, temperature q and the
Laplace variable s. Further improvement can take place by adding a parasitic
branch in parallel as shown in Figure 10.7(a). The parasitic branch represents the
irreversible reactions that often present in the battery. The voltage across this
branch is separated from the terminal voltage by a resistance, R0. The current Ip
circulates in this branch. It is a part of the total current I and is a lost current that
cannot be restored. The charge stored in the battery is only the integral of Im (the
current entering the battery in the main reversible reaction branch).
The equivalent circuit in Figure 10.7(a) can be adapted to model a specific type
of batteries, for example, lead acid battery as shown in Figure 10.7(b), where all
elements are functions of SoC, charging/discharging currents and electrolyte tem-
perature q. The model includes two main parts: the main branch, which approx-
imates the battery dynamics, and the parasitic branch, which accounts for the
battery behaviour at the end of a charge process in addition to a terminal resistance,
R0. The main branch consists of the following [8,9]:
● Internal electromotive force (EMF) or open-circuit voltage of one cell, Em in
volts. It is assumed to be constant when the battery is fully charged and is
given by
Em ¼ Em0 KE ð273 þ qÞð1 SoC Þ (10.5)
where Em0 is the open-circuit voltage at full charge in volts, KE is a constant in
V/ C and q is the electrolyte temperature in C.
● Resistance R1, which is varied with depth of charge (DoC) and is a measure of
battery’s charge adjusted for the discharge current. It is calculated by
R1 ¼ R10 lnðDoC Þ (10.6)
where R10 is a constant in ohm.
● Capacitance C1 to model a voltage delay when battery current changed. Its

value is given by
t1
C1 ¼ (10.7)
R1
where t1 is a main branch time constant in seconds.

● Main branch resistance R2 in ohm, which affects the battery during charging
and becomes relatively insignificant for discharge current. It can be computed
by the relation
exp½A21 ð1 SoC Þ
R2 ¼ R20 (10.8)
1 þ exp A22 IIm
where R20 is a constant in ohm, A21 and A22 are constants, Im is the main branch
current in ampere and I* is the nominal battery current in ampere.
The parasitic branch current, Ip, is the loss current when the battery is being
charged. It depends on the electrolyte temperature and the voltage across the
parasitic branch. Therefore, representing the parasitic branch by linear model as
shown in Figure 10.7(a) is unrealistic as the dependence of Ip on the voltage VPN is
highly non-linear. However, experiences and experiments yield that Zp can be
assumed to be a resistor that is independent of Laplace variable s and SoC. Thus, it
is preferred to express this branch in a way that satisfies the algebraic and non-
linear nature as shown in Figure 10.7(b). Ip is very small under most conditions
except during charge at high SoC and is given by [10]

VPN = tp s þ 1 q
Ip ¼ VPN Gp0 exp þ Ap 1 (10.9)
VP0 qf
where VPN is the voltage across the parasitic branch, Gp0 is a constant in seconds, tp
is a parasitic branch time constant in seconds, VP0 is a constant in volts, Ap is a
constant, q is the electrolyte temperature in C and qf is the electrolyte freezing
temperature in C.
The terminal resistance R0 in ohm is assumed constant at all temperatures, and
varied only with SoC according to the relation
R0 ¼ R00 ½1 þ A0 ð1 SoC Þ (10.10)
where R00 is the value of R0 in ohm at SoC ¼ 1 and A0 is a constant.
10.4 Comparison of BES technologies
Selection of battery technology as an energy storage is based on both battery

characteristics, investments needed and application requirements. Therefore, tech-
nical and economical comparison can help the engineers to select the proper type of
battery to achieve the desired performance at as lowest a cost as possible, that is,
the cost-effective option.
10.4.1 Technical comparison

EcES systems can be distinguished by the mechanisms used to store energy.
Batteries such as rechargeable Li- or Na-based store energy within the electrode
structure through charge transfer reactions. Compared to fuel cells, which are not
rechargeable, the energy is stored in the reactants fed externally to the cells. In
addition, both batteries and fuel cells differ from RFBs, which store energy in the
redox species that are continuously circulating through the cells.
Batteries have desirable features (pollution-free, high round-trip efficiency,
flexible power and energy characteristics) to meet different grid functions, long
cycle life, low maintenance and renewable resources integration. Various types of
batteries can be used for the applications of power quality, transmission and dis-
tribution (T&D) support, for example, load shifting and bridging power at different
discharging duration and rated power as illustrated in Figure 10.8 [11] with power
and energy densities as shown in Figure 10.9. Advantages and disadvantages of
each EcES technology as well as the field of applications are given in Table 10.1.
10.4.2 Economical comparison

For economical comparison, the cost of different battery technologies should be
analysed as it is one of the most important factors in the criterion of EcES selection.
Power quality Bridging power Bulk power management
100
management
Bulk power
10 NaS
VRFB
Li-ion
ZnBr
Discharge time (h)
1 PbA
Bridging NiCd
0.1 power
NiMH
Power quality
0.01 NaS
0.001
0.0001
0.001 0.01 0.1 1 10 100 1,000
Rated power (MW)
Figure 10.8 Comparison of discharge time and power rating for different
BES technologies [11]
300
Li-ion Li-ion
LiFePO2-C Li(TM)O2-C
Specific power (W/kg)
200
LiMP
NiMH
NaMCl
100
PbA
NiCd
NaS
0 VRFB
0 50 100 150 200
Specific energy (Wh/kg)
Figure 10.9 Power and energy densities for different rechargeable batteries [12]
In the literature, two methodologies exist to study the cost of each type: total
capital cost (TCC) and life cycle cost (LCC) [13–15].
● TCC: It is expressed as cost per kWh or cost per output power rating (kW) and
evaluates all expenses of the purchase, installation and delivery of an EcES
system including the following:
– Cost of power conditioning system (PCS) in per unit of power capacity
(cost/kW).
– Energy storage related costs comprise all costs that should be covered
for installations of the storage system and expressed in per unit of stored
or delivered energy (cost/kWh).
– Balance of power costs including costs of grid interface and integration
facilities, cost of land and accessibility and other services that are not
comprised in both CPCS and Cstor and expressed in per unit of power
(cost/kW) or energy (cost/kWh).
● LCC: It includes TCC in addition to the following:
– Fixed operation and maintenance (O&M) expenses.
– Variable O&M expenses.
– Replacement cost.
– Disposal and recycling costs.
LCC can be introduced in levellized annual costs (cost/kW year).
A typical economical comparison of various EcES technologies is given in
Table 10.2.
Table 10.1 Overview of electrochemical storage technologies
Class Storage Main Main Power Energy

technology advantage disadvantage application application
HT- LA Low capital Limited cycle life ■ ●
batteries cost
aLA-Cee Low capital Low energy density ■ ■
cost
NaS High power Cost and requirements ■ ■
and energy to run at HTs
density
Li-ion High power Cost and increased ■ ▲
and energy control circuit needs
density
ZEBRA High power Withstand limited ■ ■
and energy overcharge and
density discharge
NiCd- Low capital Cost and requirements ▲ ■
NiMH cost to run at HTs
Flow ZnBr Independent Medium energy ▲ ■
batteries power and density
energy
VRBr Independent Medium energy ▲ ■
power and density
energy
■ Acceptable and fully capable. ▲ Acceptable for this application. ● Feasible but not practical or
economical.
Table 10.2 Typical cost of different battery technologies
Battery type Conventional battery Flow battery

PbA NiCd NiMH Li-ion NaS NaNiNl Zn-air VRFB HFB
Power cost 1.75–6 5–15 6–18 12–40 32–40 1.5–3 1–2.5 14–37 18–20
Energy cost 1.5–4 6–24 7.2–28.8 4–25 3–5 2.3–3.45 0.6–1.6 5–8 1–7
Power and energy costs are multiplied by 100 $/kW and 100 $/kWh, respectively.
References
[1] Aneke M. and Wang M. ‘Energy storage technologies and real-life appli-
cations – A state of the art review’. Applied Energy. 2016;179:350–77.
[2] Beaudin M., Zareipour H., Schellenberglabe A. and Rosehart W. ‘Energy
storage for mitigating the variability of renewable electricity sources: An
updated review’. Energy for Sustainable Development. 2010;14:302–14.
[3] Vetter M. ‘Battery system technology’. Workshop; Berlin, Germany, April

2011. Fraunhofer ISE; 2011, pp. 1–34.
[4] Park J., Nam G. and Cho J. Perspective of Zn-air batteries [online]. 2015.
Available from https://www.csm.ornl.gov/BLI8/presentations/J.%20CHO-
UNIST-BYI8.pdf [Accessed 8 Feb 2020].
[5] Leao J. F. A., Hartmann L. V., Correa M. B. R. and Lima A. M. N. ‘Lead-
acid battery modeling and state of charge monitoring’. Twenty-Fifth Annual
IEEE Applied Power Electronics Conference and Exposition (M. B. R.
APEC); Palm Springs, CA, USA, Feb 2010. IEEE; 2010, pp. 239–43.
[6] Sparacino A. R., Reed G. F., Kerestes R. J., Grainger B. M. and Smith Z. T.
‘Survey of battery energy storage systems and modeling techniques’. PES
General Meeting; San Diego, CA, USA, July 2012. IEEE; 2012, pp. 1–8.
[7] Dogger J. D., Roossien B. and Nieuwenhout F. D. J. ‘Charaterization of Li-
ion batteries for intelligent management of distributed hybrid-connected
storage’. IEEE Transactions on Energy Conversion. 2011;26(1):256–63.
[8] Jackey R. A. ‘A simple, effective lead-acid batter modeling process for
electrical system component selection’. SAE International Transactions.
2007;116:219–27.
[9] Moubayed N., Kouta J., El-Ali A., Demayka H. and Outbib R. ‘Parameter
identification of the lead-acid battery model’. 33rd IEEE Photovoltaic
Specialists Conference; San Diego, CA, USA, May 2008. IEEE; 2008,
pp. 1–6.
[10] Ceraolo, M. ‘New dynamical models of lead-acid batteries’. IEEE
Transactions on Power Systems. 2000;15(4):1184–90.
[11] Lucas A. and Chondrogiannis S. ‘Smart grid energy storage for frequency
regulation and peak shaving using a vanadium redox flow battery’.
Electrical Power and Energy Systems. 2016;80:26–36.
[12] Dunn B., Kamath H. and Tarascon J. ‘Electrical energy storage for the grid:
A battery of choices’. Journal of Science. 2011;334(6058):928–35.
[13] Zakeri B. and Syri S. ‘Electrical energy storage systems: A comparative
life cycle cost analysis’. Renewable and Sustainable Energy Reviews.
2015;42:569–96.
[14] Chen H., Cong T. N., Yang W., Tan C. and Li Y. ‘Progress in electrical
energy storage system: A critical review’. Progress in Natural Science.
2009;19(3):291–312.
[15] Nadeem F., Hussain S. M., Tiwari P. K., Goswami A. K. and Ustun T. S.
‘Comparative review of energy storage systems, their roles and impacts on
future power systems’. IEEE Access. 2018;7:4555–85.
Chapter 11
Electrostatic, magnetic and thermal
energy storage
11.1 Electrostatic energy storage systems

Electrostatic energy storage systems are used for storing electrical energy in the
form of charge and then producing electrical power back when required.
Supercapacitors (SCs) are devices that can store energy in the form of elec-
trostatic field. SCs are also known as electric double-layer capacitors (DLCs) or
ultra-capacitors. Supercapacitor energy storage (SCES) has similarities with
both batteries and conventional capacitors. However, unlike batteries, SCs store
electrical energy, not chemical energy. Also, unlike conventional capacitors,
SCs contain moving ions. SCs take the form of cells in which the energy is
stored in the form of electrostatic field between two electrodes separated by a
thin layer of insulator or dielectric medium. This is the same principle as the
classical capacitors except that the insulating material is replaced by liquid
electrolyte ionic conductor between the electrodes, which are made of porous
carbon [1].
11.1.1 Working principle of SCES

The principle of SCs is based on the creation of a double electrochemical layer via the
accumulation of electrical charge at the interface between an ionic solution (electro-
lyte) and an electronic conductor (electrode). Unlike batteries, there is no reaction of
oxidoreduction. A deposit of activated charcoal on an aluminium film is used to
obtain a large surface and, thus, a high specific capacity. The electrodes are sub-
merged in an aqueous or organic electrolyte (e.g. potassium hydroxide or sulphuric
acid). The electrical charges are stored at the electrode–electrolyte interface and a
porous membrane through which ions transfer between the electrodes takes place to
prevent their physical contact as shown in Figure 11.1(a). Thus, a DLC is created
between each electrode (A and B) and electrolyte. The corresponding equivalent cir-
cuit is depicted in Figure 11.1(b), where it comprises a series combination of two
double-layer capacitances, CA and CB, with electrode ohmic resistances, RA and RB,
respectively, in parallel. The two double layers are separated by the ohmic resistance
Porous carbon Porous carbon

electrode A Electrolyte electrode B
CA CB
RS
RA RB
Separator
Double layer
(a) (b)
Figure 11.1 SC representation: (a) typical physical representation and

(b) the equivalent circuit
of the electrolyte-filled gap, RS, that is as narrow as possible. The capacitance, C, of a

plate capacitor is calculated by the following [2]:
A
C ¼ e0 er (11.1)
d
where e0 is the dielectric permittivity of free space, er is the electrolyte permittivity,
A is the anode surface area and d is the thickness of the double layer.
It is to be noted that (i) the use of liquid electrolyte and porous carbon elec-
trodes yields high specific surface area, as well as (ii) the extremely short gap
(~1 nm) between them results in the capability of producing high capacitance
values (many thousand farads).
11.1.2 Electrical performance

Capacitance of dielectric capacitors is measured in microfarads, nanofarads or
sometimes in picofarads. Electrochemical capacitors tap the full range of up to
several thousand farads because of their very high energy density. Farad is defined
as the ability to store 1 C of electric charge per volt of potential difference between
the plates, that is, 1 F ¼ 1 C per volt or 1 A per volt. The term ‘capacity’ used in
battery technology indicates the charge storage Q, in units of coulombs or
ampere hours.
Specific energy. Depending on the state of charge (SoC), the voltage of a
capacitor increases or decreases, respectively, on charging or discharging. The
energy stored, W, by a capacitor is calculated in terms of the stored electric charge,
Q ¼ CV , by the following relation:
1 1 Q
W ¼ QV ¼ CV 2 & C ¼ (11.2)
2 2 V
The capacitance per volume of SCs is 10,000-fold higher than that achievable
with the best regular capacitors. Such SCs deliver specific energy of 4–9 Wh/kg
(14–32 kJ/kg) and specific peak power in the range of 1–50 kW/kg [2].
Electrostatic, magnetic and thermal energy storage 263
Discharge behaviour. Like the most battery chemistries, less than 10% of the
stored energy can be used without affecting cycle life. SCs cannot deliver the full
charge in a short time as the voltage of an SC drops exponentially with time from
rated voltage to zero. However, about 75% of the stored energy can be used at
constant current discharge in the range from 100% to 50% of the rated voltage.
Power. SCs are the components dedicated to power storage rather than energy
storage. Their specific power density is 10 to 100 times higher than that of con-
ventional batteries (only very high-power lithium batteries can reach the same
specific power density). They have significant mass power, theoretically of the
order of 10 kW/kg, supporting between 500,000 and 1 million charge/discharge
cycles. Several elements, each with a maximum voltage of 2.7–2.85 V, can be
connected in series and parallel to achieve the desired voltage value and capacity.
In addition, fast discharge rate of SCs is substantially higher than that of most
batteries. Electric power delivered by the capacitor to a load is given by
P ¼ VI ¼ I ðV0 IRi Þ (11.3)
and the thermal loss is expressed as
Pth ¼ I 2 Ri (11.4)
where V0 is the rated voltage, I is the battery current and Ri is the internal
resistance.
Efficiency. For batteries and SCs, the electrical efficiency, hel , is given by the
voltage efficiency, hV , multiplied by the Coulombic efficiency, hC , that is,
hel ¼ hV hC (11.5)
hC ¼ Q=Q0 (11.6)
hV ¼ V =V0 (11.7)
where Q is the electric charge removed during discharge, Q0 is the amount of
charge that has been supplied during charge (mostly, hC 99% as there is no role
for side reactions during charging), V is the average voltage during discharge and
V0 is the average voltage during charging.
The discharge efficiency, a, and the pulse power of an SC, at constant current
discharge in the voltage window between the fully charged and fully discharged
conditions at ¾V 0 , depend on internal resistance as indicated by the following
relations [2]:
IR
a¼1 (11.8)
ð3=4ÞV0
and
ð3=4V0 Þ2 3
P ¼ ð1 aÞ ¼ V0 I (11.9)
R 4
For a battery,
V IR
a¼ ¼1 (11.10)
V0 V0
and
V02
P ¼ a ð1 a Þ (11.11)
R
Reversibility and cycle life. The ideal electrostatic charge commonly takes
place at the electrode/electrolyte interface, and the material is not subject to any
irreversible process and phase changes. Unlike batteries, SCs provide theoretically
unlimited cyclability. Cycle lives up to several millions have been observed even
for deep discharges (more than 75% of the stored energy), fast rates of voltage
change and at low temperatures. Because of the distributed electrolytic and contact
resistances within the pore structure of high-area electrode materials in-depth
charging and discharging encounter rate limitations. This is opposite to batteries in
which slow Faradaic processes involve phase changes of the electroactive materi-
als, which limit cycle life to some thousand charge/discharge acts, depending on the
current and temperature.
11.1.3 Lifetime aspect

The lifetime of an SC can reach one million cycles (or 10 years of operation)
without any degradation, except for the solvent used in the capacitors as it dete-
riorates in 5–6 years regardless of the number of cycles. In addition to the lifetime,
other terms used in practice are the following:
● Shelf life, which is the maximum storage time under open circuit conditions. It
depends on self-discharge.
● Service life or calendar life, which is the time between initial use and the end
of useful life, including periods of rest and standby operation.
● Operating life, which denotes the actual time of operation.
● Float life, which reflects the lifetime under a certain voltage for long periods
of time.
● Cycle life, which is determined by the number of useful charge/discharge
cycles.
11.1.4 Applications
For low- or high-power applications that require fast charging, SCs are used.
Examples of such applications are alarm systems, portable power supplies, auto-
mated meter reading equipment, flashlights and so on. Electric propulsion systems
in all-electric cars, plug-in hybrids and forklifts require electric short-time storage
for load levelling. The capacitor provides peak power density for acceleration or
uphill driving, whereas the battery system covers the base load for constant speed.
The SC can be advantageously recharged by regenerative braking, where the
wheels drive a dynamo which generates electricity.
Furthermore, SCs are also adequate for moderate power demand applications
such as rechargeable stand-alone power supplies and motor start-up for large
engines (e.g. diesel trucks and railroad locomotives).
In power systems incorporating microgrids, SCES systems offer important
benefits to the power system operation such as the following [3]:
● Increasing the penetration of renewable energy resources.
● Decreasing the investment required for new transmission and distribution lines
or building new power plants, which in turn lowers the operational cost.
● Providing reserve capacity that can help improve system’s stability and relia-
bility in addition to improving the power quality as the start speed is faster than
generators.
● Providing an effective mechanism for load management aiming to reduce the
cost of power supply at peak hours, and thus lower the electricity price.
Moreover, electricity price can be stabilized as the SCES system can avoid
energy disruption trading and the loss caused by prediction errors.
● Renewable resources such as wind and solar energy generation are not dis-
patchable because of their dependency on the weather. With the aid of SCES
system, a microgrid that includes such resources becomes dispatchable. This
yields more economical energy exchange planning between the microgrid and
the utility.
11.1.5 Advantages and disadvantages

SCs offer the following advantages:
● Long lifetime.
● Operation over a wide temperature range.
● Lifetime reaches one million cycles (10 years of operation).
● Durablity.
● High reliability.
● Environmentally friendly.
● Easy to recycle or neutralize.
● The efficiency is around 90%.
● Discharge times are in the range of seconds to hours.
● High specific power density.
● Large number of charge/discharge cycles.
On the other hand, the disadvantages of SCs are the following:
● Low energy density.
● Short storage duration.
● High self-discharge loss.
● High investment costs.
11.2 Superconducting magnetic energy storage

Magnetic coils can store energy in the form of electromagnetic field, analogous to
the capacitors that store the energy in the form of electrostatic field. Large flowing
currents in the coils are necessary to store a significant amount of energy and
consequently the losses, which are proportional to the current squared, will also be
high. Thus, the focus on superconducting coils is important as the resistance of the
coils becomes zero in the superconductivity state. To achieve this state, the
superconducting material must be cooled to a very low temperature, typically
below 9.25 K for classical superconducting materials. The cooling process can be
achieved with liquid helium (liquefaction point at 4.2 K). However, the transition
temperature to get the material into the superconductivity state and cooling liquid
depend on the type of material itself, for example, liquid nitrogen for materials with
a transition temperature above 77 K.
A schematic design of a superconductive magnetic energy storage (SMES)
device is shown in Figure 11.2. The entire SMES unit includes the following
subsystems:
● large superconducting coil with the magnet (SCM), which is contained in a
Dewar at the desired cold temperature,
● power conditioning system (PCS) to interface the AC utility and SCM by
which the power is converted from AC to DC or inversely allowing the SMES
system to respond in milliseconds,
● the cryogenic system to cool the SCM below the critical temperature at a
desired value, and
● the controller to perform different functions: controlling the modes of charge/
discharge/standby by controlling the voltage across the SCM, controlling
cryogenic system and controlling the PCS.
PCS Super-
Icoil conducting coil
Transformer Power
Coil protection
AC conversion
utility unit Vcoil
grid
Bypass
switch Dewar vessel
Cryogenic
Controller system
Figure 11.2 Typical SMES unit components

The core element of an SMES system is the SCM of high inductance (Lcoil in
henrys). It stores energy in the magnetic field generated by DC current (Icoil
in amperes) flowing through the coil. Therefore, the inductively stored energy (E in
joules) and the rated power (P in watts) are the common specifications for SMES
devices. They are expressed as
1 2
E ¼ LIcoil (11.12)
2
dE dIcoil
P¼ ¼ LIcoil ¼ Vcoil Icoil (11.13)
dt dt
During the charging process, the superconducting coil is fed with direct current
from the inverter. The electric current induces a constant magnetic field in which
the energy is stored. The discharging process begins with connecting the coil to an
external load by a switching system. The energy is then supplied by the magnetic
field that drives a current. The magnetic energy and the current are decreased
during discharging.
11.2.1 Applications
SMES systems are short-term storage systems for very short periods (power
storage). They can be used in three scales of energy storage: (i) as an
uninterruptable power supply (UPS) of the order of several kWh [4], (ii) smoothing
of production or demand on a local scale (1–1,000 MWh) [5,6] and (iii) stabiliza-
tion of grids by modulating power transmission, flexible AC transmission, of the
order of 1,000 MWh or greater [7–9].
SMES systems offer advantages for these applications compared to flywheels
and SCs, especially with respect to efficiency, response time and lifetime as given
in Table 11.1 [10]. However, SMES systems are costly and have low storage power
density.
11.3 Thermal energy storage
Thermal energy storage (TES) is a technology that stocks thermal energy by

heating or cooling a storage medium so that the stored energy can be used later
for heating and cooling applications and power generation. TES systems have
Table 11.1 Comparison of alternative storages
Storage Efficiency Power density Response time Lifetime

system (kW/m3)
Flywheel 0.7–0.9 10–100 Dozens of minutes 100,000 cycles
SMES >0.9 0.1–5 Milliseconds to a few seconds 20–30 years
SC 0.9–0.95 1–10 Seconds to a minute 500,000 cycles
advantages of very high round-trip efficiency exceeding that of other storage

options, better reliability and less pollution of the environment. TES offers high
storage efficiency for multiple hours at the modest cost as the investment and
running costs are reduced. Additionally, it allows stored power to match peak
demand periods, making solar thermal energy technologies competitive with gas
turbines as capacity power sources [11].
TES is becoming important for electricity storage in combination with con-
centrating solar power (CSP) plants, which are explained in Chapter 3, Section 3.3,
where solar heat can be stored for electricity production when sunlight is not
available. Therefore, it was developed to assist with several applications of energy
production and heating, including supplementing the loading of utility scale solar
thermal power plants. Based on the method of storage, TES technologies can be
classified into three types: sensible heat storage (SHS), latent heat storage (LHS)
and heat of reactions as illustrated in the next section.
11.3.1 Classification and characteristics of TES

TES requires the increase or decrease in the internal energy of a substance by
heating or cooling. Thermal energy is stored in materials that are classified by one
of three methods by which they store energy as heat (Figure 11.3).
● SHS involves a temperature change of the storage material.
● LHS requires an isothermal phase change of the storage material (melting,
freezing, vaporization, fusion and crystallization).
● Heat of reaction or thermochemical heat (TCH) is the result of a reversible
thermochemical reaction by the storage material.
Typical parameters (capacity, power, efficiency, storage period and cost) that
can characterize each type are given in Table 11.2 [12]. In general, to be eco-
nomically competitive in commercial solar systems, it is desirable for TES to use
materials that possess a large change in internal energy per unit volume and/or mass
to minimize the space needed to store the desired amount of energy. However, the
choice of storage medium depends on the nature of the process. As an example, for
water heating energy storage as sensible heat of stored water is logical.
TES
medium of storage technology
Storage Method Storage
Thermo-
Thermal
chemical
SHS LHS Heat of reaction Heat pump
Liquids Solids Hybrid Solid–liquid Liquid–gas Solid–solid
Figure 11.3 Types of TES

Table 11.2 Typical parameters of TES systems
TES Capacity Power Efficiency Storage Cost

system (kWh/t) (MW) (%) period (€/kWh)
SHS 10–50 0.001–10.0 50–90 Days/months 0.1–10
LHS 50–150 0.001–1.0 75–90 Hours/months 10–50
TCR 120–250 0.01–1.0 75–100 Hours/days 8–100
11.3.2 Sensible heat storage

It is the simplest method of TES as well as it is cheap, and there are no risks
associated with the use of toxic materials. It can be hot or cold potential energy
with respect to the ambient temperature, stored in a solid, liquid (e.g. water, sand,
molten salts, bricks, concrete or rocks), or dual medium that consists of some
combination of liquid and solid media. For residential and industrial applications,
water is the most popular and commercial heat storage medium.
The amount of stored energy, Ws, in sensible heat depends on mass, specific
heat and temperature change of the storage medium during the process of charging/
discharging:
ð T2
Ws ¼ mCp dt ¼ mCp ðT2 T1 Þ (11.14)
T1
where Ws represents the heat energy stored in kilojoules, m represents the mass of
the material in kilograms, Cp is the specific heat capacity (kJ/kg K) and T1 and T2
are the initial and final temperatures of the material, respectively, in kelvin.
SHS materials chosen for TES with sensible heat must be thermally stable and
undergo no phase change during temperature extremes. The substance should also
have a high heat capacity, high density and an acceptably low vapour pressure. To
be economically viable, it must be inexpensive.
11.3.3 Latent heat storage

It is also known as phase-change storage. Latent heat is the amount of energy in the
form of heat released or absorbed by a material during a change in physical state or
phase transition such as solid to liquid (melting) or liquid to gas (vaporization). The
heat is mainly stored in the phase-change process. As heat is added to the material,
its temperature increases until a phase change occurs. During the phase change, the
solid bonds are broken at a constant temperature as heat is added to the material.
This is the fundamental difference between sensible and latent heat. When the
phase change is completed (e.g. from solid to liquid), the material again undergoes
an internal energy and temperature increase with the application of heat. LHS is
attractive for phase-change materials (PCMs) that undergo a significantly high
change in internal energy during a phase change. It has the advantages of high-
energy storage density and the isothermal nature of the storage process.
The stored capacity, Ws (in joules), with a PCM medium that undergoes a
phase change from solid to liquid or liquid to gas or vice versa is given by
ð Tm ð Tf
Ws ¼ mCps dt þ mf Dq þ mCp1 dt
Ti Tm (11.15)

¼ m Cps ðTm Ti Þ þ f Dq þ Cp1 ðTf Tm Þ
where Tm is the melting temperature ( C), m is the mass of PCM medium (kg), Cps is
the average specific heat of the solid phase between Ti and Tm (J/(kg K)), cp1 is the
average specific heat of the liquid phase between Tm and Tf (J/(kg K)), f is the melt
fraction and Dq is the latent heat of fusion (J/kg). For instance, Glauber’s salt has Cps
1;950 J=ðkg CÞ, cp1 3;550 J=ðkg CÞ and Dq ¼ 2:43 105 J=kg at 34 C.
LHS via PCMs. The energy associated with a change of phase for a given
material can be used to store energy. The phase change may be one of melting or
evaporating, or it may be associated with a structural change, e.g. in lattice form or
content of crystal-bound water. When heat is added or removed from a given
material, several changes may take place successively, or in some cases simulta-
neously, involving phase changes as well as energy storage in thermal motion of
molecules (i.e. both latent and sensible heat). The total energy change, which can
serve as energy storage, is given by the change in enthalpy [11].
The primary advantage of PCMs is their ability to store energy at reduced
temperatures and in reduced quantities. Furthermore, their high heat of fusion and
other thermal properties allow freezing with little supercooling, no segregation,
chemical stability and a sharp melting point. One advantage is that some PCMs
have low thermal conductivity in the solid state, making high heat transfer rates
necessary during the freezing cycle. They are often flammable, requiring the
addition of safety features to the design of storage vessels. In addition, latent heat is
much more difficult to transfer than sensible heat [11,13].
11.3.4 TCH energy storage

Thermochemical energy is stored as the bond energy of a chemical compound.
During a thermochemical reaction, atomic bonds are broken through a reversible
chemical reaction and are catalysed by an increase in temperature, which allows the
energy to be stored. After thermochemical separation occurs, the constituents are
stored apart until the combination reaction is desired. Recombination of the bonds
between atoms releases the stored thermochemical energy.
Advantages of thermochemical storage include high energy density, long-term and
low-temperature storage capability. However, the thermochemical process is complex,
and the thermochemical materials are often expensive and can be hazardous [13].
Reversible reactions with reactants and products that can be stored easily as
liquids and solids are of practical interest for energy storage because of their high
density. Reactions that produce two distinct phases such as a solid and a gas are
also desirable since the separation of products to prevent back-reaction is

facilitated.
The turning temperature, T0 , is defined as the temperature at which the reac-
tants and products are present in approximately equal quantities. When T > T 0 , the
endothermic storage reaction is favoured as it means that heat is necessary for and
absorbed during the reaction. Conversely, for T < T 0 , the exothermic reaction
dominates, and heat is a product of the reaction.
11.3.5 Storage systems

Two-tank direct storage. In a two-tank direct storage system, the material used to
store thermal energy is the same as the heat transfer fluid (HTF) used to collect the
thermal energy. The fluid is stored in two tanks: one at a high temperature and the
other at a lower temperature as shown in Figure 11.4. The process follows the steps
given below:
● fluid from the lower temperature tank flows through a solar collector or
receiver;
● where solar energy heats it to the high temperature and it then flows back to the
high-temperature tank for storage;
● fluid from the high temperature tank flows through a heat exchanger;
● where it generates steam for electricity production; and
● the fluid exits the heat exchanger at a low temperature and returns to the low-
temperature tank.
High-
temperature
tank
Parabolic Steam 4
solar
collector 3 Generator
Heat
exchanger Turbine
Low-
temperature
1 tank
Figure 11.4 Schematic diagram of a two-tank direct storage

It is recommended to use molten salt as an HTF rather than the common fluids,
for example, oils. This allows the solar field to operate at higher temperature, which
in turn avoids the need for expensive heat exchangers. The use of a cheap heat
exchanger and molten salt as an HTF will substantially reduce the total cost of the
TES system.
Two-tank indirect storage. The two-tank indirect system operates in the same
manner as the two-tank direct system except that different fluids are used for heat
transfer and storage. This system is beneficial especially when the HTF is very
expensive or not suited for use as the storage fluid. As shown in Figure 11.5, the
cycle of the process includes the following: (i) the storage fluid from a low-
temperature tank flows through an extra heat exchanger. Thus, it is heated by the
high-temperature HTF; (ii) the high-temperature storage fluid then flows back to
the high-temperature storage tank; (iii) the fluid exits the heat exchanger at a low
temperature and returns to the solar collector or receiver, where it is heated back to
the high temperature; and (iv) storage fluid from the high-temperature tank is used
to generate steam in the same way as the two-tank direct system. So, the indirect
system requires an extra heat exchanger. This adds cost to the system and decreases
the overall TES efficiency.
Single-tank thermocline storage. A single-tank thermocline storage system
stores thermal energy in a solid medium, usually silica sand, located in a single
tank. At any time during operation, a portion of the medium is at high temperature
and a portion is at low temperature. The hot and cold temperature regions are
High-
temperature
tank
Parabolic Steam
solar
collector Generator
Heat
exchanger Turbine
HTF
Low-
temperature
tank
Figure 11.5 Concept of two-tank indirect storage

separated by a temperature gradient or thermocline. High-temperature HTF flows

into the top of the thermocline and exits at the bottom at low temperature. This
process moves the thermocline downward and adds thermal energy to the system
for storage. Reversing the flow moves the thermocline upward and removes ther-
mal energy to generate steam and electricity. Buoyancy effects create thermal
stratification of the fluid in the tank, which helps stabilize and maintain the
thermocline.
The basic thermocline storage tank concept in which hot and cold materials are
stored inside the tank is shown in Figure 11.6 [14]. Thermocline between the two
cut-off temperatures, cold and hot, Tc,cut-off and Th,cut-off, respectively, is unusable
energy. For T > Th,cut-off, the thermocline cannot be charged because the HTF is
very hot to flow into the solar field for heat collection. When T < Th,cut-off, the HTF
temperature is not high enough to support thermal power cycle. It is to be noted that
the effective capacity of thermocline TES is less than the actual size of the tank.
The thermocline technology has proved advantageous as the reduction of
materials used for constructing the tank and storing heat decreases cost and energy
input. Using a solid storage medium for only one tank reduces the cost relative to
the two-tank systems. The thermocline system can be used at the solar power tower,
where steam and mineral oil are used as the HTF and storage fluid, respectively.
Heat pump storage. A true electricity storage system can be obtained via
thermal pumping (pumped heat electricity storage). This concept consists of using a
heat pump to store energy in thermal form in cheap solid materials with a high
expected efficiency. During the load phase, electricity is used to compress the
operating fluid and store its heat in a high-temperature container. At discharge,
pressurized, high-temperature fluid is used to power a turbo-generator. The effi-
ciency per cycle is estimated at around 70% [12].
Hot fluid During discharging During charging
Stored
Temperature
energy
Stored
Th,cut-off
L energy
Th,cut-off
Thermocline
Unusable Unusable
(temperature
gradient) Tc,cut-off
Tc,cut-off
Cold Hot Cold Hot Time

Cold fluid
Figure 11.6 Thermocline storage tank and temperature profiles

11.4 TES performance

Three efficiencies can typically be used as a measure of TES performance: storage-
tank effectiveness, first-law thermal efficiency and second-law exergy conversion
efficiency [14].
Storage-tank effectiveness. It is also called ‘discharge efficiency’ and defined
as a ratio of the amount of heat discharged to the total thermal energy stored when
the tank is fully charged; that is, the system capability of providing complete dis-
charge or some residual is necessary. For example, in a two-tank storage system,
about 20% of the hot or cold tank storage fluids cannot be used because of the
pump need and requirement for submerging the pump in the liquid. In the case of
thermocline TES, the tank effectiveness is determined by the usable portion of the
stored fluid outside of the mixed-temperature thermocline region. Therefore, the
storage-tank effectiveness, heff , can be written as
Wdis
heff ¼ (11.16)
WTES
where Wdis is the energy able to be discharged and WTES is the total energy in TES.
So, the tank should be oversized to accommodate the unusable residual thermal
energy.
First-law efficiency, hTES,FL: It is defined as the ratio of the energy extracted
from the storage to the energy stored in it and can be written in a form as follows:
energy extracted from the storage mCp ðT Tc Þ
hTES;FL ¼ ¼ (11.17)
energy stored in the storage mCp ðTh Tc Þ
where mCp is the total heat capacity of the storage medium. T and Tc are the hot and
cold salt temperature, respectively, of the storage during discharging, Th is the
maximum temperature at the end of charging period.
Second-law efficiency, hTES,SL: It measures the system ability to generate
the same amount of power from TES as from the original energy used to charge the
system, W. This means that the second-law efficiency measures the energy quality
degradation caused by both charging/discharging temperatures and thermal losses.
Thus, exergy transfer through a system can be written as
DG ¼ DW Ta DS (11.18)
where DG is the storage fluid exergy change, DW is the change of the available
energy being used (charging/discharging), Ta is the ambient temperature as a
reference point (25 C or 298 K) and DS is the entropy change of the storage fluid.
The round-trip efficiency of TES based on (11.18) can be written as

DGd ðDW Ta DS Þd
hTES;SL ¼ ¼ (11.19)
DGc ðDW Ta DS Þc
where subscripts c and d denote TES charging and discharging, respectively. More
explanation about losses estimation in TES systems can be found in [14].
References
[1] Nadeem F., Hussain S. M., Tiwari P. K., Goswami A. K. and Ustun T. S.
‘Comparative review of energy storage systems, their roles and impacts on
future power systems’. IEEE Access. 2018;7:4555–85.
[2] Moseley P. T. and Garche J. Electrochemical energy storage for renewable
sources and grid balancing. Amsterdam: Elsevier Science; 2015. p. 493.
[3] Wei H., Xin W., Jianhua Z. and Jingyan Y. ‘Discussion on application of
super capacitor energy storage system in microgrids’. Sustainable Power
Generation and Supply, 2009, International Conference on; Nanjing, China,
April 2009, pp. 1–4.
[4] Kim H. J., Seongm K. C., Cho J. W., et al. ‘3 MJ/750 kVA SMES system for
improving power quality’. IEEE Transactions on Applied Superconductivity.
2006;16(2):574–7.
[5] Nakayama T., Yagai T., Tsuda M. and Hamajima T. ‘Micro power grid
system with SMES and superconducting cable modules cooled by liquid
hydrogen’. IEEE Transactions on Applied Superconductivity. 2009;19(3):
2062–5.
[6] Said S. A., Aly M. M. and Abdel-Akher M. ‘Application of superconducting
magnetic energy storage (SMES) for voltage sag/swell suppression in dis-
tribution system with wind power penetration’. Harmonics and Quality of
Power (ICHQP), 2014 IEEE 16th International Conference on; Bucharest,
May 2014, pp. 92–6.
[7] Irie F., Okada H., Ezaki T., et al. ‘A field experiment on power line stabi-
lization by a SMES system’. IEEE Transactions on Magnetism. 1992;28(1):
426–9.
[8] Li Z. and Weizin M. ‘Application of SMES system with a comprehensive
control for enhancing transient stability’. Proceedings of the 5th
International Conference on Advances in Power System Control, Operation
and Management, APSCOM 2000; Hong Kong, Oct 2000, pp. 225–9.
[9] Huang X., Zhang G. and Xiao L. ‘Optimal location of SMES for improving
power system voltage stability’. IEEE Transactions on Applied
Superconductivity. 2010;3:1316–19.
[10] Robyns B., Francois B., Delille G. and Saudemont C. Energy storage in
electric power grids. London, UK: ISTE Ltd. and John Wiley & Sons; 2015.
[11] Sorensen B. Renewable energy conversion, transmission and storage. 3rd edn.
Amsterdam: Elsevier, AP; 2007, pp. 234–60.
[12] Sarbu I. and Sebarchievici C. ‘A comprehensive review of thermal energy
storage’. MDPI Journals, Sustainability. 2018;10(1):Article number 191.
[13] Barnes F. S. and Levine J. G. Large energy storage systems. Boca Raton, FL:
CRC Press; 2011.
[14] Ma Z., Glatzmaierm G. C., Wagner M. and Neises T. ‘General performance

metrics and applications to evaluate various thermal energy storage tech-
nologies’. Proceedings of ASME 2012 6th International Conference on
Energy Sustainability collocated with 9th International Conference on Fuel
Cell Science, Engineering and Technology; San Diego, CA, USA, July 2012,
pp. 345–51.
Chapter 12
Energy storage system application:
electric vehicles
12.1 Introduction
These days, the trend towards developing electric vehicles technologies, aiming to
enlarge the scale of electric vehicles use, is growing rapidly. The catalyst that
makes the engineers and researchers motivated to work on this trend is the great
concern about the environment, particularly noise and exhaust emissions, in addi-
tion to the continuous progress in batteries and fuel cells manufacturing and tech-
nology. It is, therefore, essential to understand the principles behind the design of
electric vehicles as well as the relevant technological and environmental issues.
Of course, internal combustion (IC) engine is the most popular and pre-
dominant type used for vehicles at present. However, electric vehicles are looking
as a strong contender for transport in the future. IC engine vehicles have greater
success as the specific energy of the petroleum fuel is about 300 times its value for
lead acid battery (typically 9 kWh/kg for petroleum and 0.03 kWh/kg for battery).
Even with accounting for the efficiency, the difference is still large. For instance,
the efficiency of IC engines, gearbox and transmission is about 20%. It means that
a useful energy of 1.8 kWh/kg at gearbox shaft can be obtained. In contrast, effi-
ciency of electric motors is around 90%, which only results in a useful energy of
0.027 kWh/kg. Therefore, to reach the value of useful energy of IC engines using
battery, the mass of battery should be about 66.7 times that of IC engine [1].
Moreover, the charging time of lead acid battery, as an example, is con-
siderably longer than it takes to fill a tank with fossil fuel, and its lifetime is limited
(~5 years). This requires further investment periodically to replace the batteries.
Thus, battery electric vehicles (BEVs) are more expensive compared to corre-
sponding IC engine vehicles.
As their range is limited, the application of BEVs is, to some extent, limited.
They are used as mobility devices for the elderly and physically handicapped, city
cars, trolleybus or trams and electric bicycles.
Recently, there have been technical developments in electric vehicles design
and improvements to motors, controllers and rechargeable batteries with reduced
charging time. Another type of electric vehicle is known as a ‘hybrid electric
vehicle’. It is likely to become the most common type in the near future [2]. It
combines an IC engine with a battery, an electric motor and generator. Generally,
these two types of electric vehicles (battery electric and hybrid electric) are toge-
ther referred to as plug-in electric vehicles (PEVs) and denoted as PBEVs and
PHEVs, respectively. In addition to PEVs, there are other types with different
features, depending on the primary drive, such as photovoltaic electric vehicles
(PVEVs) and fuel cell electric vehicles (FCEVs). The photovoltaic (PV) and fuel
cells are used rather than rechargeable batteries in these vehicles.
In addition, electric utility systems are proposing to incorporate PEVs into the
power system at the distribution level to help manage the load. PEVs can be dealt with
as generation source in the parking-discharging state or as loads in charging state.
12.2 PEV technologies

Batteries are the primary power source that supply PEVs with electric energy. They
help reduce the need for fossil fuels, but on the other hand, they need to be plugged
in to recharge.
A plug-in battery electric vehicle (PBEV) is powered only by an on-board
battery that supplies an electric motor with electric energy through a charge con-
troller electronic device. The electric motor provides the mechanical power
required for the car wheel through a transmission gear (Figure 12.1). Initially, it is
expected for go up to 160 km (100 miles) on a full charge and then the batteries are
required to be recharged by plugging into a power source.
A plug-in hybrid electric vehicle (PHEV) is equipped with both an electric
motor and an IC engine to power the car. An on-board computer is installed to
Plug-in rechargeable Charge controller Electric Transmission Wheel

battery Electronic device motor gear
(a)
Rechargeable Charge Electric Transmission

battery controller motor system
Wheel
(b)
Figure 12.1 PBEV drive train configuration: (a) typical representation and
(b) the corresponding block diagram
Energy storage system application: electric vehicles 279
control the electric and IC engine drives to give the car the power to cope effi-
ciently with the overall distance travelled. The electric motor is supported by an on-
board plug-in rechargeable battery. The battery also uses the regenerated energy
from the breaking and deceleration to recycle it through the battery to power the
car. The car reverts to the standard hybrid operation in case of battery depletion. In
PHEVs, a useful variable called the ‘degree of hybrid station’ (DOH) is used. It is
defined as
Power of electric motor
DOH ¼ (12.1)
Power of electric motor þ Power of IC engine
The car uses a smaller IC engine operating close to its optimum efficiency for a
greater proportion of the time when the DOH is high.
Three configurations for PHEVs exist: series-hybrid, parallel-hybrid and
combined series/parallel hybrid [3].
● Series PHEV: Source of electrical energy is either the battery or a generator
driven by an IC engine. The electrical energy is supplied to the electric motor,
which is the only power provider to the transmission system (Figure 12.2(a)
and (b)). The main disadvantage of this configuration is that all the electrical
energy must pass through both the generator and the motor, resulting in a
considerable increase in cost.
● Parallel PHEV: Both the electric motor and the IC engine provide power in
parallel to the transmission system (Figure 12.3(a) and (b)).
● Combined series/parallel PHEV: This type allows running the car using the
electric motor only or both the IC engine and electric motor together depend-
ing on the driving conditions that help decide which one of the two options is
in an optimal way. It combines the features of series and parallel hybrid. A
power split is used to share the output mechanical power from the IC engine to
the drive shaft or to the charge controller, as shown by the block diagram in
Figure 12.4.
In all configurations, the electric motor works as a generator when used as
regenerative brake.
12.2.1 Operation modes

The energy mode to drive PEVs can be determined by what is called state-of-charge
index, which is the percentage of the remaining energy capacity in the battery. PEVs
have the capability of driving in basically two different modes. First, charge
depleting operation mode, which is suitable for urban driving. The vehicle, like all
PBEVs, is powered only by the energy stored in the battery. Second, charge sus-
taining mode, where the battery is recharged over a driving profile to maintain the
state-of-charge level. PEVs use the battery pack to a large extent, but they need the
IC engine to supplement the power and getting better performance as in PHEVs.
Therefore, the selection criterion of the operation mode is mainly based on
(i) the availability of charging facilities from the grid and (ii) the driving distance.
Charge controller
Electronic device
Electric motor
Transmission
Plug-in rechargeable gear Wheel
battery Electric
generator
Fuel
system
IC engine
(a)
Fuel Charge Rechargeable Electric Transmission

ICE
tank controller battery motor system
Wheel
(b)
Figure 12.2 Series PHEV drive train configuration: (a) typical representation
and (b) the corresponding block diagram
12.2.2 PEV design features

As shown in Figures 12.1–12.3, all types of PEVs use electrical energy as a primary
source. They differ from IC engine vehicles, where the fossil fuel is burned by the
IC engine to power the vehicle. The most important design features in a PEV drive
train are the following.
Large battery packs to store the energy required to power the PEV. Different
types of battery packs can be used such as lead acid, NiMH and Li-ion, which is
becoming the most dominant pack in PEVs.
Electric motors to convert the electrical energy stored in battery packs to
mechanical energy that propels the vehicle forward or backward. The motors may
be AC or DC motors.
Electric generators are used to convert the mechanical energy into electrical
energy. For instance, the generator converts its input mechanical power coming
from the IC engine to the electrical energy being supplied to the electric motor as in
the series PHEV.
Inverters: In case of using AC motor, the motor is coupled with the battery
pack using an inverter, which converts the battery DC current to AC current.
Electric motor
Power from motor

or IC engine Transmission
Plug-in rechargeable Charge controller gear Wheel
battery Electronic device IC engine
Fuel
system
(a)
Fuel
ICE
tank
Transmission
Charge system
Rechargeable Electric
battery controller motor Wheel
(b)
Figure 12.3 Parallel PHEV drive train configuration: (a) typical representation
and (b) the corresponding block diagram
Fuel ICE Power

tank split
Transmission
system
Charge
controller Wheel
Rechargeable Electric
battery motor
Figure 12.4 Block diagram of combined series/parallel PHEV configuration
Chargers: A charger converts the AC supply into DC and charges the battery
pack. The AC supply may be the output of the generator or the utility grid. The
chargers are electronic devices, usually, fitted with a controller to optimize the
charging process.
12.3 Batteries
The battery is the main source of DC electric supplied to the electric motor in an
electric vehicle. The battery performance and cost are of crucial importance in the
development of BEVs. The essential performance parameters include power,
energy, lifetime, safety and cost.
Power: The desired peak power of battery generally remains constant and
depends on the operation mode, mainly the range on charge depleting mode as well
as energy management and the total weight of the vehicle.
Energy: An increase in the energy storage capacity of the battery results in a
longer distance for electric vehicle to travel, and meanwhile the mass of battery
pack may be getting larger. So, the decision of making the energy storage capacity
higher depends on the design features including specific energy (energy per kilo-
gram mass) and energy density (energy per volume) of the battery.
Lifetime: The lifetime of the battery is influenced by some factors such as
necessity of fast recharge, availability of recycling, desired vehicle mass, com-
pliance with safety standards, withstanding degradation over time, operation mode
and thermal management requirements.
Cost of batteries: There are several types of batteries. The cost, generally,
differs from one type to another. It includes the cost of: material, manufacturing,
transportation, marketing and development. However, high mass production of a
specific type of battery can reduce the cost. More details about batteries as energy
storage systems are discussed in Chapter 10.
12.4 Electric motors

The electric motor is one of the main components in the drive train of an electric
vehicle. It receives either AC or DC electric energy and converts it into mechanical
energy to be delivered to the transmission gear of the electric vehicle drive train.
AC induction motors are commonly used in case of supplying the motor with
AC electric energy. The battery produces DC energy, and thus an inverter is used
for DC/AC conversion to provide the input to the induction motor. More detail
about induction machines is provided in Chapter 2, Section 2.4.
DC electric motors, as an alternative to AC motors, are used as well. The battery
DC output energy is directly supplied to the DC motor through a controller. Different
types of DC electric motors can be used in electric vehicle application. The simplest
and classical type that can run with any electronic control is the ‘brushed’ DC motor.
Other advanced types such as brushless DC motors and switched reluctance (SR)
motors with sophisticated control are also used in traction applications.
12.4.1 Brushed DC motors

The brushed DC motor is of vital importance and very widely used in several
applications including domestic appliances, automated windows and doors, and
still used as traction motor.
Operating principles: The motor is equipped with permanent magnets and

brushes. North (N) and south (S) poles are arranged alternatively around the cir-
cumference of the stator and are fully employed (Figure 12.5). As a result of this
arrangement, it, of course, becomes necessary to employ a commutator to get the
possibility to group a desired number of conductors in series. The current passing
through the coils near the magnet causes a force to be generated in the coil.
To understand the principle of operation, it is necessary to first investigate the
constructional feature of the DC motor. As a very basic construction, one conductor
of a smoothed-cored armature is assumed, i.e. an armature in which the conductor
is not housed in slots, and in which it is, in consequence, acted on directly by the air
gap flux. The current carrying armature is connected to the supply end through
commutator segments and placed within N and S poles of a permanent or an
electromagnet (Figure 12.6(a)). Applying Fleming’s left-hand rule, the current
carrying conductor is placed in a magnetic field, B, which is perpendicular to the
direction of the current, I, and then the conductor experiences a force, F, in the
direction perpendicular to both the direction of field and current in the conductor. It
is to be noted that, in practice, several conductors in series constituting a coil lie in
slots as shown in Figure 12.6(b) to increase the torque. The main flux is shown
concentrated into tufts that pass into the tops of the teeth, whereas the armature flux
is shown by the dotted fluxes embracing the slots. Supposing that the excitation of
the field and the direction of the armature current are as depicted in this figure, the
magnetic force on the wires will be set up in an anticlockwise direction and under
its influence the armature will rotate. Also, more pair poles may be used to increase
the turning force [1,4,5].
The magnitude of the force, F, can be determined by considering the diagram
in Figure 12.7. It is known that when an infinitely small charge dq is made to flow
at a velocity, v, under the influence of an electric field, E, and a magnetic field, B,
then the Lorentz force dF experienced by the charge is given by
dF ¼ dqðE þ vBÞ (12.2)
Stator magnets Magnetic flux

Stator
Windings S N
Armature Brushes
Rotor
Current
Current going in
coming up N S
Commutator
Terminals
(a) (b)
Figure 12.5 Brushed DC motor: (a) physical structure of two-pole motor and (b) the
cross section of four-pole motor. The motor torque is anticlockwise
South pole North pole

Armature N
I I F Main field
B flux
F
S N
Brush I Brush
I
Armature flux
Commutator Commutator
segment segment
(a) (b)
Figure 12.6 (a) Basic construction of a DC motor. (b) Mechanism of

torque production
I
dL
Figure 12.7 Determination of force magnitude
Considering E ¼ 0 for the operation of DC motor and substituting v ¼ dL/dt,

(12.2) becomes
dL
dF ¼ dq B ¼ IdL B (12.3)
dt
where dL is the length of the conductor (m) carrying charge q and I ¼ dq/dt (A) and
B is the magnetic field strength (Wb/m2).
Integrating both sides of (12.3) and considering the magnetic field component
that is perpendicular to the direction of the current, B sin q, obtains
F ¼ BIL sin q (12.4)
It is seen from Figure 12.5(a) that the construction of a DC motor is such that
the direction of the current through the armature conductor at all instances is per-
pendicular to the field (i.e. q ¼ 90 ). Consequently, the force acts on the armature
in the direction perpendicular to both the uniform magnetic field and current.
Therefore, if the current passing through the right-hand side of the armature con-
ductor is I, the current at its left-hand side will be –I because they are flowing in the
opposite direction with respect to each other. Then, the force on the right-hand side
and that on the left-hand side are equal in magnitude and in the upward and
downward directions, respectively, and separated by armature turn width, w (m).
This produces a torque resulting in a rotation of the coil in anticlockwise direction.
As given by (12.4), the magnitude of the force on each side of the armature con-
ductor can be expressed as
F ¼ BIL sin 90 ¼ BIL (12.5)
Due to the rotation of the armature, the torque expression should be examined
when the armature turn is at an angle a with its initial position, that is, an angle
between the plane of armature turn and the plane of reference that is taken along the
magnetic field direction (Figure 12.8). The torque produced, T, is the multiplication
of the force tangential to the direction of armature rotation and the distance w.
Hence,
T ¼ F cos a w ¼ BILw cos a (12.6)
The torque is not the same at all positions as it varies with the variation of the
angle a. The arrangement shown in Figure 12.6(a), where the plane of armature
F
F cos(α)
I
B
I α
S N
F
I
I
Figure 12.8 Rotation of armature turn with respect to the initial position
turn coincides with the direction of magnetic field B, can be initially considered as
a reference position, that is, a ¼ 0, and then, the torque is at its maximum value
(T ¼ BILw) that helps in overcoming the initial momentum of the armature and sets
it into rotation. Because of the armature motion, the angle a is increasing in the
path of its rotation until it reaches 90 from the initial position. During this period,
the torque is decreased and given by (12.6). The instant at which a ¼ 90 ; the
torque is equal to zero, but the armature does not come to a standstill because of its
momentum of motion which is designed in such a way that the armature crosses
over this null torque point. After that, the angle a decreases and the torque is acting
again on the armature.
Supposing that the armature consists of n turns, and j represents the total flux
passing through the coil, i.e. j ¼ B area ¼ BLw, (12.6) can be rewritten as
T ¼ njI cos a (12.7)
It is to be noted that the torque does not only depend on the number of turns
and their relative position with respect to the magnetic field but also depends on the
number of pole pairs and other motor design aspects. So, the average torque can be
expressed in terms of the current and magnetic flux by the relation
Tav ¼ Km jI (12.8)
where Km is known as the motor constant and depends on the number of turns in
each coil, number of pole pairs and other aspects of motor design.
It is seen from (12.8) that the torque is directly proportional to the armature
current I that depends on the source voltage Vs and armature coil resistance Ra.
With turning the motor at velocity v, the armature moves in a magnetic field and
works as a generator producing a voltage called back electromotive force (BEMF),
Eb, which by considering the two sides of the coil is given by
Eb ¼ nBLww ¼ Km jw (12.9)
where w is the angular velocity (v ¼ ww/2) and Km is the same constant as that in
(12.8).
The BEMF acts to reduce the current in the motor as it opposes the source
voltage Vs. So, the armature current is produced due to the voltage difference
‘Vs – Eb’ and can be obtained by
Vs Eb Vs Km j
I¼ ¼ w
Ra Ra Ra
Substituting into (12.8) gets the average torque as
Km jVs ðKm jÞ2 1

T¼ w¼ KVs K2 w (12.10)
Ra Ra Ra
where K ¼ Km j.
Torque
Vs/Ra
θ= tan−1(− 2/R )
a
No-load
speed
Speed
Figure 12.9 Brushed DC motor torque–speed relation
It is noted that in (12.10) the torque needed to overcome the losses due to
friction in bearings and at the commutator and windage losses is ignored. However,
for this type of constant flux DC motor, the torque–speed relation is a linear rela-
tion and the torque decreases as the speed increases with a maximum at zero speed
(Figure 12.9).
However, it is seen that the torque is dependent on both the supply voltage, Vs,
and the magnetic flux, j. Two methods can be applied to provide the brushed DC
motor with magnetic flux, first, using permanent magnets as shown in Figures 12.5
and 12.6, and second, using electromagnets. For electromagnets, field windings
(wdgs) are placed in the stator, and by passing a current through these windings, the
magnetic field is created. So, electromagnets give the facility to easily change the
flux by varying the current and controlling the torque with great flexibility.
Meanwhile, the field windings consume electric current causing additional copper
losses. However, brushed DC electric motors with electromagnets are still com-
monly used as traction motors in electric vehicles application, whereas for smaller
motors permanent magnets may be used.
Field windings can be connected either in series, parallel (shunt) or separate
with motor armature windings to provide the required magnetic field
(Figure 12.10). The separate excited brushed DC motor is more advantageous than
the other two types when it is used in this application of electric vehicles. It allows
to independently control both the source voltage and the magnetic flux that can be
carried out by controlling the voltage across the field windings. So, obtaining the
required torque at any speed is ultimately facilitated.
Vf
Field wdgs
Field wdgs
Field wdgs Armature Armature Armature

wdgs wdgs wdgs
Vs Vs Vs
Series excited motor Shunt excited motor Separate excited motor
Figure 12.10 Types of excitation for DC motors
In addition, one of the main characteristics that should be determined is the

motor losses which in turn affect the efficiency of the motor. Brushed DC motors
have the same sources of losses as those of other electric motors. These losses are
mainly copper losses, iron losses (hysteresis and eddy currents) and mechanical
losses (friction and windage) as well as constant losses representing, for example,
the power needed for the electronic control circuits.
Copper losses, PcuLoss, result from the electrical resistance of the windings, R,
when the current, I, passes through them and causes heating. Copper losses can be
calculated by
PcuLoss ¼ I 2 R (12.11)
PcuLoss ¼ KcuLoss T 2 (12.12)

where KcuLoss has a value depending on R and the magnetic flux j.
Iron losses, PirnLoss, are produced due to exposing the rotor to changeable
magnetic field as the rotor rotates passing, frequently, N pole and then S pole. So,
the magnetic field effect on the rotor iron resulting in iron losses which are the sum
of what is called ‘hysteresis’ loss and ‘eddy currents’ loss. Hysteresis loss is the
energy required to magnetize and demagnetize the iron, whereas eddy currents are
generated in the iron by rotating the rotor under changeable magnetic field and
result in iron heating.
Therefore, iron losses are proportional to the frequency with which the mag-
netic field changes, and consequently they are proportional to the angular rotor
speed, w. Thus,
PirnLoss ¼ KirnLoss w (12.13)
where KirnLoss is a value affected by the magnetic field strength and depends on the
method of providing the magnetic field (with permanent magnets, it is nearly a
constant value).
Mechanical losses, PmecLoss, consist of (i) friction loss that results from
the presence of friction torque, Tfr, in bearings and brushes of the motor, it equals to
Tfr w, and (ii) windage loss, which is caused by rotor wind resistance (proportional
to w2 ), and so the power loss is proportional to the cube of angular speed; that is, it
can be expressed as KwLossw3 , where KwLoss is a constant depending on the size and
structural design of the motor. Thus,
PmecLoss ¼ Tfr w þ KwLoss w3 (12.14)

Constant losses, CLoss, are the power needed, for example, to supply the
electronic control circuits. These losses independently vary on both speed and
torque; that is, they occur even if the motor is stationary.
Therefore, the sum of losses from (12.12) to (12.14) in addition to the constant
losses gives the total losses, PtLoss as
PtLoss ¼ PcuLoss þ PirnLoss þ PmecLoss þ CLoss

¼ KcuLoss T 2 þ ðKirnLoss wÞ þ Tfr w þ KwLoss w3 þ CLoss (12.15)
By calculating the total losses as in (12.15) and considering the output power,
which is the torque T multiplied by the angular speed w (i.e. Tw), it is easy to
determine the motor efficiency, h, by
Tw
h¼ (12.16)
T w þ PtLoss
12.4.2 Dynamic and regenerative braking in DC motors

Dynamic braking means slowing down the motor using a shunt resistance by which
the negative torque produced when disconnecting the energy source can be con-
trolled. When a battery of negligible internal resistance is supplying its energy at
voltage Vs to a DC motor, the motor starts rotation until reaching its steady state at a
speed w providing a torque T given by (12.10) (Figure 12.11(a)). If the battery is
disconnected, the motor continues to rotate, because of its momentum, resulting in
a voltage to be generated with a value given by (12.9). Applying this voltage to a
resistance RL, as shown in Figure 12.11(b), causes a current I to flow out of the
motor and can be obtained by the following [1]:
Kw
I¼ (12.17)
Ra þ RL
The current I causes a negative torque that will slow down the motor. From
(12.10) and substituting Vs ¼ 0, as there is no battery, the value of the negative
torque is
K2 w
T ¼ (12.18)
R a þ RL
Vs Ra Ra
Vs
+ +
RL RL
M
–
–
(a) (b)
Figure 12.11 Dynamic braking of a DC motor. (a) DC motor at normal operation.

(b) DC motor at braking condition
From (12.18), at constant RL the negative torque falls rapidly as the speed
might decrease exponentially to zero. Also, the resistance RL can be changed
aiming to control the negative torque. It is noted that the kinetic energy of the motor
is totally converted into heat the same as normal friction brakes. So, it will be
worthy to exploit this energy in a useful manner such as storing the electrical
energy produced by the motor in a capacitor or battery.
Regenerative braking: A capacitor or a battery is used rather than a resistor to
store the braking energy generated by the motor. The problem in this case is that
connecting the battery with the motor will cause a current, Ibat, depending on the
difference between the motor voltage, Vm, and battery voltage, Vbat, and is given by
Vm Vbat Kw Vbat
Ibat ¼ ¼ (12.19)
Ra Ra
By examining (12.19), it is noted the following:
● Vm depends on a varying flux, j, and varying motor speed, w, and conse-
quently the braking torque which is proportional to the current Ibat will be
varying.
● The motor is slowing down (w is decreased) until reaching an instant at which
Vm ¼ Vbat resulting in a torque decline and no more braking effect unless the
battery voltage is very low.
● For low battery voltage, use of stored energy is difficult.
So, a voltage converter (DC/DC) can be used to solve this problem
(Figure 12.12). The input to the converter is the braking power, Pbrak, delivered by
the motor. It equals VmIm, where Vm and Im will change with speed and the desired
braking torque, respectively. The converter produces an output, Pconv, of the same
value of the input power, assuming its 100% efficiency, but at an increased voltage
and reduced current to be matched with the storing battery (or capacitor). In prac-
tice, the efficiency is not 100% as some of the braking electrical power will be lost.
DC/DC converter
Im Ibat
Ra
Vm Vbat
Figure 12.12 DC motor regenerative braking
Gears Energy
Charge Electric Wheels
to move
controller motor
Rechargeable vehicle
battery
Flow direction at normal forward driving
Flow direction at regenerative braking
Figure 12.13 Energy flows in a PBEV at normal and braking conditions
However, the function of the converter can be imagined as a transformer trans-

mitting a power at different turns ratios and a relation in terms of converter effi-
ciency, hconv , can be written as
Pconv ¼ hconv Pbrak
Hence,
Vbat Ibat ¼ hconv Vm Im (12.20)
The battery might be the same that provides the electrical energy to the motor.
So, the energy flows, for example, in a classical PBEV, are shown in Figure 12.13
at the normal forward driving and regenerative braking.
The power produced by the brushed DC motor can be increased by increasing

the supply voltage. This increase is limited not only by the insulation of copper
wires but also by the heat generated because of the losses above certain power and
occurs at the middle of the rotor. So, the heat generated is much more difficult to
remove, and then the heat removal is the main limiting factor of power increase.
For this reason, in addition to the problem of gradual wearing of both brushes and
commutator due to the friction between them, other types of DC motors may be
used. Brushless DC and SR motors are commonly used in electric vehicle appli-
cation where the great majority of the heat generated by losses is in the outer stator,
which in turn leads to the heat being removed easily, and thus allows smaller
motors to be used.
12.4.3 Brushless DC motors

Brushless DC motors (BLDC motors) are AC motors as the current passing through
the motor alternates. The name BLDC comes from the facts that (i) the AC current
must be variable frequency and so derived from a DC supply and (ii) the torque–
speed characteristic is very similar to that of a brushed DC motor [6–9].
12.4.3.1 Operational theory

The basic principle of BLDC motor operation can be explained by considering that
the rotor consists of a permanent magnet whereas the stator has a coil wound either
clockwise or anticlockwise along each stator arm. Switches are used to direct the
current from a DC source through the coil. In Figure 12.14(a), the current is flowing
in the direction of magnetizing the stator to turn the rotor clockwise. The rotor
passes between the stator poles until being alignment with stator magnetic axis and
the stator current is switched off (Figure 12.14(b)). Momentum carries the rotor on
and the stator coil is re-energized, but the current and its resulting magnetic field
are reversed as shown in Figure 12.14(c). Thus, the rotor is turning clockwise, and
the process continues with switching the current ON–OFF–ON (reversed). So, this
switching must be synchronized with the position of the rotor. To do that, the
BLDC motor accomplishes commutation electronically using rotor position feed-
back to determine when to switch the current. Feedback usually entails an attached
Hall sensor or a rotary encoder.
Stator Stator Stator
(a) (b) (c)
Figure 12.14 Basic principles of BLDC motor operation. The current is switched
(a) ON, (b) OFF and (c) ON reversed
Permanent
A1
magnet rotor Field windings
C2 B1
N
Rotor
Magnet
S
Field coils C1
B2
(electronically
controlled) A2
Rotary encoder
(a) (b)
Figure 12.15 Typical examples of BLDC motors; (a) single-phase motor and
(b) cross section of three-phase motor
In practice, the rotor may consist of a shaft and a hub with permanent magnets
arranged to form between two to eight pole pairs that alternate between N and S
poles. The stator has single-phase, two-phase or three-phase windings. Single- and
three-phase windings are the most widely used. A BLDC motor with single-phase
stator winding forming two pole pairs with electronically controlled field coils is
typically shown in Figure 12.15(a). A simplified cross section of three-phase stator
windings forming three pole pairs is depicted in Figure 12.15(b). A rotating mag-
netic field is created by the sequential excitation of the pole pairs by a DC pulse.
Two types of stator windings may be used, trapezoidal and sinusoidal, which
refers to the shape of BEMF signal. The shape of this signal is determined by the
coil interconnections and the air gap distance. The phase current also follows a
trapezoidal and sinusoidal shape.
Motor operation is based on the attraction or repulsion between magnetic
poles. Using the three-phase motor shown in Figure 12.16, the process starts when
current flows through one of the three stator windings and generates a magnetic
pole that attracts the closest permanent magnet of the opposite pole. The rotor will
move if the current shifts to an adjacent winding. Sequentially charging each
winding will cause the rotor to follow in a rotating field. The torque in this example
depends on the current amplitude and the number of turns on the stator windings,
the strength and the size of the permanent magnets, the air gap between the rotor
and the windings, and the length of the rotating arm.
Therefore, the stator windings work in conjunction with permanent magnets on
the rotor to generate a nearly uniform flux density in the air gap. This permits the
stator coils to be driven by a constant DC voltage (hence the name brushless DC),
which simply switches from one stator coil to the next to generate an AC voltage
waveform with a trapezoidal or sinusoidal shape.
12.4.3.2 Switching process

Electric switches are used in BLDC motors to realize current commutation, and
thus continuously rotate the motor. These switches are usually connected in an
H-bridge structure for a single-phase BLDC motor, and a three-phase bridge
Permanent magnet Stator

rotor windings
A1 A1 A1
B2 C2 B2 C2 B2 C2
C1 B1 C1 B1 C1 B1
A2 Rotating magnetic A2 A2
field
Figure 12.16 Basic principle of three-phase BLDC motor operation
Three-phase
BLDC motor
U
SW1 SW3 SW1 SW3 SW5
M U M
V W V W
SW2 SW4 SW2 SW4 SW6

H-bridge Three-phase bridge
Figure 12.17 Electric driver circuit
structure for a three-phase BLDC motor as shown in Figure 12.17. High-side

switches usually are controlled using pulse-width modulation (PWM), which con-
verts a DC voltage into a modulated voltage to easily and efficiently limit the start-
up current and control speed and torque. Generally, PWM losses are increased when
raising the switching frequency, and on the other hand, lowering the switching
frequency limits the system’s bandwidth and can raise the ripple current pulses to
the point where they become destructive or shut down the BLDC motor driver.
12.4.3.3 Equivalent circuit and general equations

The equivalent circuit in the dynamic state is shown in Figure 12.18(a) as a per-
phase diagram. lm is the flux linkage of stator winding per phase due to the per-
manent magnet. For steady-state conditions, both the voltage, v, and electromotive
force, e, are assumed sinusoidal at frequency, w. The equivalent circuit becomes as
in Figure 12.18(b), where X ¼ wL, and V, I, E and lm are phasors with rms
amplitudes. The steady-state circuit equation can be written as
V ¼ E þ ðR þ jwLÞI (12.21)
R L R X = ωL
i I
v e = dλm/dt V E = jωλm
(a) (b)
Figure 12.18 Equivalent circuits of BLDC motors: (a) dynamic equivalent circuit
and (b) steady-state equivalent circuit
For maximum mechanical power at a given speed, I and E are in phase. This
also gives maximum torque/ampere (minimum current/N m). A brushless DC
motor has position feedback from the rotor via Hall devices, optical devices,
encoder, etc., to keep an angle between V and E, as E is in phase with rotor position
and V is determined by the inverter supply to the motor. If wL R, when I is in
phase with E, V will also be in phase with E. Thus, the circuit can be analysed using
magnitudes of E, V and I as if it were a DC circuit.
It is noted that when E and I are in phase, the motor mechanical power output
(before friction, windage and iron losses), i.e. the electromagnetic output power is
Pem ¼ mjEjjIj ¼ mwjlm jjIj (12.22)
where m is the number of phases, |E|, |I| and |lm| are the amplitudes of phasor E, I
and lm, and the electromagnetic torque is
Pem mwjlm jjI j
Tem ¼ ¼ (12.23)
wr wr
where wr ¼ 2 w/p is the rotor speed in rad/s, and p the number of poles.
Hence,
mp
Tem ¼ jlm jjI j (12.24)
2
The load torque is the actual shaft output torque derived by subtracting the
losses in the torque resulting from friction, windage and iron losses. That is,
Tload ¼ Tem Tlosses (12.25)
Dropping the modulus signs in (12.24) gives
mp
Tem ¼ lm I (12.26)
2
The voltage, E, in terms of rotor speed can be written as

p
E ¼ wr lm (12.27)
2
Assuming V and E are kept in phase by the position feedback and wL R,

(12.21) can be simplified and rewritten in an algebraic form as
V ¼ E þ RI (12.28)
Incorporating (12.26), (12.27) and (12.28) gives

p 2R
V ¼ wr lm þ Tem
2 mplm
Thus,
mðplm =2Þ mðplm =2Þ2 m

Tem ¼ V wr ¼ KV K2 wr (12.29)
R R R
where K ¼ plm =2.

It is seen from (12.29) that the torque speed curve is very similar to that of
brushed DC motors given in (12.10).
The efficiency is calculated as the ratio of output power, Pout, and input power,
Pin. Hence,
Pout
h¼ (12.30)
Pin
where Pin ¼ mVI and Pout ¼ Tloadwr.

The input power can be obtained by adding the power losses to the output
power. The loss calculations are the same as those of brushed DC motors,
Section 12.4.1. A comparison of features between the brushed DC motor and
BLDC motor is given in Table 12.1.
12.4.4 SR motors
The basic operation of the SR motor is based on magnetizing the iron stator and
rotor by a current through the coil on the stator. As shown in Figure 12.19(a), the
rotor is out of line with the magnetic field, and so a torque will be produced to
reduce the reluctant in the path of the magnetic field by minimizing the air gap and
making the magnetic field symmetrical as well. When the rotor is aligned with the
stator, the switch will be OFF and no current is supplied (Figure 12.19(b)). Because
of the rotor momentum, the rotor continues turning to the position shown in
Figure 12.19(c), where the switch at that instant is in ON position to re-apply the
magnetic field in the same direction. Consequently, the field exerts a torque to
minimize the air gap and make the field symmetrical. The rotor is pulled forward
Table 12.1 Comparison between brushed DC motor and BLDC motor
Feature Brushed DC motor BLDC motor

Life Shorter Longer
Maintenance Periodically required Less required due to the absence of brushes
Commutation Brushed commutation Electronic commutation
Efficiency Moderate High (no voltage drop across brushes)
Output power Moderate as large heat High as the heat is produced in the
is produced in the middle outer stator and easy to remove
of rotor and not easy to
remove
Torque–speed Moderately flat as brush Flat and enables operation at all speeds
curve friction increases at higher with rated load
speeds
Speed range Lower because of mechanical Higher because of absence of
limitations by brushes mechanical limitations by brushes
and commutator
Rotor inertia Higher leading to limiting the Low as it has permanent magnets on the
dynamic characteristics rotor and the dynamic response is
improved
Electric noise Higher and results from the Low
arcs in the brushes
Control Simple and inexpensive Complex and expensive
Control necessity It is essential to keep the It is not needed for fixed speed. Only
motor running required if variable speed is desired
Stator Stator Stator
(a) (b) (c)
Figure 12.19 Basic principle of SR motors: (a) and (c) power supply is ON, stator
and rotor are magnetized and the magnetic field turns rotor to
minimize air gap; (b) power supply is OFF and momentum keeps
rotor turning
until it is aligned with the stator where the current is switched off and so on. Thus,
the SR motor is quite simple and has the following advantages [10,11]:
● The rotor is a piece of magnetically soft iron.
● No need to alternate the current in the coil. So, the cost is potentially reduced.
● Speed control can be carried out by changing the time interval between two
successive switching (ON and OFF). The controller should be designed care-
fully and based on accurate information about the rotor position that is usually
provided by sensors or recently inferred from voltage and current patterns. A
simplified schematic diagram illustrating the feedback signals from both cur-
rent detector and position detector to the controller is shown in Figure 12.20.
● No BEMF is generated as in case of BLDC motor because the rotor of SR
motor is not a permanent magnet, i.e. enabling higher speeds.
● Higher power density is possible particularly when using a core of high mag-
netic permeability that leads to higher torque produced within a given volume.
The torque produced by the magnetic field can be made much smoother by
multiplying the stator coils as well as using a rotor made of laminated iron and has
salient poles (Figure 12.21(a)). In this figure, the SR motor consists of four salient
pole rotor and six stator coils. Coil A is magnetized and exerting a clockwise force
on the rotor. When the salient poles are aligned with coil A, the current in A is
switched off. At this instant, two other salient poles are nearby the line of coil B,
which is energized keeping the rotor smoothly rotating. The number of salient poles
will mostly be two less than the number of coils. Clearly, correct current switching
ON and OFF is needed. A typical physical structure of an SR motor with eight
salient pole rotor and ten stator coils is shown in Figure 12.21(b).
Motor
Power Converter
supply SR
DC/DC
Current
detector
Controller
Position
detector
Figure 12.20 Simplified block diagram of SR motor controller

A1
C2
B2
Rotor
C1 B1
A2
(a) (b)
Figure 12.21 SR motor structure: (a) cross section of four salient pole rotor and
six stator windings and (b) typical physical structure of eight salient
pole rotor and ten stator coils
12.5 Electric vehicles performance
Derivation of equations that express the performance of electric vehicles necessi-

tates the determination of the forces acting on the vehicle during its motion [12,13].
As shown in Figure 12.22, a vehicle of mass m is moving at speed v to climb a hill
with a slope of angle Q, as a general case. Obviously, if the vehicle is moving on a
horizontal plane the angle will be zero. The vehicle is exposed to forces as follows:
● A force propelling the vehicle drive train forward, Ftrain.
● A force in opposite to the direction of movement due to the friction of tire on
the road, friction in bearings and the gearing system. This force is called
‘rolling resistance force’, Froll, and is proportional to vehicle weight, Thus,
Froll ¼ mroll mg (12.31)
where g is the gravitational acceleration and mroll is the rolling resistance
coefficient that depends on the type of tire and the tire pressure, with typical
values of 0.015 for radial ply tire or about 0.005 for tires especially developed
for electric vehicles.
● The drag force, Fdrg, due to the friction of the vehicle body moving through the
air. It acts in a direction opposite to the direction of velocity and it is a function
of frontal area, body shape, duct and air passages, spoilers, etc. The formula
expressing the value of this force is
Fdrg ¼ ½rACdrg v2 (12.32)

where Fdrg is the force in Newtons, r is the air density of reasonable value of
1.25 kg/m3, A is the frontal area in m2, Cdrg is the drag coefficient of a typical
value around 0.19 depending on the vehicle design and v is the velocity (m/s).
Fhill
Fdrg
Ftrain
θ
Froll
mg θ
Figure 12.22 Forces affecting the vehicle motion
● Hill climbing force, Fhill, which is the force needed to overcome the compo-
nent of vehicle’s weight acting down the slope. It is given by
Fhill ¼ mg sin q (12.33)
and has a negative sign if the vehicle is going downhill.
● A force to accelerate the vehicle is needed and is added to the other forces
because the velocity of the vehicle is changing. The acceleration is not only a
linear acceleration a, provided by a force, Flin-acc, that is given by
Flin-acc ¼ ma (12.34)
but also an angular acceleration that needs a force, Fang-acc, to turn the rotating
parts faster. It can be derived as given below.
● The axle torque ¼ Ftrainr, where r is the radius of the tire. As the motor torque
is transmitted to the tire through a transmission gear with gear ratio ngear, the
motor torque T can be obtained by
Ftrain r
T¼
ngear
that is,
ngear
Ftrain ¼ T (12.35)
r
The axle angular speed ¼ (ngear/r) rad/s and accordingly the motor angular
speed w and its angular acceleration w_ are given by
v a
w ¼ ngear and w_ ¼ ngear (12.36)
r r
Therefore, this angular acceleration needs to provide a torque of a value as

a
T¼ Ingear (12.37)
r
where I is the moment of inertia of the motor rotor.
Incorporating (12.35) and (12.37) gives the corresponding force, Fang-acc, at the
wheels, as
n 2
gear
Fang-acc ¼ Ia (12.38)
r
T accounts for the gear system efficiency, hgear, and (12.38) should be written in a
more accurate form as
1 ngear 2
Fang-acc ¼ Ia (12.39)
hgear r
It is to be noted that both Flin-acc and Fang-acc will be negative if the vehicle is
slowing down.
Therefore, the force provided the vehicle drive train must be the sum of all
other forces, that is,
Ftrain ¼ Froll þ Fdrg þ Fhill þ Flin-acc þ Fang-acc (12.40)
Substituting the various terms given in (12.31) to (12.35) and (12.39) in (12.40) gives
" ! #
ngear I n 2 dv
2 gear
T ¼ mroll mg þ 0:625ACdrg v þ mg sin q þ m þ
r hgear r dt
(12.41)
where a ¼ dv/dt.
Equation (12.41) can be solved as a first-order differential equation to obtain
the value of vehicle acceleration which is a key performance indicator.
12.6 Battery charging infrastructure
Range anxiety problems associated with electric vehicles have been obviated by
introducing PHEVs and development of higher density batteries capable of storing
more energy in the same space. These days charging capacity is of more concern
than the range anxiety as it helps in increasing the popularity of electric vehicles.
So, suitable chargers and the charging infrastructure necessary to support the
growth in electric vehicles population must be provided. The current-carrying
capacity of chargers and charging time must be suitable for the application.
However, high charging rate may accelerate battery ageing, which is the reason for
the manufacturers to recommend the charging time being between 2 and 3 h. In
addition, the charger can be placed on-board of the vehicle (Figure 12.23), or off-
board, for example, charging stations [14].
Fuel Clutch
ICE
Wheels
tank
Coupler
Inverter and
Battery Electric
switching
charger motor Clutch Transmission
Battery device
gear
Figure 12.23 On-board battery charger for PHEV
Chargers provide a DC charging voltage from an AC source whether from a

common socket outlet or more recently from a purpose-built DC charging station.
Most important are the methods of controlling the charge and protecting the battery
from over-voltage, over-current and over-temperature. These charger functions are
integrated with and unique to the battery.
Chargers for electric bikes are usually low-cost, separate units. To save weight,
they are not usually mounted on the bike and charging takes place at home. Their
power handling capacity is only enough for charging the relatively low-power bike
batteries and entirely unsuitable for passenger car applications.
Chargers for passenger cars are normally mounted inside the car. This is
because the vehicle may be used a long way from home, further than the range
possible from a single-battery charge. For this reason, they must carry the charger
with them on board the vehicle. Charging can be carried out at home from a stan-
dard domestic electricity socket outlet, but the available power is very low, and
charging takes a long time, possibly 10 h or more depending on the size of the
battery. As charging is usually carried out overnight this is not necessarily a pro-
blem, but it could be if the car is away from its home base. Such low-power
charging is normally used in an emergency and most cars are fitted with a higher-
power charging option which can be used in commercial locations or with a higher-
power domestic installation. In many countries, this higher-power facility is
implemented by means of a three-phase electricity supply.
Commercial electric vehicles need bigger batteries which need higher power
charging stations to achieve reasonable charging times, but they also have extra
options. Many of them follow prescribed delivery routes within a limited range
from base and return to base in the evening. In these cases, off-board charging is
possible saving weight and space on the vehicle. Such applications can also be
adapted to battery swap options. Each vehicle may have two batteries with one
being charged while the other is in use. When used in long-distance shuttle appli-
cations, this can double the effective range of the vehicle. The vehicle depletes the
battery during each journey and picks up a fully charged battery at the terminus
leaving the discharged battery to be recharged ready for the next trip. This shuttle
option, however, needs three batteries per vehicle.
12.6.1 Charging stations

Charging stations merely deliver the energy to the vehicle, usually in the form of a
high-voltage AC or DC supply. They do not normally have the functions of the
charger which must transform the electrical energy into a form that can be applied
directly to the battery.
From the wide range of potential consumers noted above, the electric vehicle
(EV) community needs several power supply options. Three different power levels
have been outlined.
Level 1 refers to single phase AC using grounded receptacles as used in
domestic applications (Figure 12.24). In North America, this typically means 16 A
at 120 V delivering 1.9 kW of power. In Europe, it may be 13 or 16 A at 240 V
delivering 3 kW of power. The EV may incorporate a standard domestic power
cord to connect the vehicle to a domestic socket outlet or a Level 1 charging station.
Level 2 delivers up to 20 kW of power from either single- or three-phase AC
source of 208–240 V at up to 80 A. The connector and charging cable used in
Level 2 applications must follow the standards. These cables are permanently fixed
to Level 2 charging station rather than the vehicle with the male connector being
mounted in the vehicle itself. The connector is also commonly called a ‘coupler’.
Individuals can install a Level 2 charging station at home, whereas businesses and
Solar power
Grid
powe
r
Distribution
board
Charge
EV
Discharge
Figure 12.24 Domestic charge–discharge EV

local government can also install Level 2 charging stations in kerbside, car parking
and so on (Figure 12.25).
Level 3 refers to DC charging, or ‘fast charging’. To achieve very short char-
ging times, Level 3 chargers supply very high currents of up to 400 A at voltages
up to 600 V DC delivering a maximum power of 240 kW. Hybrid coupler allows
charging from either a conventional 15-A AC wall outlet or a DC connection of up
to 90 kW. These systems are being proposed for public fast charging stations
(Figure 12.26).
12.6.2 Chargers
The diagram shown in Figure 12.27 illustrates how the key functions of Level 1, 2
and 3 chargers are implemented. All charging systems take AC power from the grid
and convert it to DC power at a suitable voltage for charging the battery. In EV
applications, except for bicycles, Level 1 and Level 2 chargers are completely
contained within the vehicle. In Level 3 charging systems, however, the charging
functions are split between the charging station and the vehicle’s on-board charger.
12.6.2.1 Level 1 and Level 2 chargers

In low-power, Level 1 and Level 2, applications, the power conditioning which
includes the AC to DC conversion, the power control unit which delivers a variable
DC voltage to the battery, and various filtering functions are all carried out within
the charger and can be implemented at a relatively low cost. The battery manage-
ment system (BMS) is tightly integrated with the battery. It monitors the key bat-
tery operating parameters of voltage, current and temperature and controls the
charging rate to provide the required constant current/constant voltage (CC/CV)
charging profile and it triggers the protection circuits if the battery’s operating
limits are exceeded, isolating the battery if needed. The charger will also most
likely incorporate controller area network (CAN) bus functionality to integrate with
other vehicle systems but not necessarily with the charging station.
(a) (b)
Figure 12.25 Charging station Level 2: (a) charging station in car park and
(b) charging station in kerbside
Figure 12.26 Fast charging station
Level 1 and Level 2 AC chargers

Public charging
Monitoring
Charging
Controls BMS
station
Safety Power control
AC interlocks DC Variable DC Protection CC&CV
unit
supply Fixed Rectifier DC/DC circuits Charging
AC supply AC converter
Battery
Metering and Level 3 DC charger

billing
Monitoring
Charging
CAN bus control Controls BMS
station
Safety
AC interlocks Variable DC controlled current and Protection CC&CV
supply Variable circuits
Charging
Level 3 DC supply controlled voltage
Battery
Figure 12.27 Main elements of the charger
Safety measures in the relatively low-power Level 1 charging station are

simple and may be limited to a ground fault sensing device and a circuit inter-
rupting device or ‘circuit breaker’. However, the charger itself will usually incor-
porate more comprehensive safety measures in addition to the standard BMS
functions including safety interlocks and isolators to prevent power being con-
nected if there is a fault in the battery or the charger, as well as measures to prevent
misuse, electric shocks and inadvertently driving away with the power cord still
plugged in.
Level 1 charging works from a single-phase AC power outlet and is suitable for
private, domestic installations and these do not need authentication and billing.
If the charger is designed to work with public charging stations, as many
Level 2 installations will be, it will most likely need to incorporate further intelli-
gence to communicate with the charging station to verify that the user is authorized
to draw power from that particular source and to allow it to bill the customer for the
energy transferred unless charging is installed at home or as a free service in the
workplace or shopping mall.
Level 2 charging stations may use either single-phase or three-phase AC power
from the grid.
This self-contained system used for Levels 1 and 2 gives the charger the
flexibility to connect to different AC charging sources.
12.6.2.2 Level 3 chargers

Level 3 chargers have the same functionality as the Level 1 and 2 chargers but with
the very high-power levels used, the AC/DC conversion and the power conditioning
and control circuits become very large and very expensive requiring heavy-duty
components. It makes sense to carry out these functions in the charging station rather
than the charger so that the equipment can be shared by many users. This allows major
cost and weight savings in the vehicle’s on-board charger and with a bigger budget for
the charging station, more efficient designs may be possible. In this case, the BMS
must communicate with the charging station to control the voltage and current
delivered to the battery since power control is not implemented within the battery. The
charging station, however, does not control how the batteries on the vehicle are
charged. That is the function of the charger on the vehicle itself, and its BMS. The
charger provides this control by means of the CAN bus which communicates its
demands to the charging station. Safety functions for isolating the battery and pro-
tecting the vehicle’s occupants are also managed by the charger and the BMS.
While putting some of the traditional charger functions into the charging sta-
tion permits cost savings in the vehicle, the charging station will be correspond-
ingly more expensive. On top of this, there is the cost of providing access to the
grid. Because of the high-power requirements of up to 240 kW per station, they
cannot just be connected to the grid anywhere. The electricity generating utility
must provide a dedicated supply line capable of delivering the high currents
demanded. One drawback of this system is that the vehicle can only be charged at
compatible public charging stations.
12.6.3 Charging times

Charging rates for very large batteries may be limited by the capacity of the power
supply resulting in excessive charging times, and quick chargers have been
developed to address this problem. However, just because a fast charging station is
available it does not mean that the charging time can be reduced by connecting to
it. First, fast charging systems use high-power DC charging systems and, unless the
vehicle has a compatible charger on-board, the charging station will be of no use to
the user. But more importantly, no matter what voltage or current the charger can
deliver, the charging time is ultimately controlled or limited by the battery, and the
cells from which it is made and how much current they can accept. Try powering a
light bulb with a higher voltage than it was designed for and see how quickly it
expires.
Renewable energy sources, such as PV solar panels, wind turbines and fuel
cells, can be used to charge the vehicle’s battery in addition to the power supplied
by utility grid. The controllers of each source can be connected through a wireless
network for system management (Figure 12.28) [15]. PV solar arrays can be also
PV solar arrays
Utility grid
MPPT-PV
Energy
Load shift
storage
system
system
MPPT-WT
An EV
charging slot
WTs
Figure 12.28 Charging station supplied from utility grid, PV panels and
wind turbines equipped with maximum power point tracking
(MPPT) system
(a) (b)
Figure 12.29 Installation of PV arrays: (a) on-board PV panels and (b) PV arrays
on the roof of a station
installed on-board of the vehicle or on the roof of the charging station or homes
(Figure 12.29(a) and (b)).
References
[1] Larminie J. and Lowry J. Electric vehicle technology explained. 2nd edn.
West Sussex, England: John Wiley & Sons Ltd.; 2012.
[2] Singer M. Consumer views on plug-in electric vehicles – National bench-
mark report. NREL/TP, Technical Report, 2016.
[3] Chakraaborty S., Simoes M. G. and Kramer W. E. (eds.). Power electronics
for renewable and distributed energy sources. London: Springer; 2013.
[4] Clayton A. E. and Hancock N. N. The performance and design of direct
current machines. 3rd edn. London: Sir Isaac Pitman & Sons; 1961.
[5] Cotton H. Electrical technology. 4th edn. London: Sir Isaac Pitman & Sons;
1961.
[6] Brown W. Brushless DC motor control made easy. Chicago, USA:
Microship Technology Inc.; 2002.
[7] Yedamale P. Brushless DC (BLDC) motor fundamentals. Chicago, USA:
Microship Technology Inc.; 2003.
[8] Zhao J. and Yu Y. Brushless DC motor fundamentals application note. MPS/
AN047 Rev. 1.0 Report, 2014.
[9] Singh B. and Singh S. ‘State of the art on permanent magnet brushless DC
motor drives’. Journal of Power Electronics. 2009;9(1):1–17.
[10] DiRenzo M. T. Switched reluctance motor control-basic operation and
example using the TMS320F240. Application Report, Texas Instruments
Inc., 2000.
[11] Ahn J. W. Switched reluctance motor. Rijeka, Croatia, IntechOpen; 2011,
pp. 201–52.
[12] Abdel-Kader F. E. S., Elsherif M. Z., Abdel-Rahim N. M. B. and Fathy M. M.
‘Control methods of the switched reluctance motor in electric vehicle during
acceleration’. Leonardo Electronic Journal of Practices and Technologies.
2012;20:127–46.
[13] Abdel-Kader F. E. S., Elsherif M. Z., Abdel-Rahim N. M. B. and Fathy M. M.

‘Performance characteristics of the switched reluctance motor in electric
vehicle during acceleration at variable turn on and turn off angle’. International
Journal of Science and Emerging Technologies 2012;4(2):143–53.
[14] Rajakaruna S., Shahnia F. and Ghosh A. (eds.). Plug in electric vehicles in
smart grids: Integration techniques. London: Springer; 2015.
[15] Yan Y., Qian Y., Sharif H. and Tipper D. ‘A survey on smart grid commu-
nication infrastructures: Motivations, requirements and challenges’. IEEE
Communications Surveys & Tutorials. 2012;15(1):5–20.
Part III
Renewables integration
Chapter 13
Power conditioning systems
13.1 General
Power conditioning systems (PCSs) are power electronics devices/circuits that act as
electrical interface between the utility power grid or demand and renewable sources
or energy storage systems [1]. A PCS is a dedicated device for power processing to
output a voltage or current in a form adequate for the end user. The processing is
based on one or a combination of some of the following functions [2–8]:
● AC–DC conversion (e.g. rectifiers and converters),
● DC–DC conversion for changing the voltage level (e.g. DC–DC converter),
● DC–AC conversion (e.g. inverters),
● AC–AC conversion (e.g. matrix inverter),
● elimination of voltage wave ripples by filters and
● monitoring and control.
For instance, a PCS for interfacing an energy storage system and utility power
grid has three functions: (i) AC–DC conversion for charging the storage device,
(ii) DC–AC conversion during storage discharge, and (iii) controlling the power
during conversion process to avoid any damage occurring to the storage facility or
end user’s electrical system [9]. To perform its function, a PCS has a controller
integrated with a power electronic device (Figure 13.1).
Each type of storage technology has its own operation methodology during
charging/discharging process. Accordingly, a PCS must be properly designed for
each type to keep its desired lifespan.
13.2 AC–DC conversion

Rectification is the process of conversion from AC to DC, and the conversion
device is called ‘rectifier’. For an ideal rectifier, the output voltage should be a pure
DC signal without any ripple, and the current in phase with the voltage with no
harmonics. Depending on the power electronic device adopted for rectification,
rectifiers can be divided into categories, such as (i) uncontrolled rectifiers
with diodes, (ii) phase-controlled rectifiers with thyristors and (iii) pulse-width
modulation (PWM)-controlled rectifiers with insulated gate bipolar transistors
(IGBTs) or metal oxide semiconductor field-effect transistors (MOSFETs).
Power processing
PCS
Power in Power out

Power
electronic
devices
Controller
Figure 13.1 Power processing for interfacing

Output voltage
Input voltage
D io
+
vs R Vo
−
0 2 3 4 0 2 3 4
ωt (rad) ωt (rad)
(a) (b) (c)
Figure 13.2 Uncontrolled rectifier with diode: (a) single-phase bridge, (b) input
voltage waveform and (c) output voltage waveform
13.2.1 Uncontrolled rectifiers

The basic form of rectifiers comprises a diode (Figure 13.2(a)). The diode allows
only the positive half cycle of the input sinusoidal AC voltage, Figure 13.2(b), to
pass through to provide a DC output voltage in a form shown in Figure 13.2(c). The
output voltage encompasses a significant number of ripples and, consequently, the
source current waveform is not a pure sinusoid due to the presence of harmonic
components. To improve the waveform of both output voltage and source current, a
single-phase bridge rectifier is employed by connecting four diodes in a bridge
form (Figure 13.3). Compared to the half-wave rectifier with one diode, both half
cycles of the AC source voltage are passed to form the DC output voltage that is
applied to the load.
Power conditioning systems 315
io
Output voltage
Input voltage
is D1 D3
+
vs Load Vo
D4 D2
−
0 2 3 4 0 2 3 4
ωt (rad) ωt (rad)
(a) (b) (c)
Figure 13.3 Uncontrolled full-wave rectifier: (a) full-wave bridge, (b) input
voltage waveform and (c) output voltage waveform
io Vab Vac Vbc Vba Vca Vcb Vo
Output voltage
Input voltage
Va Vb Vc
va ia
D1 D3 D5
vb ib +
vc ic Load Vo
D4 D6 D2
−
0 2 3 4 0 2 3 4
ωt (rad) ωt (rad)
(a) (b) (c)
Figure 13.4 Uncontrolled three-phase full-wave rectifier: (a) six-diode bridge,

(b) three-phase input voltage and (c) output voltage waveform
This symmetry in the two halves of the cycle reduces both the ripples in the
output voltage and harmonic components in the input current. The DC output
voltage, Vo, of this full-wave rectifier is given by
ð
1 p pffiffiffi
Vo ¼ 2Vs sin wt dwt (13.1)
p 0
where Vs is the RMS value of the source voltage.
For three-phase supply applications, the full-wave bridge converter comprises
six diodes (Figure 13.4(a)). The pair of diodes with the highest applied instanta-
neous line voltage conducts periodically in the order D1; D2 ! D2, D3 ! D3,
D4 ! D4, D5 ! D5, D6 ! D6, D1 ! D1, D2 for 120 electrical. Thus, the DC
output voltage is the envelope of the line voltages with six ripples each completed
cycle (Figure 13.4(b) and (c)), and the DC output voltage performance is enhanced.
The DC output voltage is given by
ð pffiffiffi
1 p=6 pffiffiffi pffiffiffi 3 6
Vo ¼ 2 3Vs cos ðwtÞ d ðwtÞ ¼ Vs ¼ 2:34Vs (13.2)
p=6 0 p
where Vs is the RMS value of the AC source phase voltage.
13.2.2 Phase-controlled rectifiers

Uncontrolled diode rectifiers can only supply load with constant output DC
voltage. The average output DC voltage, Vo , is fixed for a given load and given
source (mains voltage ‘vs ’). If an autotransformer is used, adjustable DC output
voltage can be obtained by varying the source AC voltage. However, this involves a
costly and bulky autotransformer. Alternatively, by replacing the diodes in the

circuits shown in Figures 13.2–13.4 by thyristors, the output voltage, Vo , can be
controlled by varying the thyristor switching angle, and an adjustable output DC
voltage can be obtained. Phase-controlled rectifiers with an efficiency greater than
90% are used in many industrial implementations, especially in variable-speed
drives.
A phase-controlled single-phase full-bridge rectifier is shown in Figure 13.5(a).
The principle of operation is that during the positive half cycle of the source voltage,
thyristors T1 and T2 are forward biased. If the gate signal is applied to thyristors T1
and T2 at wt ¼ a, the thyristors are turned on. Hence, the load current flows through
T1 and T2 and the output/input voltage waveforms are as shown in Figure 13.5(b). In
case of a large inductive load, thyristors T1 and T2 remain conducting even when the
AC source voltage becomes negative. Likewise, during the negative half cycle of the
AC source voltage, thyristors T3 and T4 are forward biased and thyristors T1 and T2
become backward biased. When the gate signal is applied to thyristors T3 and T4, the
thyristors T3 and T4 are turned on whereas thyristors T1 and T2 are turned off. The
load current is provided through T3 and T4 instead of T1 and T2.
For positive half cycle of AC source voltage vs during the period
wtða wt pÞ, the source current is and source voltage vs are positive and thus
the energy flows from the source to the load. Nevertheless, during the period
ðp wt p þ aÞ, is is positive and vs is negative. Hence, the energy (stored in the
large inductor) flows backwards. The line current waveform is depicted in
Figure 13.5(c). The DC output voltage is calculated by
ð pffiffiffi
1 aþp pffiffiffi 2 2
Vo ¼ 2Vs sin ðwtÞ d ðwtÞ ¼ Vs cos a (13.3)
p a p
pffiffi
It can be varied from 0 to 2 p 2 Vs by varying the firing angle a from p=2 to 0. If
the energy stored in the inductor is not enough to keep the load current flowing
continuously, the thyristors turn off as the load current becomes discontinuous. It is
to be noted that the converter can operate in the inversion mode and feed energy
into the grid if a negative source voltage is applied to the DC bus. By varying a
io
+ Vo
Input current
T1 T3 L
is
Voltage
vs Vo
T4 T2 R Vs
−
0 2 3 4 0 2 3 4
ωt (rad) ωt (rad)
(a) (b) (c)
Figure 13.5 Phase-controlled rectifier with a large inductive load when a ¼ p=3:
(a) single-phase full bridge, (b) input and output voltage waveform
and (c) line current
between p=2 and p, energy is fed back to the grid. The corresponding waveforms
when a ¼ 2p=3 are exhibited in Figure 13.6.
For high-power industrial applications, three-phase controlled bridge rectifiers
with thyristors as exhibited in Figure 13.7(a) are often employed. The thyristors are
triggered at the firing angle a at intervals of p/3 rad. On applying the triggering
pulse to the thyristors that are forward biased, the corresponding applied line-to-
line voltage is passed to the load. The waveforms of output voltage for a ¼ p=6
and a ¼ p=2 are presented in Figure 13.7(b) and (c), respectively. The DC output
voltage is given by
ð p pffiffiffi
1 aþ2 pffiffiffi pffiffiffi p 3 6
Vo ¼ 2 3Vs sin wt þ d ðwtÞ ¼ Vs cos a (13.4)
p=3 aþp6 6 p
pffiffi
The output voltage can be varied from 0 to 3 p 6 Vs by changing the firing angle
a from p/2 to 0. In a similar manner, the converter operates in the inversion mode
when negative supply voltage is applied to the DC bus to deliver energy to the grid
(Figure 13.8).
13.2.3 Diode rectifiers cascaded with a boost converter

A significant portion of the source currents of both the uncontrolled diode and
phase-controlled rectifiers is comprised of harmonic components, which results in a
Input current
Vs
Voltage
Vo
0 2 3 4 0 2 3 4
(a) ωt (rad) (b) ωt (rad)
Figure 13.6 Phase-controlled rectifier operated in the inversion mode when

a ¼ 2p=3 with a negative DC bus voltage present: (a) input and
output voltage waveform and (b) line current
io
Vab Vac Vbc Vba Vca Vcb Vo Vab Vac Vbc Vba Vca Vcb Vo
va ia T1 T3 T5
+
Voltage
Voltage
vb ib
Load Vo
vc ic
−
T4 T6 T2
α = /6 α = /2
0 2 3 4 0 2 3v 4
ωt (rad) ωt (rad)
(a) (b) (c)
Figure 13.7 Phase-controlled three-phase rectifier with a large inductive load:

(a) three-phase bridge rectifier, (b) a ¼ p=6 and (c) a ¼ p=2
Vab Vac Vbc Vba Vca Vcb Vo Vab Vac Vbc Vba Vca Vcb Vo
Voltage
Voltage
α = 2/3 α = 5/6
0 2 3 4 0 2 3 4
Figure 13.8 Phase-controlled three-phase rectifier operated in the inversion

mode when a negative DC bus voltage is present: (a) a ¼ 2p=3
and (b) a ¼ 5p=6
iL + VL − io vs
PLL ωt
|sin(ωt)|
L |sin|
is D1 D3 D ic +
vs Q Load Vo voref Q
C − PI x
Hysteresis
controller
D4 D2 − −
vo iL
(a) (b)
Voltage and current
Vs iL
Current
io
is
0 2 3 4 0 2 3 4
(c) ωt (rad) (d) ωt (rad)
Figure 13.9 Diode rectifier cascaded with a boost converter: (a) diode bridge
rectifier, (b) basic hysteresis control (PLL ¼ phase-locked loop),
(c) input voltage and input current and (d) iL and io
low power factor. Several current control techniques have been developed to obtain
a variable output voltage besides high-quality source currents at the same time
[10,11].
One solution is to apply a boost converter in cascade with the DC bus of the
uncontrolled bridge rectifier (Figure 13.9(a)). By controlling the boost converter in
combination with the basic hysteresis current control technique, the source current
can be brought in phase with the source voltage as well as regulate the output
voltage (Figure 13.9(b)). The source current becomes sinusoidal while the load
current is mainly DC, and the inductor current is full-wave rectified. It is to be
noted that the average inductor current is higher than the load current because of
the boosted DC output voltage. PWM pulses generated by the hysteresis current
controller are used to turn ON/OFF the switch Q according to the difference
between the inductor current iL and its estimated reference, which is in phase with
the required output voltage of the uncontrolled rectifier and hence the source
voltage.
Thus, the actual inductor current iL continuously follows the estimated refer-
ence current within a hysteresis band (HB) and the source current is in phase with
the source voltage [2]. The corresponding circuit waveforms are exhibited in
Figure 13.9(c) and (d).
13.2.4 PWM-controlled rectifiers

A bridge rectifier in cascade with a boost converter can enhance the quality of the
source current but the power flow is unidirectional from the supply to the load. This
problem can be solved by utilizing a bi-directional converter with fully PWM-
controlled power semiconductor switches.
A PWM-controlled converter can be used as either a rectifier or an inverter,
and a bi-directional flow of power can be achieved if energy is available on the load
side. A bi-directional PWM-controlled single-phase H-bridge converter is shown in
Figure 13.10(a). As it is mainly operated as a boost converter, an inductor is con-
nected to the supply voltage side. The aim of the control technique applied,
Figure 13.10(b), is to make the source current in phase with the source voltage. The
difference between the actual output voltage, vo ; and the reference output voltage,
voref , is fed into a PI controller to produce the required amount of the current being
absorbed from the supply. It is then multiplied by sin ðwt) of the per-unit source
voltage value as a synchronization signal to produce the reference source current.
io vs ωt |sin(ωt)|
Hysteresis Q1
PLL |sin|
Q1 Q3
is + controller Q4
vs L C Load Vo voref iL
PI x
− Q
Q4 Q2 Hysteresis 3
Q
vo controller 2
(a) (b)
Voltage and current
vs
Voltage
is
0 2 3 4 0 2 3 4
Figure 13.10 PWM-controlled single-phase full-bridge rectifier:

(a) PWM-controlled single-phase H-bridge rectifier,
(b) basic control strategy, (c) input current and voltage
waveform and (d) output voltage waveform
The source current of the converter is controlled through a hysteresis current con-
troller to track the estimated reference source current.
The principle of the hysteresis current control, Figure 13.11, is that when the
actual current is below the lower margin of an HB around the estimated reference
current, the upper and lower switches are turned ON and OFF, respectively, causing
the actual current to increase. When the actual current exceeds the upper margin of
the HB, the upper and lower switches again are turned ON and OFF, respectively,
causing the actual current to decrease. Consequently, the PWM pulses for the upper
switch Q1 are generated as follows:
8
> 1
< ON if iact < iref HB
Q1 ¼ 2 (13.5)
: OFF if iact > iref 1 HB
>
2
The PWM pulses for Q4 are complementary, and the PWM pulses for Q2 and
Q3 can be generated accordingly. The waveforms of both source current, input
voltage and output voltage of the converter are shown in Figure 13.10(c) and (d).
The source current and source voltage are in phase. The output voltage is main-
tained satisfactory despite the existence of some ripple, which can be handled by
other techniques. It is worth noting that the capability of changing the source power
factor to any desired value by controlling the phase of the synchronization signal.
Three-phase PWM-controlled converter, Figure 13.12(a), can be achieved with the
same principle and a slight change in the control strategy, Figure 13.12(b), to
accommodate the other two phases. The source and output voltage waveforms are
shown in Figure 13.12(c) and (d). The three-phase source currents are controlled to
be sinusoidal and in phase with the corresponding phase voltages. In a three-phase
system, the power input is constant, and the ripples in the output voltage waveform,
Figure 13.12(d), of the three-phase converter are much lower than those in a single-
phase converter (Figure 13.10(d)).
13.3 DC–DC conversion
The DC–DC converter is used to convert the output voltage level of a fixed DC
source. According to the ratio between the output and source voltages, the
Hysteresis band
Upper band
Reference current Actual current
Lower band
0 /2 0 /4 /2 3/4

Figure 13.11 Principle of hysteresis control: (a) HB, actual and reference
currents; (b) gate signal
5/6
|sin(ωt − 2/3)|
|sin|
io
vab |sin(ωt)|
PLL |sin|
Q1 Q3 Q5 ωt + /6
va i La
+ Q
a
/6 x Hysteresis 1
vb ib Lb
C Load Vo controller Q4
vc ic Lc − ia
voref Q5
Q4 Q6 Q2 PI Hysteresis
Q
controller 2
ic
vo Q3
Hysteresis
x controller Q6
(a) (b)
ib
Va Vb Vc
Voltage and current
Output voltage
ia ib ic
0 2 3 4 0 2 3 4
Figure 13.12 PWM-controlled three-phase full-bridge rectifier: (a) three-phase

PWM-controlled rectifier, (b) basic control strategy, (c) input
voltage and current and (d) output voltage
converter can be designed to step up or step down the output voltage, a function
similar to an AC transformer. In this case, the ratio between the source and output
voltage is called the conversion ratio a. For a < 1, the converter is called a ‘buck
converter’, whereas it is called a ‘boost converter’ for a > 1. If the converter can
increase/decrease the voltage level, that is, can operate for both a < 1 and a > 1,
the converter is called a ‘buck-boost converter’ [4,5].
13.3.1 Buck converter

A buck converter is a step-down DC to DC converter. A typical buck converter,
which comprises two switches (semiconductor switch and a diode), a capacitor and
an inductor, is shown in Figure 13.13(a). The capacitor and inductor act as passive
filters to enhance both the quality of the load current and the output voltage. The
switch Q is turned ON and OFF periodically.
If the switching period and the duty cycle are T and k, respectively, then the
ON time of one period is kT and the OFF time is (1 k)T. The circuit can operate
in two operation modes based on the state of the switch Q: mode #1 and mode #2
when turning ON and OFF, respectively. The corresponding equivalent circuits are
shown in Figure 13.13(b) and (c). For mode #1, the inductor current iL increases
linearly as
dil
L ¼ vs vo (13.6)
dt
is Q L io iL L io iL L io
+ + + + +
vs D C Load Vo vs C Load Vo D C Load Vo
− − −
− −
(a) (b) (c)
Inductor current
Output voltage
0 kT T (k + 1)T 0 kT T (k + 1)T
(d) Time (e) Time
Figure 13.13 Typical buck converter: (a) buck converter topology, (b) mode #1
(Q: ON), (c) mode #2 (Q: OFF), (d) inductor current and
(e) capacitor voltage
where vs and vo are approximately constant and positive. For mode #2, the inductor
current freewheels through the diode and decreases linearly as
dil
L ¼ 0 vo (13.7)
dt
where vo is approximately constant and negative. The related inductor current

waveform is exhibited in Figure 13.13(d).
Ideally, the ripple current should pass through the capacitor and the related
voltage waveform is presented in Figure 13.13(e). In the steady state, the net energy
exchange in the inductor during one period is almost zero; that is, the increase in
current in mode #1 and the decrease in current in mode #2 should be equal. Thus,
kT ð1 k ÞT
ðv s v o Þ ¼ vo (13.8)
L L
and the output voltage is
vo ¼ kvs (13.9)
So, it is a buck converter as a ¼ k.
13.3.2 Boost converter

The configuration of a typical boost converter (also known as a step-up converter)
is shown in Figure 13.14(a). The output voltage is higher than the source voltage,
and because of the power balance the output current is lower than the source cur-
rent. Similar to the buck converter, the boost converter also has two operation
iL L D io iL L io iL L io
+ + VL − ic + + + VL − ic + + + V −
L
ic +
Vs Q C Load Vo vs C Load Vo vs C Load V
o
− − − −
− −
(a) (b) (c)
Inductor current
Output voltage
0 kT T (k + 1)T 0 kT T (k + 1)T
Time Time
(d) (e)
Figure 13.14 Typical boost converter: (a) boost converter topology, (b) mode #1
modes when the switch Q is turned ON and OFF. The equivalent circuits of the two
modes are shown in Figure 13.14(b) and (c), respectively. In mode #1, the inductor
current increases linearly as
dil
L ¼ vs 0 (13.10)
dt
and the energy from the power source is stored in the inductor, whereas the capa-
citor discharges to supply the load. In mode #2, the energy stored in the inductor
and the energy supplied by the power source are both transferred to the capacitor
and the load. The inductor current decreases linearly as
dil
L ¼ vs vo (13.11)
dt
As in the buck converter, the net exchange of energy in the inductor during one
period in the steady state should be zero; that is, the value of both current increase
in mode #1 and current decrease in mode #2 should be equal. Thus,
kT ð1 k ÞT
ðv s Þ ¼ ðv o v s Þ (13.12)
L L
from which the output voltage can be expressed as
1
vo ¼ vs (13.13)
1k
Hence, it is a boost converter as a ¼ 11 k > 1 for ð0 k 1Þ. The inductor
current and the capacitor voltage waveforms are exhibited in Figure 13.14(d) and
(e), respectively.
is Q is D
D − + − −
vs L C Load Vo vs L C Load Vo L C Load Vo
iL ic io + − iL ic io + iL ic i +
o
(a) (b) (c)

Inductor current
Output voltage
0 kT T (k + 1)T 0 kT T (k + 1)T
(d) Time (e) Time
Figure 13.15 Typical buck-boost converter: (a) converter topology, (b) mode #1
13.3.3 Buck-boost converters

A typical configuration of buck-boost converter is depicted in Figure 13.15(a). The
output voltage polarity is opposite to that of the source voltage. Like the buck and
boost converters, this converter has two operation modes based on the state of
switch Q. The equivalent circuits are shown in Figure 13.15(b) and (c) and the
corresponding waveforms of the inductor current and capacitor voltage are shown
in Figure 13.15(d) and (e).
The output voltage is
k
vo ¼ vs (13.14)
1k
and the converter operates in the buck mode when k < 0:5 and in the boost mode
when k > 0:5.
13.4 DC–AC conversion

The device that adopts the process of DC–AC conversion is known as an inverter.
Depending on the type of the DC supply, there are different types of inverters, such
as (i) voltage-source inverter (VSI) if the supply is a voltage source in parallel with
a large capacitor across the DC bus and (ii) a current-source inverter (CSI) if the
supply is a current source with a large inductor connected in series with the DC
supply. Similarly, depending on the type of the inverter output, the inverter is
called voltage-controlled if the output is controlled to be a voltage source and
current-controlled if the output is controlled to be a current source. Thus, there are
voltage-controlled VSIs, current-controlled VSIs, voltage-controlled CSIs and
current-controlled CSIs [2]. According to the structure of VSI inverter, it can be
classified into two-level VSI and multi-level VSI [12] as illustrated in Figure 13.16.
The operation of a three-phase two-level VSI is demonstrated in Section 13.4.3.
The output voltage amplitude and frequency of the inverter can be set to a
desired value depending on the application. This can be easily accomplished with
PWM techniques [13–18]. The main objective of PWM controller is to modulate a
triggering signal with certain amplitude into a chain of pulses with adjustable widths
to trigger switches and obtain sinusoidal output voltage or current depending on the
control mode.
There are different PWM techniques that can be used with DC–AC inverter,
such as sinusoidal PWM (SPWM), space vector PWM (SVPWM), selective har-
monic elimination PWM [19] and HB controller (as described in Section 13.4.2).
The most widely used technique, SPWM, is highlighted in the next subsection.
13.4.1 Sinusoidal PWM

It is required of most inverters that they produce a pure sinusoidal voltage output
with a desirable frequency, which is typically much lower than the switching fre-
quency. The desired pure sinusoidal output voltage, used as a modulating signal,
can be compared with a triangular carrier wave at the switching frequency to pro-
duce a sequence of pulses, as exhibited in the left column of Figure 13.17. The
harmonic components of this signal are generally around the multiples of the
switching frequency. The output voltage waveform has the same shape as the
generated trigger pulses.
The carrier frequency, that is, the switching frequency, is designed to be high
so that the harmonic components can be simply filtered out using a low-pass filter
(commonly, an LC or LCL filter). It is noted that higher switching frequency causes
a higher switching loss; thus, switching frequency must be properly designed to
Pulse generation technique DC–AC

conversion
SPWM
+ Volta
SVPWN ource s g
paral e source
ent s e
Selective harmonic Curr or in seri lel ca
pacit
+
du ct or
elimination PWM CSI in VSI
HB controller
Output control
Current-controlled CSI Two-level VSI Multi-level VSI

(current source)
Output control
Output control
Current-controlled VSI
Voltage-controlled CSI
(current source)
(voltage source)
Voltage-controlled VSI
(voltage source)
Figure 13.16 Types of DC–AC conversion devices

eliminate the harmonic component without increasing the switching losses. This
type of modulation technique is called SPWM. The frequency of the reference
signal controls the frequency of the output voltage. Its peak amplitude determines
the modulation index and in turn the RMS value of the output voltage.
Consequently, the amplitude and frequency of the output voltage can be simply
controlled by changing the modulating signal. The SPWM is bipolar, as the carrier
alters its sign during the positive or negative half cycle. If the carrier does not alter
its sign during the positive or negative half cycle, the resulting SPWM is unipolar,
as exhibited in the right column of Figure 13.17.
In both cases, the upper and lower switches on the same leg are operated in a
complementary way. Likewise, for three-phase applications, three modulating
signals can be compared with the carrier signal to produce the gate-driving signals,
as exhibited in Figure 13.18.
Modulating signal Carrier
Carrier Modulating signal

(a)
(b)
(c)
Figure 13.17 SPWM for a single-phase inverter: bipolar (left column) and
unipolar (right column). (a) Sketch of modulation. (b) Gate signal
for the upper switch. (c) Gate signal for the lower switch
va vb vc
Carrier
(a) (b)
Figure 13.18 SPWM for a three-phase inverter: (a) sketch of modulation and
(b) three-phase gate signals (for upper switches)
13.4.2 Operation of single-phase inverters

A single-phase inverter with a DC voltage source is shown in Figure 13.19. To
illustrate the principle of operation of the inverter, the DC-bus voltage is split into
two halves and the mid-point of the DC bus is considered as a reference point for
the two-phase legs.
iin
Q1 Q3
VDC L io
2 a
N C
Vo
VDC
b
2
Q4 Q2
Figure 13.19 Single-phase VSI
VDC/2 VDC/2
−VDC/2 −VDC/2
0 /2 3/2 2 0 /2 3/2 2

(a) ωt (rad) ωt (rad)
VDC/2 VDC/2
−VDC/2 −VDC/2
0 /2 3/2 2 0 /2 3/2 2

ωt (rad) ωt (rad)
(b)
vab vab
vo
VDC VDC
vo
−VDC −VDC
0 /2 3/2 2 0 /2 3/2 2

(c) ωt (rad) ωt (rad)
Figure 13.20 Unipolar operation of a single-phase inverter, with only one leg
operated at the switching frequency (left column) and both legs
operated at the switching frequency (right column): (a) vaN , (b) vbN ,
(c) vab and vo
The inverter can be operated to produce unipolar and bipolar SPWM signals
for vab . In case of operation with the unipolar SPWM signal presented in the right
column of Figure 13.17, the voltages vaN , vbN , together with vab and vo are exhibited
in the left column of Figure 13.20. The phase-leg a is operated according to the
unipolar SPWM signal and phase-leg b is operated according to the polarity of the
voltage signal at its frequency. For example, Q1 and Q4 are turned ON and OFF
when the modulating voltage v is positive according to the unipolar SPWM signal,
whereas Q2 is always ON and Q3 is always OFF when v is negative. Q1 and Q4 are
turned ON and OFF according to the SPWM signal whereas Q2 is always OFF and
Q3 is always ON. Because the second leg is operated at the frequency of the vol-
tage, depending on the type of the DC supply the switching losses can be decreased.
It is possible to operate the inverter to attain a unipolar SPWM for vab even if
the phase legs are driven by bipolar SPWM signals, as presented in the right col-
umn of Figure 13.20. In such a case, the modulating voltage v and its opposite are
compared with the carrier waveform to produce two sets of bipolar SPWM signals
to drive the two-phase legs. Consequently, the voltages vaN and vbN are 180 apart
from each other. The voltage vab , as a difference between the two voltages vaN and
vbN , is unipolar at the doubled switching frequency. The switching losses are high
as both phase legs are operated at the same high switching frequency. Nevertheless,
the output voltage quality is better than the case exhibited in the left column of
Figure 13.20 because the resulting vab has a doubled switching frequency. So, the
two-phase legs are operated as two separate phases, which are 180 apart from each
other. When both legs of the inverter are operated with the bipolar SPWM as shown
in the left column of Figure 13.17, the resulting waveforms are as shown in
Figure 13.21. The same SPWM signal is sent to the two-phase legs in a com-
plementary way. That is, (Q1, Q2) and (Q3, Q4) are two sets of a pair of switches for
VDC/2 VDC/2
−VDC/2 −VDC/2
0 /2 3/2 2 0 /2 3/2 2

vab vo
VDC
−VDC
0 /2 3/2 2
(c) ωt (rad)
Figure 13.21 Bipolar operation of a single-phase inverter: (a) vaN , (b) vbN and
(c) vab and vo
each. Switches of each set are operated simultaneously. As a result, vbN ¼ vaN
and vab ¼ 2vaN .
The amplitude of vab is VDC for the three operation modes and the maximum
attainable amplitude is the same as the DC bus voltage.
13.4.3 Operation of three-phase inverters (two-level)

For three-phase inverters, Figure 13.22(a), three sets of bipolar PWM signals are
generated by comparing the three-phase voltages with the carrier waveform,
Figure 13.18, to drive the three-phase legs individually. The related waveforms are
iin
Q1 Q3 Q5
La
VDC va
a
2 Lb
N vb
b
Lc
VDC c vc
2 Q4 Q6 Q2
Ca Cb Cc
(a) N
va vaN
VDC/2
−VDC/2
vbN vb
VDC/2
−VDC/2
vcN vc
VDC/2
−VDC/2
0 /2 3/2 2
(b) ωt (rad)
Figure 13.22 Operation of a three-phase inverter: (a) topology and

(b) vaN ; vbN ; vcN and the three-phase voltages va , vb and vc
shown in Figure 13.22(b). The maximum amplitude of the phase voltages is half of
the DC-bus voltage.
13.5 AC–AC conversion

AC–AC conversion can be achieved either indirectly through AC–DC–AC with
adding a DC bus in between or directly without using a DC bus. The indirect AC–
AC conversion (back-to-back) is mainly a combination of AC–DC conversion and
DC–AC conversion, as discussed in Sections 13.2 and 13.4. The direct AC–AC
conversion implies bi-directional switches such as thyristors connected in anti-
parallel or triacs. One approach to implement direct AC–AC conversion is to
operate matrix converters to produce AC outputs with arbitrary amplitude and
frequency. The topology of a single-phase AC–AC converter is shown in
Figure 13.23(a), where two main control methods can be used: ON–OFF control
and phase control.
13.5.1 ON–OFF control

Two thyristors are connected in antiparallel to pass both half cycles of the supply to
the load when they are triggered. Triggering signals are provided by an ON–OFF
control. The thyristors are turned ON to pass the source voltage to the load and
turned OFF when the source voltage is not pffiffipassed
ffi to the load. It can be found p
that
ffiffiffi
the RMS value of the output voltage is k V and the source power factor is k ,
T1
is io
Input voltage
vs T2 Load vo
0 2 3 4 5 6 7 8
(a) (b) ωt (rad)
Output voltage
Output voltage
0 2 3 4 5 6 7 8 0 2 3 4 5 6 7 8
ωt (rad) ωt (rad)
(c) (d)
Figure 13.23 Single-phase AC–AC converter: (a) topology, (b) input voltage,
(c) output voltage under ON–OFF control and (d) output voltage
under phase control
where k is the ratio of the number of ON cycles to the number of total cycles in an
operational period and the RMS value of the supply is V. For the source voltage
sketched in Figure 13.23(b), the related output voltage is as exhibited in
Figure 13.23(c) for one OFF cycle. The thyristors are triggered when the source
voltage crosses 0.
13.5.2 Phase control

For the circuit shown in Figure 13.23(a), the voltage phase can be controlled when
both half cycles are turned ON through triggering the thyristors at a specific time. It
is found that compared to a case with thyristor bridge rectifiers there is no big
difference, aside from the fact that both half cycles can be passed to the load. The
source voltage as sketched in Figure 13.23(b) produces an output voltage as shown
in Figure 13.23(d) for a triggering angle of p=6 rad. The output voltage encom-
passes harmonics and the switches are not triggered when the voltage crosses 0.
13.6 Output filters

The output voltage and current waveforms of the inverter often contain harmonics
in a multiple of the switching frequency as a result of the effect of PWM pulses.
Consequently, a low-pass filter is required to be connected to the output of the
inverter to filter out the harmonics of the output voltage. Different configurations of
the passive low-pass filter can be used in conjunction with inverters to filter out
harmonics, for example, L, LC and LCL filters [20–22].
The L filter is also called choke filter. It consists of an inductor L connected
between the inverter output and the load. The inverter output voltage waveform
contains AC components in addition to DC components. The inductor provides a
high-impedance to the high-frequency components and low-impedance to the low-
frequency components. Thus, the high-frequency components of the output voltage
are blocked, and only the fundamental frequency component is applied to the load.
LC and LCL filters are briefly described in the following subsections.
13.6.1 LC filters
The circuit of an LC filter includes an inductor connected in series with the load
and a capacitor connected in parallel across the load. The inductor offers a high-
impedance to the high-frequency component and allows fundamental component to
flow through to the load, while the capacitor filters out any high-frequency com-
ponents flowing through the inductor. A passive LC filter can be represented by a
circuit shown in Figure 13.24. The equivalent internal resistances of the capacitor
and the inductor may be neglected during the design procedure of the filter due to
their small values. Practically, the internal resistances can dampen high-frequency
components. The capacitance and the inductance should be designed properly to
filter out the switching effects considering several factors [23], for example, the
cut-off frequency fc [24,25], the voltage total harmonic distortion (THD), size, the
cost function [26–28], resonance damping, efficiency [29], power level and so on.
The cut-off frequency fc of the filter is derived as
1
fc ¼ pffiffiffiffiffiffiffi (13.15)
2p LC
The most important consideration is the filtering out of switching harmonics.
Although it is desirable to keep the cut-off frequency much lower than the
switching frequency, it should be selected such that it still provides enough band-
width for the controller. It is recommended that the cut-off frequency is selected
within 13 to 12 of the switching frequency fsw [24] as in the following:
fsw fsw
fc (13.16)
3 2
It is to be noted that this may cause resonance in the output impedance of the
inverter near the cut-off frequency. This amplifies the harmonic current oscillations
near the cut-off frequency and might cause high THD in the output voltage. It is,
therefore, necessary to not select the cut-off frequency around the band where the
major harmonic components of the load current exist. The cost function [26–28] of
the LC filter can be expressed as
2QL þ Qc
COST ¼ Pn (13.17)
h¼1;odd jVoh Ih j
and
X
n
QL ¼ jIh j2 XLh (13.18)
h¼1;odd
X n
jVCh j2
QC ¼ (13.19)
h¼1;odd
XCh
L RL
L
C
C
RC
(a) (b)
Figure 13.24 Circuit model of a passive LC filter: (a) without considering the
internal resistance and (b) with the internal resistance considered
Heuristically, the size of the capacitor C should be small for high-voltage

applications and that of the inductor L small for high-current applications main-
taining the same cut-off frequency. The values of both capacitance C and induc-
tance L should be small in situations, for example, under low switching frequency
where, respectively, large amount of voltage and current harmonics exist. When the
internal resistances of the filter, RC and RL , are considered or purposely increased,
the LC resonance within the cut-off frequency is dampened. Nevertheless, the
existence of RC and/or RL results in excessive power losses, which is considered as
a challenge in the LC filter design for high-power applications. One possible
solution is to implement control techniques to add virtual internal resistors to obtain
the same results [30,31]. In any case, the ratings, current for the inductor and vol-
tage for the capacitor, must meet the specified requirements of the inverter design.
13.6.2 LCL filters

The passive LCL filter, represented by the circuit shown in Figure 13.25, is com-
monly adopted in grid-connected inverters. The increase in the order of the LCL
filter by 1, because of the additional grid-side inductor Lg, means that better
attenuation of the current harmonics can be achieved by an LCL filter than an LC
filter. It also provides an additional mechanism for limiting the current harmonics
that may be caused by harmonics in the grid voltage.
It is very important to pay attention to selecting the value of the switching
frequency in designing the LCL filter. For high switching frequency applications,
the LC filter is designed first and then the grid-side inductor is added. For very low
switching frequency applications at MW level, it is important that extra attention
should be devoted to check the amount of reactive power of the filter capacitor,
such that it does not significantly affect the rating of the inverter.
L Lg
Figure 13.25 Circuit model of a passive LCL filter

13.7 Case studies

The importance and performance of the power electronics-based converters are
illustrated through two case studies: Case 1, the application of grid-connected
photovoltaic (PV), and Case 2, the application of grid-connected doubly fed
induction generator (DFIG) wind turbine (WT).
The two cases are studied using MATLAB/Simulink software with various
output results.
13.7.1 Case 1: grid-connected PV

A PV module is interfaced to a grid through power electronics two-level converter
connected to the secondary of step-up distribution transformer. The PV design
parameters are listed in Table 13.1. The converter objective is to obtain the max-
imum power point of PV and inject the PV-generated power to the grid. The block
diagram of the system is shown in Figure 13.26.
The sun irradiance is varied as shown in Figure 13.27 to validate the converter
performance in tracking the maximum power supplied by PV. The PV-generated
power is injected to the distribution network using two-level converter as shown in
Figure 13.28. The DC-link voltage and the waveform of converter output current/
voltage are shown in Figures 13.29, 13.30 and 13.31, respectively.
The RMS values for the converter output current and voltages are exhibited in
Figures 13.32 and 13.33. The converter is keeping the balance of the RMS values of
the output voltages and within the allowable range during the variation of the light
irradiance.
13.7.2 Case 2: grid-connected DFIG WT

The WT is connected to a DFIG to generate electrical power using energy extracted
from the wind. As the wind speed is varying, the rotor is connected to a back-to-
back converter to control the phase and magnitude of rotor voltage and, thereby,
Table 13.1 PV system parameters
System quantity Value Unit

PV module open circuit voltage (voc) 85.3 V
PV module voltage at maximum power point (vmp) 72.9 V
PV module short circuit current (isc) 6.09 A
PV module current maximum power point (imp) 5.69 A
PV module max. power (pmax) 414.8 W
Cells per module 128 Cell
Series modules 11 Module
Shunt modules 17 Module
DC-link voltage 760 V
DC-link capacitance 1,000 mF
AC
Tr. 2 Grid
T.L. 30 km 220/22 kV 220 kV
15 MVA
DC-link PV Tr. 1 T.L. 15 km

LC filter
75 kW PV interface 0.4/22 kV
75 kVA
Load Load
50 kW 500 kW
Figure 13.26 Block diagram of a grid-connected PV
1,200
1,000
Irradiance (W/m2)
800
600
400
200
0
0.5 1 1.5 2 2.5 3
Time (s)
Figure 13.27 Sun irradiance during the study
control the magnetic field frequency or speed. It forms an AC excitation with a

variable frequency to match the actual wind speed. This operation gives the DFIG
the benefits of a synchronous generator. These benefits include the separate control
of reactive and active power, or the control of the WT’s power factor. A block
diagram of the studied system is shown in Figure 13.34. The system parameters are
listed in Table 13.2.
The wind speed is assumed to be varied to assure the converter performance
during the different wind speeds (Figure 13.35).
80
Active power (kW)
70 Reactive power (kVAr)
60
Converter output power
50
40
30
20
10
0 0.5 1 1.5 2 2.5 3

Time (s)
Figure 13.28 PV-injected power
1,000
800
DC-link voltage (V)
600
400
200
0
0.5 1 1.5 2 2.5 3
Time (s)
Figure 13.29 DC-link voltage
The active power output generated from the WT during the variation of the
wind speed is shown in Figure 13.36. The active power varies according to the
wind speed and the converter adjusts the excitation of the DFIG to generate power
during low wind speed.
The DC-link voltage of the back-to-back converter during the variation of wind
speed remains constant at the pre-specified level as shown in Figure 13.37.
200
150
Output converter current (A)

100
50
0
–50
–100
–150
–200
0.9 0.95 1 1.05 1.1 1.15 1.2
Time (s)
Figure 13.30 Waveforms of the converter output current
400
300
Output converter voltage (V)
200
100
−100
−200
−300
−400
0.9 0.95 1 1.05 1.1 1.15 1.2
Time (s)
Figure 13.31 Waveforms of the converter output voltage
100
RMS converter current (A)
80
60
40
20
0.5 1 1.5 2 2.5 3

Time (s)
Figure 13.32 RMS values of the converter output current

250
RMS converter voltage (V)
240
230
220
210
200
0.5 1 1.5 2 2.5 3
Time (s)
Figure 13.33 RMS value of the converter output voltages
AC
T.L. 30 km
Grid Tr. 2
220 kV 220/22 kV
16 MVA
DFIG
Gear box
Tr. 1
3 MW WT 0.4/22 kV
(6 * 0.5 MW) 4.5 MVA
Rotor DC-link Grid LC filter

converter converter
Figure 13.34 Block diagram of a grid-connected DFIG WT
Table 13.2 WT system parameters
System quantity Value Unit

DFIG power 500 kW
DFIG stator voltage 400 V
No. of Connected WTs 6 Number
Total connected WT power 3 MW
Wind speed at maximum efficiency 11 m/s
DC-link voltage 1,150 V
DC-link capacitance 5,000 mF
Wind speed (m/s) 15
10
0
1 2 3 4 5 6 7 8 9 10
Time (s)
Figure 13.35 Wind speed variation
2.5
Active power (MW)
Reactive power (MVAr)
2
1.5
WT output power
0.5
−0.5
1 2 3 4 5 6 7 8 9 10
Time (s)
Figure 13.36 WT-generated power
The RMS values of the WT output current during different stages are shown
in Figure 13.38. The converter maintains the three-phase RMS current values
balanced during the different stages. The three-phase RMS voltage values are kept
balanced and within the allowable range as shown in Figure 13.39.
The waveforms of the WT output currents and voltages are maintained balanced
and sinusoidal during different generation stages as shown in Figure 13.40(a) and (b).
1,500
1,000
DC-link voltage (V)
500
1 2 3 4 5 6 7 8 9 10
Time (s)
Figure 13.37 Back-to-back converter DC-link voltage
3,500
3,000
RMS output WT current (A)
2,500
2,000
1,500
1,000
500
1 2 3 4 5 6 7 8 9 10
Time (s)
Figure 13.38 RMS values for WT output current

250
RMS WT output voltage (V)
240
230
220
210
200
1 2 3 4 5 6 7 8 9 10
Time (s)
Figure 13.39 RMS values for WT output voltage
4,000 400
3,000 300
Output WT voltage (V)
Output WT current (A)
2,000 200
1,000 100
0 0
−1,000 −100
−2,000 −200
−3,000 −300
−4,000 −400
3 3.1 3.2 3.3 3.4 3.5 3.6 3 3.1 3.2 3.3 3.4 3.5 3.6
Time (s) Time (s)
(a) (b)
Figure 13.40 (a) Waveforms of WT output current and (b) waveforms of WT

output voltage
References
[1] Chakraborty C., Iu H. H. and Lu D. D. ‘Power converters, control, and
energy management for distributed generation’. IEEE Transactions on
[2] Bimal B. K. Modern power electronics and AC drives. Vol. 123. NJ, USA:
Prentice Hall; 2002.
[3] Erickson R. W. and Maksimovic D. Fundamentals of power electronics.
New York: Springer ScienceþBusiness Media; 2007.
[4] Mohan N. A first course on power electronics and drives. New York: John
Wiley & Sons Inc.; 2012.
[5] Rashid M. H. Power electronics: Circuits, devices and applications. 4th edn.
London: Pearson Press; 2013.
[6] Fisher M. J. Power electronics. Boston, USA: PWS Kent Publishing Co; 1991.
[7] Thorborg K. Power electronics. New York: Prentice Hall; 1988.
[8] Vithayathil J. Power electronics: Principles and applications. New York:
McGraw-Hill; 1995.
[9] Baxter R. Energy storage – A nontechnical guide. Tulsa, Oklahoma, USA:
PennWell Press; 2006 p. 55.
[10] Bollen M. H. J. ‘Understanding power quality problems’. Voltage sags and
interruptions. New York: John Wiley & Sons; 2000.
[11] Sankaran C. Power quality. New York: CRC Press; 2001.
[12] Mittal N., Singh B., Singh S. P., Dixit R. and Kumar D. ‘Multilevel inverters:
A literature survey on topologies and control strategies’. Proc. Int. Conf.
Power Control Embedded Syst.; Allahabad, India, Dec 2012, pp. 1–11.
[13] Asiminoaei L., Aeloiza E., Enjeti P. N. and Blaadjerg F. ‘Shunt active-power-
filter topology based on parallel interleaved inverters’. IEEE Transactions on
Industrial Electronics, 2008;55(3):1175–89.
[14] Holtz J. ‘Pulsewidth modulation – A survey’. IEEE Transactions on
[15] Lascu C., Asiminoaei L., Boldea I. and Blaabjerg F. ‘High performance
current controller for selective harmonic compensation in active power fil-
ters’. IEEE Transactions on Power Electronics. 2007;22(5):1826–35.
[16] Holmes D. G. and Lipo T. A. Pulse width modulation for power converters:
Principles and practice. Hoboken, NJ, USA: Wiley-IEEE Press; 2003.
[17] Alper Ç. and Ermis M. ‘VSC-based D-STATCOM with selective harmonic
elimination’. IEEE Transactions on Industry Applications. 2009;45(3):1000–15.
[18] Wong P., Xu P., Yang P. and Lee F. C. ‘Performance improvements of
interleaving VRMs with coupling inductors’. IEEE Transactions on Power
Electronics. 2001;16(4):499–507.
[19] Shehu G. S., Kunya A. B., Shanono I. H. and Yalcinoz T. ‘A review of
multilevel inverter topology and control techniques’. Journal of Automation
and Control Engineering. 2016;4(3):233–41.
[20] Chang G. W., Chu S. Y. and Wang H. ‘A new method of passive harmonic
filter planning for controlling voltage distortion in a power system’. IEEE
Transactions on Power Delivery. 2006;21(1):305–12.
[21] Das J. C. ‘Passive filters-potentialities and limitations’. IEEE Transactions
on Industry Applications. 2004;40(1):232–41.
[22] Hamadi A., Rahmani S. and Al-Haddad K. ‘A hybrid passive filter config-
uration for VAR control and harmonic compensation’. IEEE Transactions
on Industrial Electronics. 2010;57(7):2419–34.
[23] Pasterczyk R. J., Guichon J. M., Schanen J. L. and Atienza E. ‘PWM inverter
output filter cost-to-losses tradeoff and optimal design’. IEEE Transactions
on Industry Applications. 2009;45(2):887–97.
[24] Hatua K., Jain A. K., Banerjee D. and Ranganathan V. T. ‘Active damping
of output LC filter resonance for vector-controlled VSI-fed AC motor
drives’. IEEE Transactions on Industrial Electronics. 2012;45(2):334–42.
[25] Michels, L., de Camargo R. F., Botteron F., Grudling H. A. and Pinheiro H.
‘Generalised design methodology of second-order filters for voltage-source
inverters with space-vector modulation’. IEE Proceedings-Electric Power
Applications. 2006;153(2):219–26.
[26] Dewan S. B. and Ziogas P. D. ‘Optimum filter design for a single-phase
solid-state UPS system’. IEEE Transactions on Industry Applications.
1979;15(6):664–9.
[27] Dewan S. B. ‘Optimum input and output filters for a single-phase rectifier
power supply’. IEEE Transactions on Industry Applications. 1981;17(3):
282–8.
[28] Kim J., Choi J. and Hong H. ‘Output LC filter design of voltage source
inverter considering the performance of controller’. PowerCon 2000.
Proceedings of 2000 International Conference on Power System Technology;
Perth, WA, Australia, Dec 2000. IEEE; 2000, Vol. 3, pp. 1659–64.
[29] Strom J. P., Korhonen J., Tyster J. and Silventoinen P. ‘Active du/dt – New
output-filtering approach for inverter-fed electric drives’. IEEE Transactions
on Industrial Electronics. 2011;58(9):3840–7.
[30] Dahono P. A. ‘A method to damp oscillations on the input LC filter of
current-type AC–DC PWM converters by using a virtual resistor’.
Proceedings of the 25th International Telecommunications Energy
Conference, 2003. INTELEC’03; Yokohama, Japan, Oct 2003. IEEE; 2003,
pp. 757–62.
[31] Guo S. and Liu D. ‘Analysis and design of output LC filter system for
dynamic voltage restorer’. 2011 Twenty-Sixth Annual IEEE Applied Power
Electronics Conference and Exposition (APEC); Fort Worth, TX, USA,
March 2011. IEEE; 2011, pp. 1599–605.
Chapter 14
Integration of distributed energy resources
14.1 Introduction
To solve problems such as high energy costs or low electric power reliability at
consumer’s facilities, distributed energy resources (DERs) could be the solution
that energy managers are looking for.
DERs can deliver the same electricity services provided by centralized
resources, including large-scale generators, and transmission and distribution net-
work assets. Many DERs can be deployed at different scales and exhibit econom-
ical and technical benefits. For instance, solar PVs can be deployed at the kilowatt
scale on residential rooftops, at the scale of several hundred kilowatts to 10 MW on
commercial rooftops or ground-mounted arrays, or tens to hundreds of megawatts
rating in the utility-scale solar farms. DER technology is developing at a fast rate,
and it can provide electrical energy where required. In addition, DER systems can
be either connected to the local electric grid or isolated from the grid in stand-alone
applications [1].
DER technologies are mainly classified into distributed generation (DG)
resources (renewable and non-renewable) and energy storage systems (ESSs) as
depicted in Figure 14.1 [2]. It includes the following:
● As explained in Part I, renewable energy sources, such as wind, solar, ocean
and biomass energy, (i) can be directly connected to the consumers’ load on
the distribution systems at the medium- or low-voltage level, (ii) can be dis-
patched to the grid based on the request of the utility grid operators or (iii) can
operate as a stand-alone system to serve a particular load or an area.
● Non-renewable energy sources have the technologies that use fossil fuels (e.g.
natural gas, coal and oil) to produce energy for various operations. They can be
used for power generation and thermal energy through a co-generation
arrangement. Examples of non-renewable technologies are the following:
– Microturbines, which run on natural gas at high speed (~90,000 rpm) to
drive an electric generator with power rating of 30–250 kW for small
industrial and commercial applications. They have the advantage of very
low emissions.
– Gas turbines (or combustion turbines), which use gas or liquid fuel. They
are suited for peaking and load-following applications as well as for base-
load operation in large sizes.
Distributed energy resources

(DERs)
DG
Energy storage
systems
Renewable energy Non-renewable
resources energy resources Mechanical
Chemical
Electrochemical
Wind energy Microturbines Electrostatic
Solar energy Gas turbines Magnetic
Ocean energy Diesel engine Thermal
Biomass energy generator set
Dual-fuel engine
generator set
CHP
Figure 14.1 Classification of DERs
– Diesel engine generator sets: They use diesel as fuel and consist of a
compression ignition reciprocating engine as a prime mover coupled to an
electrical generator. They operate at a relatively high compression ratio
and relatively low speed. Diesel engines are manufactured in a wide range
of sizes, from 1 kW to 10 MW, and can be cycled frequently to operate as
peak-load or load-following power plants. In stand-alone operation, diesel
engines can be run in base-load mode. Major drawbacks are very high
level of both emissions and noise, but on the other hand diesel engines
generator sets are cost-effective, reliable and a widely used technology. To
reduce the level of emissions, the engine is modified using a mixture of
diesel (low percentage) and natural gas (as high percentage) fuel. Diesel
fuel allows the use of compression ignition whereas the natural gas results
in low emissions. The engine generator set with this capability is called
‘dual-fuel engine generator set’.
– Cogeneration (combined heat and power (CHP)): Many types of DG can
provide useful and valuable thermal energy by capturing excess heat
energy produced during electricity generation and using it to heat or cool
water or interior space. The excess heat can be also used as energy input to
an industrial process, in lieu of burning fuel. This energy capture and reuse
process is known as CHP or cogeneration. So, plants including CHP units
are designed to produce thermal energy (heating and/or cooling) and
electricity, simultaneously, from a single source of energy in one process.
CHP units are sized to suit large-scale commercial buildings, institutional
facilities and communities with ways to reduce energy costs and total
emissions as well as providing more resilient and reliable electric power
and thermal energy.
Integration of distributed energy resources 347
● ESSs are responsible to store the imported electricity from a power grid at off-
peak periods in a form that can be converted back into electrical energy when
it is required. ESS technologies are classified into mechanical, chemical,
electrochemical, electrostatic, magnetic and thermal ESSs as explained in
Part II.
It is to be noted that some DER technologies, such as engines and combustion
turbines, produce AC power directly and can be connected to the electric grid or
AC loads in a straightforward manner. Other technologies, such as photovoltaics
(PVs) and fuel cells, produce DC power, which must be passed through a power
electronics device called an inverter or power conditioning unit to produce AC
output that can then be connected to the AC system. Electric power generated by
wind-driven turbines is generally variable frequency and requires rectifier/inverter
to convert to constant frequency power.
14.2 Powering DG systems
Different types of fuel based on availability, price and storage can be used to power
DG systems as in the following.
14.2.1 Non-renewable resources

Diesel fuel is readily available at a reasonable cost; however, it requires an on-site
storage tank and periodic deliveries, and spills or fumes are always possible.
Biodiesel fuel can be an alternative for diesel engine generator sets in areas
where it is available. Biodiesel is made by reacting natural oils and fats with an
alcohol. It can significantly reduce CO2, SO2 and matter emissions, but it may
increase NOx in some cases.
Natural gas is a cleaner alternative to diesel, and comparable in terms of cost.
Gas delivery and a storage tank may be required only when it is expensive or
difficult to run a gas line. Propane can be also used as an alternative to diesel fuel if
the natural gas is not available.
14.2.2 Renewable resources

Biological materials such as agricultural crops, residues from the food, fiber, trees,
grasses and forest product industries can be used in place of fossil fuels to generate
electricity. In addition, municipal solid waste facilities and sewage treatment plants
can also produce fuel to power DG renewable technologies.
PV systems depend on the availability of enough sunlight in a specific area. By
paying more attention to the design and placement of the solar panels, and adding
ESSs (typically, the deep discharge batteries), the potential of solar installation can
be maximized.
Wind energy is highly dependent on location, especially on the average wind
speed in a specific locale and the daily and seasonal variations in wind speed. So,
adding storage can pay off the overall functionality of a wind system.
DC bus AC bus
Power conditioning
system DC/AC
Solar PV
arrays
AC load
Power conditioning
system DC/AC
Power conditioning
system AC/DC
Diesel
Wind generator
turbines set
ESS
Figure 14.2 PV/wind/ESS/diesel generator hybrid system
Hybrid systems, such as PV and wind in addition to storage and diesel backup
generator, Figure 14.2, may be necessary and cost-effective in situations where
high reliability and/or stand-alone operations are required. Incorporating both ESS
and diesel generator minimizes the risks associated with the uncertainty of
renewable output, and consequently, the system reliability and energy services are
improved.
14.3 Benefits of DGs
The beneficial effects that result, or are expected to result, from the installation and
operation of DG in the electric system may include the following.
Environmental benefits that can be determined by the value of emissions
offsets or other environmental benefit such as
● reduction in land-use effects,
● reduction in greenhouse gas (GHG) emission pollutants, environment-
friendliness and reduction in health costs with renewable DG.
Technical benefits:
● Generation, transmission and distribution capacity deferral, which can be
evaluated by the financial value of deferring or avoiding a capital investment
in central generation capacity, transmission system capacity and distribution
system capacity, respectively.
● Usage of CHP results in cost savings from exploiting waste heat.
● Ancillary services provided by DG such as spinning reserve, regulation or

other ancillary services yield a potential revenue.
● Providing voltage control/var production results in a potential revenue.
● Reduction of system resistive losses and better control of reactive power.
● Reliability enhancement where both capacity release and peak power are
reduced as well as the generation diversity is improved.
● Improving the quality of the power at or nearby customer sites.
● Security enhancement for critical loads and power utilities. In addition,
security risks to the grid, impacts of cyberattacks and vulnerability of terrorist
attacks are reduced.
● Standby generation, which can help provide generation capacity dispatchable
by the utility or the customer during emergencies.
Economic benefits:
● Reduction in the cost of installation, auxiliaries’ costs, loss-associated costs
and operation and maintenance (O&M) costs.
● No fuel cost with renewable DG.
● Deferments of investment in infrastructures.
● Maintaining constant running cost for longer-time period.
● Achieving significant cost savings by utilizing the flexibility of DG, which can
start up and ramp quickly.
Furthermore, DG units incorporating ESSs (i.e. DERs) can be installed not
only as stand-alone energy resources to serve a specific load or an area but can be
integrated with the utility grid. The integration of DERs in a power system is
only accepted by utilities if the net benefit exceeds the investment costs of
installation. So, some parameters such as the benefit-to-cost ratio, cost of energy,
internal rate of return, return on investment and payback period should be esti-
mated and the revenue of the benefits, which result from the integration of DERs,
should be evaluated in comparison to the conventional alternatives. However,
economics of DER integration benefits can be analysed by considering the
following.
Economic benefits of DERs integration in power systems. To evaluate the
potential of DERs integration, it is crucial to identify the costs and benefits. DERs
compete with conventional generation and network assets to provide electricity
services. They can be deployed closer to the point of electricity consumption and in
locations inaccessible to more centralized resources. Consequently, DERs can
provide value to power systems where and when they can mitigate high marginal
network losses, relieve transmission network congestion or defer impending
upgrades and enhance reliability. These benefits are highly dependent on the sizing
and location of DERs, system configuration and type of DER technology. So,
DERs should be at locations in power grids where they are most valuable. If sited at
the right locations and operated at the right times, DERs can deliver more loca-
tional value. Otherwise, this could result in negative effects such as an increase in
system losses, voltage drop, harmonics and low voltage stability.
Locational value is constituted of four major items: (i) the cost of power loss
reduction by avoiding the electrical energy to be delivered in areas that experience
high marginal losses, (ii) the cost of avoiding or delaying an upgrade when an
investment in new network capacity would otherwise be required, (iii) the value
that customers place on avoiding service interruptions due to enhancing reliability
or resilience by supplying energy to load during network failures [3] and (iv) the
recent trend of providing electricity to the end users through low-carbon electricity
systems (e.g. renewables), in combination with the base-load conventional plants,
means that the provision of flexibility and resilience will increasingly shift towards
the flexibility inherent in distributed sources provided by end consumers. Thus,
renewable DG will transform passive consumers to active prosumers (professional
consumers) who may provide energy and service flexibility to both local and entire
systems (Figure 14.3).
Therefore, flexible DG, distributed ESSs as well as demand side management
(DSM) will help achieve significant cost savings in comparison to a system in which
flexibility and security of supply are delivered by conventional generation [4].
However, economies of size (known as unit scale) should be investigated when
using DER technologies. The unit costs of energy technologies typically fall as the
technology is installed at larger scales; for example, a 500-MW system of a given
technology will typically cost less per megawatt than a 5-MW system of the same
type. Fortunately, many technologies of DERs such as solar PVs, fuel cells and
electrochemical energy storage can be deployed across a wide range of scales. In
addition, the rate at which economies of scale are reduced is much more rapid for
DER technologies than that of traditional generation technologies. For instance,
economies of two PV systems of scales one to tens of megawatts and tens to
hundreds of megawatts have often cost roughly the same per unit installation cost.
In contrast, traditional nuclear, hydroelectric, natural gas and coal-fired power
plants typically exhaust economies of scale between several hundred megawatts
and more than 1,000 MW [3].
Therefore, an optimal trade-off between the value of locational benefits and
economies of unit scale should be found.
To select the optimum DGs configuration, application of optimization techni-
ques is necessary. Integration of renewable DG into a power system commonly
Flexible
Conventional power system prosumers
Transmission Sub-transmission
Distribution
Generation
Energy flow
Flexibility flow
Figure 14.3 Energy and flexibility flow

incorporate an ESS to make it dispatchable and thus more network capacity can be
delivered. For example, unit commitment and economic dispatch techniques can be
applied to implement economic analysis on different technologies and prioritize
them according to a desired criterion.
Emissions reduction (ER) benefits with integration of renewable DG units.
One of the major environmental impacts is the GHG emissions. It is widely accepted
that integration of renewable DG technologies into a power system can to a large
extent reduce the GHG emissions in addition to reduction of other pollutants such as
NO2, SO2 and CO2 that can be produced by the traditional fossil fuel power plants.
The percentage of ER for a specific pollutant (‘) by incorporating renewable
DG can be computed as in the following [2]:
PEl=wRE
ERl ¼ 100 (14.1)
PEl=woRE
where PEl=wRE and PEl=woRE are, respectively, the quantity of emissions for the ‘th
pollutant with and without application of renewable DG, and
X
m X
n
PE l=wRE ¼ xi yli þ xj ylj (14.2)
i¼1 j¼1
X
m
PE l=woRE ¼ xk ylk (14.3)
k¼1
where
m ¼ number of conventional generating units;
n ¼ number of renewable DG units;
xi ¼ quantity of electricity produced by the ith conventional power unit with
the application of renewable DG (MWh);
yli ¼ quantity of emission of the ‘th pollutant for the ith conventional power
unit per MWh of electricity produced;
xj ¼ quantity of electricity produced by the jth DG unit (MWh);
ylj ¼ quantity of emission of the ‘th pollutant for the jth DG unit per MWh of
electricity produced;
xk ¼ quantity of electricity produced by the kth conventional power unit
without application of renewable DG (MWh); and
ylk ¼ quantity of emission of the ‘th pollutant for the kth conventional power
unit per MWh of electricity produced.
If the number of pollutants is H, the total quantity of emissions with and
without the application of renewable DG can be given by
X
H X
H
PE total=wRE ¼ PE l=wRE and PEtotal=woRE ¼ PEl=woRE (14.4)
l¼1 l¼1
Thus, (14.1) can be rewritten as

PE total=wRE
ERtotal ¼ 100 (14.5)
PEtotal=woRE
From (14.4), the saving in quantity of pollutant emissions, PE saving , is given by
PEsaving ¼ PEtotal=woRE PEtotal=wRE
XH X
H
(14.6)
¼ PEl=woRE PE l=wRE
l¼1 l¼1
Hence, the cost saving because of pollutant ER, Ctotal/saving, can be computed by
Ctotal=saving ¼ Cav PE saving (14.7)
where Cav is the average price per unit quantity of pollutant emissions.
14.4 Operation requirements for DERs integration

into power systems
DERs, especially inverter-based resources, will need to provide more grid services
and functionalities that have been supplied by other means.
14.4.1 Voltage and frequency function

Renewable DG resources (e.g. solar PV and wind) supply a fluctuating power
depending on the climatic conditions. To integrate such resources into a power
system, the voltage and frequency at the point of coupling in case of normal and
abnormal operating conditions must be stabilized. An increase/decrease of resour-
ces generation tends to cause over/under-voltage, respectively. In addition, the
balance between generation and load is necessary to support the grid frequency.
Renewable DGs are coupled to the grid through power electronic inverters,
which are different from the traditional rotating machines. So, inverters should
participate in maintaining stable and reliable operation. The most feasible way to
do that is by autonomously responding to the AC terminal voltage waveform of the
inverter in a way that voltage and frequency are stabilized [5].
Voltage regulation. Inverters have two output parameters, reactive power
and active power, available to mitigate voltage deviation through a process
called volt-var control. In this process, reactive power is modulated in proportion
to voltage deviation, absorbing or producing reactive power for high and low
voltages, respectively, with minimal impact on the produced active power.
Reducing active power can also bring down the voltage, that is, mitigate over-
voltage by a process called volt-watt control, but under this control mode, the
produced active power is directly reduced. However, volt-var control is mostly
preferred, whereas volt-watt control is only used as an option when the voltage
rise is very high.
Frequency regulation. Supporting grid frequency necessitates a balance

between generation and load, that is providing more generation or less load when
frequency is low and reducing generation or more load when frequency is high.
Currently, inverters can reduce power in response to over-frequency events by a
process called frequency-watt control. If an inverter has additional active power
available, it can increase power in response to under-frequency. A simplified block
diagram indicating the inverter associated with volt-var and frequency-watt con-
trols for solar PV resource integration, as an example, is shown in Figure 14.4.
Voltage and frequency ride through functionality. It is the most basic grid-
support function required from inverter-based DERs. This function enables DERs
to ride-through voltage and frequency disturbances in lieu of tripping offline [6].
Tripping of inverters could exacerbate power system disturbances causing island-
wide blackout. IEEE standard-1547 requires DER resources to have ride-through
capability according to one of three categories: Categories I, II and III. Each
category specifies the DER behaviour within operating regions called continuous
operation region, mandatory operation region, permissive operating region and
momentary cessation operating region. Each region is defined by upper and lower
(U/L) voltage bounds and the corresponding cumulative time tome for ride-through
and clearing time for trip as summarized in Table 14.1.
For frequency ride-through, the same methodology is applied and DG beha-
viour is as given in Table 14.2.
Additional capabilities for inverter ride-through such as high rate of frequency
and voltage change, voltage-phase-angle change, phase imbalances and multiple
successive disturbances ride-through are required. With such capabilities, inverters
can remain connected under abnormal conditions and that is crucial to grid
stability.
Inverter control
Frequency-watt Volt-var control

control
P Q
f
Vo Power
'P estimator
Ii Vi
Rectifier and AC
filter
DC Transformer
AC grid
Solar PV resource Inverter
Figure 14.4 Simplified schematic diagram of an inverter control for

PV integration
Table 14.1 DER device behaviour for voltage ride-through Categories I, II and III
DER device behaviour Voltage ride-through categories

(operation regions)
Category I Category II Category III
U/L voltage Clearing time U/L voltage Clearing time U/L voltage Clearing time
bounds (p.u.) limits (s) bounds (p.u.) limits (s) bounds (p.u.) limits (s)
Continuous operation 0.90–1.10 0.01–1,000 0.90–1.10 0.01–1,000 0.90–1.10 0.01–1,000
Mandatory operation 0.70–0.88 0.01–1.20 0.65–0.88 0.01–0.80 0.50–0.88 0.01–0.40
Permissive operation 0.50–0.70 0.01–0.30 0.30–0.50 0.01–0.3 _________ __________
1.10–1.20 0.01–1.00 1.10–1.20 0.01–1.00
Momentary cessation _________ __________ _________ __________ 0.00–0.50 0.01–1.00
operation 1.10–1.20 0.01–11.00
For all categories: at regions of U/L 0.65–0.90 p.u. with clearing time 60–1,000 s and U/L 1.20–1.30 p.u. with clearing time 0.01–0.10 s, DG behaviour is ‘may ride-
through or may trip’. At regions of U/L 0.0–0.5 p.u. with clearing time 1–1,000 s and U/L 1.10–1.30 p.u. with clearing time 10–1,000 s, DG behaviour is ‘shall trip’.
Table 14.2 DER device behaviour for frequency ride-through Categories I, II

and III
DER device behaviour/ Regions for Categories I, II and III

operation regions
Frequency Cumulative time for
range (Hz) ride-through and clearing
time for trip (s)
Continuous operation 59–61 0.01–1,000
Mandatory operation 57 to <59 0.01–500
>61–61.7 0.01–500
May ride-through or may trip 56–57 and 61.7–63 0.01–0.2
56.5 to <57 and 61.7–62.2 0.2–500
Shall trip 56–56.5 and 62.2–63 0.2–500
56–58.5 and 61–63 500–1,000
The rated frequency is 60 Hz.
Large-scale tripping of
– Lines Stability loss Partial or
Transmission
and large-scale complete
network
– Loads outage of power system
congestion
power plants collapse
– Generators
Figure 14.5 Cascade processes of blackout
14.4.2 Power system restoration

As a result of major emergencies (e.g. dangerous congestion or severe faults),
blackouts occur in transmission networks due to cascade tripping of transmission
lines, loads and generators. These can jeopardize the stable operation regarding
voltage and frequency. Thus, loss of stability yields partial or complete collapse of
the power system (Figure 14.5) [7]. However, critical infrastructure, such as hos-
pitals and communication infrastructure, has backup power sources to confront
power outages, but for a specific period. Therefore, restoring the power system to
its stable operation is of utmost importance, especially if the cascade processes
develop very fast with a speed exceeding many times the capability of personnel to
perceive the situation. Furthermore, the conventional thermal power plants need
hours to restart-up.
Restoration procedures can be carried out through three consecutive stages
(Figure 14.6) [8]:
1. Collecting all necessary and available status information from the power sys-
tem control centre, which is securely communicated with local controllers at
Activation of
Information black-start Loads
collection power plants restoration
Stage I Stage II Stage III
Figure 14.6 Restoration procedures
different desired locations, to be acquainted with the availability of power

plants and grid infrastructure.
2. Activating black-start power plants (e.g. hydropower plants) to energize the
transmission network assets, lines and transformers, and supply cranking
power to non-black-start thermal power plants to enable them to start power
generation.
3. Restoring the loads.
Role of DERs. Use of DERs, especially renewable DGs such as wind and
solar, to increase the generation capacity of the distribution grid leads to challenges
in the restoration process, and meanwhile has additional potential control cap-
abilities. The challenges include the following:
● The active power is dependent on weather conditions and, consequently, may
reduce the availability of wind and solar power plants compared to non-
renewable DGs (e.g. biomass power plants and CHP units), and that might
disturb the restoration process in weak grid situation.
● Large penetration of uncontrollable weather-dependent feed-in from DERs can
exceed the power demand, and thus, impact the stable operation of thermal
power plants regarding minimum loading requirements and ramp rates.
● Interfacing large-scale DERs through electronic power conditioning circuits
can impact the system inertia.
● Communication links might not work in blackout situation leading to unavail-
ability of information that is required in the first stage of restoration procedures.
● Dynamics of frequency-dependent function are defined by the manufacturers,
not by most grid codes.
● Limitation of some control parameters, e.g. ramp rate, by manufacturers and
grid codes.
On the other hand, DERs are expected to become major contributors to
generation-load balance of the future grid. Biomass and hydropower plants are
normally interconnected with synchronous generators. Solar and wind power plants
use power electronics for grid integration. Control capabilities of synchronous
generators include inertia and standard grid-forming modes, whereas the control
capacities of inverter-based DGs include very fast control capabilities, possibility
of virtual inertia and potential capabilities in power system restoration including

the following:
● Starting up with comparably little auxiliary power within seconds or minutes,
whereas the thermal power plants need hours.
● Supporting power plants cranking and improving frequency stability.
● Using the reactive power feed-in capabilities to support voltage stability.
● Controlling DERs to maintain generation-load balance.
● Forming grid islands from DERs to reduce blackout duration. Backup gen-
erator sets or CHP units are state of the art for critical infrastructure.
14.4.3 Confirming reliability

Reliability is one of the key performance indicators that the planners of utilities use
to investigate the power system adequacy and security. Power utilities must stra-
tegize different mitigation techniques for power system reliability improvement
such as system structure, integration of DER units, installation of protection and
switching devices, and so on.
Incorporation of DER units into the electrical distribution system with a goal
of supporting the power system reliability entails the following:
● Coordination among all generation, including DER units, in an optimal manner,
e.g. specifying the penetration level of DER units under the existing constraints.
● Capability of DER units to maintain a continuous power supply at the load
points in the event of a power outage from the utility.
● Serving all necessary distributed system control capabilities by DER units.
However, a substantial improvement in the reliability of the distribution sys-
tem with penetration of DER units should be achieved. This can be confirmed by
calculating the reliability indices of the distribution system as reported in IEEE Std.
1366TM-2003 as in the following.
14.4.3.1 Sustained interruption indices

System Average Interruption Frequency Index (SAIFI): It is the average fre-
quency of sustained interruptions per customer over a predefined area:
Total number of customers interrupted

SAIFI ¼
Total number of customers served
System Average Interruption Duration Index (SAIDI): It measures the total

duration of interruptions. It is cited in units of hours or minutes per year:
Sum of all customer interupption durations

SAIDI ¼
Customer Average Interruption Duration Index (CAIDI): It is the average

time needed to restore service to the average customer per sustained interruption:
SAIDI Sum of all customer interruption durations
CAIDI ¼ ¼
SAIFI Total number of customers interrupted
Customer Total Average Interruption Duration Index (CTAIDI): It
represents the total average time in the reporting period that customers who
experienced an interruption were without power. This index is a hybrid of CAIDI
and is similarly calculated except that those customers with multiple interruptions
are counted only once:
Sum of customer interruption durations
CTAIDI ¼
Total number of customers interrupted
Customer Average Interruption Frequency Index (CAIFI): It gives the
average frequency of sustained interruptions for those customers experiencing
sustained interruptions. The customer is counted once regardless of the number of
times interrupted for this calculation:
SNi
CAIFI ¼
Nc
where Ni ¼ number of interrupted customers for each sustained interruption event
during the reporting period and Nc ¼ total number of customers who have experi-
enced a sustained interruption during the reporting period.
Average System Availability Index (ASAI): It represents the fraction of time
(often in percentage) that a customer has received power during the defined
reporting period:
Customer hours service availability
ASAI ¼
Customer hours service demand
Customers Experiencing Multiple Interruptions (CEMIn): It indicates the
ratio of individual customers experiencing more than n sustained interruptions to
the total number of customers served:
Nn
CEMIn ¼
where Nn ¼ total number of customers that experience more than n sustained
interruptions.
14.4.3.2 Momentary indices

Momentary Average Interruption Frequency Index (MAIFI): It indicates the
average frequency of momentary interruptions:
P
Total number of customer momentary interruptions
MAIFI ¼
Momentary Average Interruption Event Frequency Index (MAIFIE): It

has the same definition as MAIFI, but it does not include the events immediately
preceding a lockout:
P
Total number of customer momentary interruption events
MAIFIE ¼
14.4.3.3 Load-based indices

Average System Interruption Frequency Index (ASIFI): It includes the magni-
tude of the load unserved during an outage:
P
kVAsustained
ASIFI ¼
kVAserved
Average System Interruption Duration Frequency Index (ASIDI): It
includes the magnitude of the load unserved during an outage.
P
kVAsustained Dsustained
ASIDI ¼
Nserved
14.4.4 Virtual inertia

Virtual inertia is also called ‘synthetic inertia’. Some renewable DG resources
may still use a thermal cycle for power production like conventional sources,
whereas others, such as solar PV systems and most modern wind turbines, are
interfaced with utility grids through inverters. This means that the inverter-based
generation is decoupled from the grid and its frequency excursions as it does not
provide any mechanical inertial response and, hence, compromises frequency
stability [9].
Imbalances between active power production and load consumption lead to
frequency variations. The frequency drops when the consumption rises rapidly, or a
large generating plant shuts down. Three main control actions are used for fre-
quency regulation: (i) primary control by which the governor response takes place
within the first few seconds (10–30 s); (ii) secondary control which is the automatic
generation control – its action takes place within minutes (10–30 min) and restores
the system frequency back to the nominal value; and (iii) tertiary control action is
the reserve deployment when actions are taken to get the resources in place to
counteract present or future disturbances in the system [10,11].
In the early stage, and before the primary control reacts to changes, the gen-
erators cannot respond instantaneously to balance the system. A physical response
called ‘inertial response’ comes from the conventional rotating generators where
the kinetic energy stored in the rotating mass is responsible for counteracting the
energy imbalance by slowing the rate of change of frequency (RoCoF) until the
primary control activates. In case of using renewable DGs, the inertial response
decreases, and thus, RoCoF is increased leading to a low-frequency nadir
(Figure 14.7).
Frequency
event
response
Secondary control Tertiary control
Inertial
Primary control
(automatic generation (long-term
(governor response)
control) reserve)
<10 s 10–30 s 10–30 min >30 min
Nominal frequency
System frequency (Hz)
High inertia
Low inertia
Reduction in
frequency nadir
Time
Seconds Minutes Hours
Figure 14.7 Frequency response and control actions
Therefore, in systems with lower inertia, the frequency nadir is lower and
RoCoF is getting higher. This can lead to under-frequency load shedding by fre-
quency relay tripping and may yield cascaded outages. Adding virtual inertia to the
system in a condition of operation in a very short time interval (<10 s) and in
autonomous fashion is the solution to such situations. Furthermore, the system
stability is improved and higher level of renewable DGs is permitted when
deploying virtual inertia system appropriately.
Various topologies used for virtual inertia implementation are presented in
literature [11–13]. They are designed to effectively emulate inertia of synchronous
generators through power electronic-based inverters. The concept of virtual inertia
topologies can be summarized as in the following.
Using the swing equation, the frequency variation in a power system after a
frequency event can be expressed as
2H dwg DP
¼ (14.8)
wg dt Sg
and
DP ¼ Pgen Pload (14.9)
where H, the inertia constant, is expressed as the kinetic energy normalized to
apparent power Sg of the connected generators in the system, wg is the system
angular speed (rad/s), Pgen is the generated power and Pload is the power demand
including losses.
In terms of frequency, f (Hz), (14.8) can be rewritten as

2H df DP
¼ (14.10)
f dt Sg
RoCoF of the system is mathematically given by df/dt, which is inversely
proportional to the system inertia. Thus, the system requires additional inertia as
more renewable DGs are integrated into the power system. The key idea of the
virtual inertia is depicted in Figure 14.8. A typical combination of inverter-based
DER units, solar PV system, wind turbines and ESS-like batteries is integrated into
a power grid. PV system and ESS have a DC–DC converter, whereas the wind
turbines are mostly variable speed wind turbines and have an AC–DC converter.
So, all units are interfaced through a DC–AC inverter in the front end and do not
contribute to the inertial response. The virtual inertia system receives voltage and
current feedback signals from the inverter output to generate gating signals by
applying an algorithm. The virtual inertia algorithm is the major part of the virtual
inertia emulation system. It presents the different DER units interfaced to the power
grid through electronic power conditioning systems as synchronous generators.
As an example, to illustrate the main components and the principle of opera-
tion of the virtual inertia system as a controller, one of the topologies called ‘Ise
lab’s topology’ is briefly discussed and its schematic diagram is shown in
Figure 14.9. This topology is a swing equation-based topology as it solves the
swing equation every control cycle to emulate inertia. The principle of operation
includes the following steps:
● The controller senses the inverter output current, i, and voltage, v, to compute
the grid frequency, wg, and the corresponding power, Pout by frequency/power
DC bus
Renewable DGs Converters Utility power grid
DC
AC bus
DC
Solar PV arrays Inverter
DC
Substation
AC
Wind AC
v i
turbines DC Gating
signals
Virtual inertia
DC control system
ESS
DC
Figure 14.8 Concept of virtual inertia

DC
AC
Gating i v
signals
PWM generator
ωg P0
v e
Pout
K Pin
ω* Control algorithm Power/frequency
1 + Td S ωg calculator
Governor model Virtual inertia control
Figure 14.9 Schematic diagram of the key elements of virtual inertia control and
governor model
calculation unit. These two parameters along with the prime mover input
power, Pin, are inputs to the control algorithm unit.
● The control algorithm unit solves the swing equation every control cycle pro-
viding the phase command, q, as well as the voltage reference, e, which is
generated through Q-v droop approach. The two parameters, q and e are inputs to
the PWM generator. The typical swing equation of a synchronous generator is

dwm
DP ¼ Pin Pout ¼ J wm þ Dp Dw (14.11)
dt
and
Dw ¼ wm wg (14.12)
where Pout is the inverter output power, wm is the virtual angular frequency, wg
is the grid/reference angular frequency, J is the moment of inertia and Dp is the
damping factor. Pin is the input power, which is like the prime mover input
power in a synchronous generator and can be computed by utilizing a governor
model. The governor is modelled, in a simplest way, as a first-order lag ele-
ment with gain and time constant of K and Td, respectively. The output of the
model, Pin , is based on the frequency deviation from a reference frequency w*
and a continuous power reference, P0 (Figure 14.9).
● The PWM generator sends the gating signals to the inverter.
14.5 Microgrids
Adding to the benefits of integrating the DER units into the power grid and coun-
teracting the challenges that can be met, the new trend is towards gathering the
DER units with the loads present in their vicinity in a separate gird called ‘micro-
grid’ (MG). Depending on the request of distribution system operator (DSO), the
MG can be connected to or disconnected from the utility distribution system
through a point of common coupling (PCC). Furthermore, the distribution system
may include more than one grid, some within the entire area covered by utility grid
and others located remotely and isolated from the utility grid (Figure 14.10).
14.5.1 MG concept
The MG is structured as a group of DERs and loads. It can be considered as a
subsystem which can be operated in three possible modes: grid-connected, islanded
and isolated. In islanded mode, the MG that is grid-connected disconnects from the
main grid, either in a planned fashion or due to fault/disturbance in the main grid. It
provides seamless transition between the two modes. The MG in the isolated mode
is designed in such a way that it is never connected to the main grid. This is the case
of remote sites (e.g. remote communities or remote industrial sites) where an
interconnection with the main grid is not feasible due to either technical and/or
economic constraints. Therefore, isolated MGs operate permanently in stand-alone
mode [14].
A typical MG system is shown in Figure 14.11. It comprises a variety of DERs
and different types of end users. Hence, it is considered as a portion of the dis-
tribution system, which is located downstream the distribution substation [15].
14.5.2 Operation functions

For each operation mode, MG provides some functions as follows: (i) grid-
connected mode, where the MG is connected to the upstream grid (MG can receive
energy, totally or partially, from the main grid depending on power sharing. On the
other hand, any excess power can be sent to the main grid when the total production
exceeds consumption); (ii) island mode, which occurs when the host grid has a
failure, or there are some planned actions (e.g., to perform maintenance actions).
(MG can smoothly move to islanded operation. Thus, the MG operates
Area covered by utility distribution system
DG
PCC DG
ESS
PCC
MG (con/discon)
DG
Load DG
DG
Distributed generation ESS DG
DG source ESS
ESS Energy storage

system MG (con/discon) MG (isolated)
Figure 14.10 Trends towards MGs

Solar PV
Storage
LC LC
LC LC LC
Load MGCC Main grid

Substation
LC PCC
LC
Diesel DG unit
LC LC
LC LC LC
CHP
Wind
Figure 14.11 Typical MG configuration
autonomously); (iii) ride-through between the two modes [15,16]. The operation of
MG whether in grid connected mode or islanded mode requires several reinforce-
ments regarding the control and protection issues.
14.5.3 Challenges facing MG implementation

In this section, the challenges countering the progress of MG deployment are pre-
sented. Although MGs have many benefits, expansion of their utilization involves
some critical issues, such as technical, regulatory and financial concerns. These
concerns are detailed next in this subsection [17,18].
Technical issues. Due to the electrical interface between multiple sources,
which may be based on conventional generators (e.g. diesel synchronous generator)
or inverter-based renewable energy (wind turbines, PV system, etc.), several
interconnection concerns have been raised. Stable operation of MG with voltage
quality at certain frequency limits is considered as a serious issue, practically with
fluctuating DG power. Moreover, the difficulty of the coordinated control may
arise in decoupling the active and reactive power during operation. In addition,
generating harmonic currents and voltage between many DGs may lead to con-
flicting control requirements. One of the most important challenges is the transition
ability of the MG from grid connected to islanded mode without losing its stability,
practically after fault provoked islanding incidents. This transition gets worse in the
presence of dynamic load (e.g. induction motors), especially with intermittent
renewable DG units. Further, reconnection with the host grid also poses challenges.
Hence, re-synchronization requires carful transition to prevent the voltage dis-
turbance. Other issues have been raised due to designing proper protection scheme
of MG. The protection scheme of MG should be able to meet protection
requirements such as sensitivity and reliability not only in grid connected mode, but
also in the islanded mode. More explanation for MG technical issues are presented
in Chapter 19.
Regulatory issues. There are many aspects of legislation that prohibit the
integration of MG with utility grid. For example, in Spain, regulations prevent
exciting any load in the same circuit between PV system and low-voltage system.
Moreover, the integration of generation and storage systems is prevented.
Unfortunately, such rules and regulations are applied in Egypt as well, which may
lead to constraining the MG concept when connecting MG to the utility grid. From
the grid operator point of view, this interconnection may be affected and disturb the
safety of the utility grid. Hence, the grid operators create guidelines and codes,
differing from one country to other, to manage this interconnection. Unfortunately,
most of these guidelines are not directly applied to the MG concept as in case of
anti-islanding and faulty regulations, which deal with MG as a potential source of
disturbance to the grid. The anti-islanding protection schemes constrain the MG’s
ability to seamlessly transition to the islanding mode and continue its function
locally, since the DGs are forced to disconnect.
Financial issues. Although most of the previous issues have been handled by
continuous development and research, financial issue is still the most common
obstacle for increasing penetration of MG. Some technologies have already become
cost-effective. However, many important technologies such as PV, fuel cells and
storage remain expensive and need some sort of financial support and legislation
from the interested countries.
14.5.4 Role of ESSs

ESS is considered as a vital part in MG infrastructure because it is responsible for
successful operation of the MG, especially with intermittent renewable energy sources.
It allows the DG units of the MG to work as dispatched units, with capability of fault
ride through during the variation of the primary energy sources (i.e. solar PV and
wind). Consequently, it improves the stability, reliability and power quality of the MG.
The benefits of storage in the latter application are of interest, because while renewable
energy resources are a pillar of the MG, without storage, their generation cannot
improve the system reliability and must be duplicated by other means of generation.
A storage unit can provide a functionality like that of the inertia of a syn-
chronous generator by absorbing temporary mismatches between power generation
and demand, especially in a low-inertia power electronic-based MG. Therefore,
system stabilization can be improved by providing voltage/frequency control in a
droop-based scheme [19].
Energy storage has applications in transmission capability improvement,
power quality enhancement, MG-islanded operation, active distribution systems
and electric vehicle technologies, and may improve dynamic stability, transient
stability, voltage support and frequency regulation [19,20]. The power grid can
substantially benefit in generation, transmission, distribution and consumption
from the availability of stored energy [21]. For example, storage can eliminate or
delay expansion of the transmission infrastructure or generation capacity. Storage

can be combined with non-dispatchable DER units such as wind and solar energy to
turn them into dispatchable units. On the consumers’ side, storage can be employed
for peak-shaving by storing the locally generated energy until it is needed. An
extensive list of applications for energy storage in transmission, distribution and
generation is presented in [22].
Despite its benefits, energy storage has not been fully utilized. Among the
limiting factors is, besides the cost, the lack of appropriate control and management
strategies [23]. Future research is needed to investigate and develop control
methodologies for the various forms of energy storage technologies.
14.5.5 Protection strategy

Some technical challenges need to be resolved by power system researchers and
engineers. Protection of MGs is one of them that requires more attention. During
the grid-connected operation mode, since the main grid provides a large short-
circuit current to the fault point, the protection can be performed by the existing
protective devices in the distribution networks, but in the islanded mode, fault
currents provided by DG units installed in the MG are significantly less than those
in the grid-connected mode. Hence, the employment of traditional overcurrent-
based protective devices in MG and sub-grids is no longer valid. Most conventional
distribution systems operate radially, where the power flows uni-directionally from
large power plants to the customers. In such systems, as the magnitude of short-
circuit currents is proportional to the fault location, the protection is done by
overcurrent-based protective devices [24]. The time-graded coordination between
them enables upstream devices to operate as backup for downstream ones [25].
In recent years, the emergence of MG has changed the structure of distribution
systems from passive networks into active ones. This change has disturbed the
operation of overcurrent-based strategies such that they can no longer protect new
structures [26,27]. The fault current contribution of inverter-based DG sources in
an MG is limited (only two to three times the maximum load current) due to the
low thermal capability of their power electronic devices. Therefore, protective
devices in an MG containing inverter-based DG sources would operate very slowly
or may not be triggered at all for fault events during islanded mode. In addition, the
considerable difference between the magnitude of short-circuit currents in the grid-
connected and islanded modes makes single-setting traditional overcurrent relays
unable to protect dual-mode operating MG [28,29]. Therefore, overcurrent pro-
tective devices cannot be used for protection of MGs and sub-grids including
inverter-based DG sources, and some new techniques should be devised.
The traditional overcurrent-based strategies do not have the ability to protect
AC MGs and sub-grids due to the drastic difference between the magnitude of fault
currents in the grid connected and islanded modes. The main protection strategies
are briefly described as follows [30]:
● The DG is automatically taken out of service following a fault on the distribution
network, and thus, protection of the distribution network is not affected.
● The capacity of DG is limited, and the distribution network remains unchanged.

● Measures for limiting the fault current are taken; for example, a fault current
limiter is used to minimize the effects of DG after a fault occurs, and the
distribution network remains unchanged.
An MG should be able to operate in parallel with the grid in islanded mode. It
shall meet the following basic requirements:
● In grid-connected operation, if a fault occurs in the MG, the MG protection
should operate reliably to clear the fault. For example, when the electrical
equipment in the LV distribution network fails, the protection of the distribu-
tion network should operate to isolate the equipment to ensure secure and
stable operation of the MG.
● When an instantaneous fault occurs in the distribution network, the protection
of the distribution network should operate quickly to clear the fault to maintain
uninterrupted operation of the MG.
● In case of power loss of the distribution network, the islanding protection of
the MG operates to disconnect the MG from the distribution network and the
MG switches to islanded operation.
● In islanded operation, when a fault occurs on the MG, the protection should
operate reliably to clear the fault to ensure secure and stable operation of the MG.
● Upon recovery of the distribution network, the MG is reconnected to the grid.
Conventional protection schemes face serious challenges when it comes to
protecting an islanded MG with inverter interfaced DG (IIDG) units. They need
major revision to find new methods based on the limited fault current to detect and
isolate the faulty portion. The various possible solutions to cope with the problem
can be broadly divided into four categories: use of inverters having high fault
current capability, that is, uprating of the inverter [31]; communication between the
inverter and protective relays; introduction of energy storage devices that are cap-
able of supplying large current in case of a fault [32]; and in-depth analysis of the
fault behaviour of an islanded MG with an IIDG unit to comprehend the behaviour
of system voltages and currents [33,34]. This will in turn help in defining alter-
native fault detection and alternative protection strategies that, in case of a fault, do
not rely on a large magnitude of the fault current but rely on other parameters, like
change in the voltage of the system [35].
14.5.6 Control of MGs

Generally, there are two levels of MG control system, central control (supervisory
control) and local control for each DG or load. Both are discussed in the following
[32,36].
Microgrid central control (MGCC). In the grid-connected mode, the MG
voltage is influenced by the host utility grid. Thus, the hosted DGs cannot play any
appreciable role in determining the MG voltage magnitude or frequency (although
they can locally affect the voltage magnitude through their reactive power or, to a
lesser extent, real power contributions). Hence, the function of the central control in
grid-connected mode is to handle the active- and reactive-power commands for the
DG systems. The central control can calculate the commands based on a variety of
criteria, such as market signals, economy of the MG, optimal operation, and host
grid conditions and requirements. For example, during hours when the grid elec-
tricity is cheap, the overall control may decide to reduce the power output of the gas
microturbine DG and charge the energy storage devices mainly through the grid
power. At the same time, in response to a grid command, it may dispatch the DG
systems in such a way that the MG draws reactive power from the host grid to
prevent an overvoltage at the PCC.
By contrast, in the islanded mode of operation, the DGs are mainly controlled
locally to regulate the MG voltage magnitude and frequency. This process must be
fast and reliable and is a function of the DG control. In the absence of a connection
to the utility grid, a sustained islanded mode operation also implies that the sum of
DG system power outputs equals the aggregate load power. Although, this is
strictly and rapidly ensured by the DG control, the central control can specify the
steady-state DG system output powers, according to the constraint mentioned
above, in such a way that the MG steady-state operation is optimized. Further, the
central control can shed loads in the islanded mode, depending on load criticalities,
MG energy reserves or other considerations [18,37,38]. For safe reconnection of
the MG to the host grid, once the operating mode is to be switched to the grid-
connected mode, the central control can also specify the commands for steady-state
voltage magnitude and frequency of the DG system if they have drifted away from
their nominal values due to the islanded-mode operation.
The central control requires a communications network to exchange informa-
tion with the DG systems and loads. Even though the communication links may not
work fast or be very reliable, their design should be such that their malfunction
shall not lead to a system collapse. The malfunction should only lead to the non-
optimal operation of the DGs [39].
Local control (LC). The main function of LC is to ensure the stability and
robustness of the MG, in transient as well as steady states. Thus, LC ensures that
● DG systems operate in synchronism with the grid in the grid-connected mode of
operation. In the islanded mode of operation, the aggregate of the DG system
power outputs tracks the aggregated load power. Moreover, some dynamics
such as the MG voltage magnitude and frequency are regulated when the DG
systems properly share the total load, both in transient and steady-state regimes.
● DG systems operate within their limits and are protected against network faults
irrespective of the operating mode.
LC also receives and enforces the commands that are issued by the MGCC. In
contrast to MGCC, LC must be fast and very reliable. Otherwise, either the MG vol-
tage magnitude and frequency undergo unacceptable excursions with respect to their
nominal values, or the MG collapses completely [39,40]. Therefore, its centralized
realization is theoretically feasible. The consensus is that the LC should be imple-
mented based on decentralized methods. The most notable decentralized method is
based on droop characteristics. One distinct merit of the droop-based control method is
that it grants plug-and-play capability to the DG systems. The DG control may also be
applied to the loads, for rapid and continuous power control, or even for load shedding.
14.5.7 Islanding detection

MGs can be viewed as small scale of the classical large centralized electric dis-
tribution system and include DERs. They have the capability of switching from one
operation mode (grid-connected) to another (islanded). Loads of MG should not be
affected when changing the mode of operation.
The interconnection of MGs and/or DERs to the utility grid is controlled by
primary (local), secondary (regional) and tertiary (national) controls to be regulated
in compliance with the published standards (e.g. IEEE-1547, IEC-62115 and AS-
4777). However, it is necessary for each DER (renewable, non-renewable and
energy storage) unit to have the ability to detect if it is connected or disconnected
from the utility grid, which is normally referred to as islanding detection. Islanding
state can be caused by faults, accidental opening of the main switch that connects
the MG/DER and the main grid, human errors or natural causes. Islanding should be
accurately detected (i.e. the non-detected zone ‘NDZ’ is as zero as possible) within a
time as small as possible (e.g. IEEE-929 std. requires islanding detection within 2 s
after the islanding state occurs). Generally, the faster the islanding is detected, the
faster the control algorithms of all MG-DERs can be adapted to effectively achieve
the desired performance, especially for MG stability and power quality.
14.6 Concluding remarks

● The power system of the future will become more complex as it may comprise
different DG units and/or ESSs connected in different locations (integrated into
transmission or distribution system) as well as several MGs downstream the
distribution system.
● Several technical problems arise because of the integration of DER units,
which have different characteristics.
● Cost-benefit analysis should be carried out to decide which type of DG and
ESS can be added to the power system.
● Information technology, computer monitoring and control technologies, cyber
security techniques, communication networks and advance protective relaying
system are essentially needed to get the power system smarter. More details
about such grids are explained in Chapter 1, Section 1.4.
References
[1] National Renewable Energy Laboratory, DOE National Laboratory.

Distributed energy resources: A how to guide for federal facility managers
[online]. 2002. Available from https://www.nrel.gov/docs/fy02osti/31570.
pdf [Accessed 12 Feb 2020].
[2] Adefarati T. and Bansal R. C. ‘Integration of renewable distributed gen-

erators into the distribution system: A review’. IET Renewable Power
Generation. 2016;10(7):873–84.
[3] Burger S. P., Jenkins J. D., Huntington S. C. and Pérez-Arriaga I. J. ‘Why
distributed? A critical review of the tradeoffs between centralized and decen-
tralized resources’. IEEE Power & Energy Magazine. 2019;17(2):16–24.
[4] Strbac G., Pudjianto D., Aunedi M., Papadaskalopoulos D., Djapic P.,
Ye Y. et al. ‘Cost-effective decarbonization in a decentralized market’. IEEE
Power & Energy Magazine. 2019;17(2):25–36.
[5] Hoke A., Giraldez J., Palmintier B., Ifuku E., Asano M., Ueda R., et al.
‘Setting the smart solar standard’. IEEE Power & Energy Magazine.
2018;16(6):18–29.
[6] Vartanian C., Bauer R., Casey L., Loutan C., Narang D. and Patel V.
‘Ensuring system reliability’. IEEE Power & Energy Magazine. 2018;
16(6):52–63.
[7] Barkans J. and Zalostiba D. ‘Blackout prevention and power system self-
restoration’. EUROCON, 2007. The Int. Conf. on ‘Computer as a Tool’;
Warsaw, Poland, Sep 2007. IEEE; 2007, pp. 1–8.
[8] Braun M., Brombach J., Hachmann C., Lafferte D., Klingmann A.,
Heckmann W., et al. ‘The future of power system restoration’. IEEE Power
& Energy Magazine. 2018;16(6):30–41.
[9] Taczi I. and Vokony I. ‘Possible role and effects of synthetic inertia on the
future’s power system’. Trivent Publishing, Engineering and Industry Series:
Power Systems, Markets and Renewable Energy Sources in South-Eastern
Europe. 2016;3:32–47.
[10] Eriksson R., Modig N. and Elkington K. ‘Synthetic inertia versus fast fre-
quency response: A definition’. IET Renewable Power Generation, Open
Access. 2018;12(5):507–14.
[11] Tamrakar U., Shrestha D., Maharjan M. and Bhattarai B. P. ‘Virtual inertia:
Current trends and future directions’. MDPI, Applied Sciences. 2017;7(7):
1–29.
[12] Waffenschmidt E. and Hui R. S. Y. ‘Virtual inertia with PV inverters using
DC-link capacitors’. IEEE, 18th European Conference on Power Electronics
and Applications; Karisruhe, Germany, Sep 2016, pp. 1–10.
[13] Gonzalez-Longatt F., Chikuni E. and Rashayi E. ‘Effects of the inertia from
wind power on the total system inertia after frequency disturbance’. IEEE
International Conference on Industrial Technology (ICIT); Cape Town,
South Africa, Feb 2013, pp. 826–32.
[14] Olivares D. E., Mehrizi-Sani A., Etemadi A. H., et al. ‘Trends in microgrid
control’. IEEE Transactions on Smart Grid. 2014;5:1905–19.
[15] Mariam L., Basu M. and Conlon M. F. ‘Microgrid: Architecture, policy and
future trends’. Renewable and Sustainable Energy Reviews. 2016;64:477–89.
[16] Lopes J. P., Hatziargyriou N., Mutale J., Djapic P. and Jenkins N.
‘Integrating distributed generation into electric power systems: A review of
drivers, challenges and opportunities’. Electric Power Systems Research.

2007;77:1189–203.
[17] Soshinskaya M., Crijns-Graus W. H., Guerrero J. M. and Vasquez J. C.
‘Microgrids: Experiences, barriers and success factors’. Renewable and
Sustainable Energy Reviews. 2014;40:659–72.
[18] Planas E., Gil-de-Muro A., Andreu J., Kortabarria I. and de Alegrı́a I. M.
‘General aspects, hierarchical controls and droop methods in microgrids: A
review’. Renewable and Sustainable Energy Reviews. 2013;17:147–59.
[19] Nourai A. and Kearns D. ‘Batteries included: Realizing smart grid goals with
intelligent energy storage’. IEEE Power Energy Magazine. 2010;8(2):49–54.
[20] Nourai A., Kogan V. I. and Schafer C. M. ‘Load leveling reduces T&D line
losses’. IEEE Trans. on Power Delivery. 2008;23(4):2168–73.
[21] Thomas R. J. ‘Putting an action plan in place: The Electricity Advisory
Committee’s energy storage report’. IEEE Power & Energy Magazine.
2009;7(4):26–31.
[22] The Electricity Advisory Committee. Bottling electricity: Storage as a
strategic tool for managing variability and capacity concerns in the modern
grid. EAC report, 2008.
[23] Hajimiragha A. and Zadeh M. R. D. ‘Practical aspects of storage modeling in
the framework of microgrid real-time optimal control’. Proceedings of IET
Conf. on Renewable Power Generat. (RPG); Edinburgh, UK, Sep 2011,
pp. 93–8.
[24] Piesciorovsky E. C. and Schulz N. N. ‘Fuse relay adaptive overcurrent pro-
tection scheme for microgrid with distributed generators’. IET Generation,
Transmission and Distribution 2017;11(2):540–9.
[25] Liu Z., Hoidalen H. K. and Saha M. M. ‘An intelligent coordinated protec-
tion and control strategy for distribution network with wind generation
integration’. CSEE Journal of Power Energy Systems. 2016;2(4):23–30.
[26] Sahoo A. K. ‘Protection of microgrid through coordinated directional over-
current relays’. Proceedings of the 2014 IEEE Global Humanitarian
Technology Conference South Asia Satellite (GHTC-SAS); Trivandrum,
India, Sep 2014. IEEE; 2014, pp. 129–34.
[27] Choudhary N. K., Mohanty S. R. and Singh R. K. ‘A review on microgrid
protection’. Proceedings of the 2014 International Electrical Engineering
Congress (IEECON); Chonburi, Thailand, 2014, pp. 1–4.
[28] Bui D. M., Chen S. L., Lien K. Y. and Jiang J. L. ‘Fault protection solutions
appropriately used for ungrounded low-voltage AC microgrids’.
Proceedings 2015 IEEE Innov. Smart Grid Technologies – Asia, ISGT ASIA
2015; Bangkok, Thailand, Nov 2015. IEEE; 2015, pp. 4066–71.
[29] Hussain A., Aslam M. and Arif S. M. ‘N-version programming-based pro-
tection scheme for microgrids: A multi-agent system based approach’.
Sustainable Energy, Grids and Networks. 2016;6:35–45.
[30] Fusheng L., Ruisheng L. and Fengquan Z. Microgrid technology and engi-
neering application. London: Academic Press; 2016.
[31] Lopes J. P., Moreira C. L. and Madureira A. G. ‘Defining control strategies

for microgrids Islanded operation’. IEEE Transactions on Power Systems.
2006;21(2):916–24.
[32] Jayawarna N., Jenkins N., Barnes M., Lorentzou M., Papthanassiou S.,
Hatziagyriou N. ‘Safety analysis of a microgrid’. Proceedings of 2005
International Conference on Future Power Systems; Amsterdam, Netherlands,
Nov 2005. IEEE; 2005, pp. 1–7.
[33] Timbus A. V., Rodriguez P., Teodorescu R., Liserre M. and Blaabjerg F.
‘Control strategies for distributed power generation systems operating on
faulty grid’. Proceedings of IEEE International Symposium on Industrial
Electronics; Montreal, Canada, July 2006. IEEE; 2006, 2: pp. 1601–7.
[34] Baran M. E. and El-Markaby I. ‘Fault analysis on distribution feeders with
distributed generators’. IEEE Transactions on Power Systems. 2005;
20(4);1757–64.
[35] Tumilty R. M., Brucoli M., Burt G. M. and Green T. C. ‘Approaches to
network protection for inverter dominated electrical distribution systems’.
Proceedings of 3rd IET International Conference on Power Electronics,
Machines and Derives (PEMD 2006); Dublin, Ireland, April 2006. IET;
2006, pp. 622–6.
[36] Dimeas A. L. and Hatziargyriou N. D. ‘Operation of a multiagent system
for microgrid control’. IEEE Transactions on Power Systems. 2005;
20(3):1447–55.
[37] Huang W., Lu M. and Zhang L. ‘Survey on microgrid control strategies’.
Energy Procedia. 2011;12:206–12.
[38] Bouzid A. M., Guerrero J. M., Cheriti A., Bouhamida M., Sicard P. and
Benghanem M. ‘A survey on control of electric power distributed generation
systems for microgrid applications’. Renewable and Sustainable Energy
Reviews. 2015;44:751–66.
[39] Vandoorn T., De Kooning J., Meersman B. and Vandevelde L. ‘Review of
primary control strategies for islanded microgrids with power-electronic
interfaces’. Renewable and Sustainable Energy Reviews. 2013;19:613–28.
[40] Green T. and Prodanović M. ‘Control of inverter-based micro-grids’.
Electric Power Systems Research. 2007;77:1204–13.
Chapter 15
Economic dispatch of hybrid thermal
and wind plants
15.1 Description of ED problem

Economic dispatch (ED) of power system could be defined as the process of allo-
cating specific generating units and the generation levels to the committed units in
the mix with the primary objective of minimizing fuel cost or emissions or both
while sustaining system constraints. This means that the total load demand plus
system losses are satisfied at the minimum cost while satisfying different technical
constraints of the network and power sources.
Prior to the widespread use of alternate sources of energy, the ED problem
involved only conventional thermal energy power generators which used
exploitable resources such as fossil fuels. It has now become apparent that there is a
need for alternatives to thermal energy power generation, and renewable energy
sources, such as wind and PV, are now seeing more widespread use. One of the
major benefits of renewable energy sources is that, after the initial cost of land,
equipment and installation, there is essentially no cost involved in the production of
power. Furthermore, environmentally their impact is usually considered to be
friendlier than that of thermal energy sources.
The high uncertainty concomitant with the power output of various renewable
distributed generations (DGs), for example, wind and solar, is considered as one of
the main challenges in the operation of the smart grids. This high uncertainty cre-
ates a major task for system operators in keeping the generation–load balance,
particularly in regions with high renewable DG penetration [1,2].
Integrating uncertain power sources such as wind power with the conventional
thermal units requires specific arrangements, so stochastic ED (SED) is considered.
SED is done by characterizing the uncertainty of wind power in the ED problem.
The primary problem associated with the SED model is the fact that wind speed,
the power source for the wind energy system, forecast at a certain time in future is
not exact. So, probability density function (pdf) of the actual wind power output
should be obtained.
Probability density function is a statistical expression that defines a probability
distribution for a continuous random variable as opposed to a discrete random
variable (Appendix A). When pdf is graphically portrayed, the area under the curve
will indicate the interval in which the variable will fall. The total area in this
interval of the graph equals the probability of a continuous random variable

occurring.
The uncertainty issue has made the researchers assess various aspects of the
problem based on stochastic principles [2–5]. The uncertainty of electrical energy
cost coefficients and load demand has been considered in the expected objective
functions of ED in the literature [6,7]. A proposed stochastic long-term security-
constrained unit commitment formulation for representing uncertainties in the
availability of generating units and transmission lines, and inaccuracies in load
forecasting has been investigated [8]. For a practical problem, like ED, modified
intelligent methods should be adequate for solving the ED with more accurate
multiple fuel cost functions and constraints.
15.2 Modelling of system comprising thermal units

and wind energy sources
ED of a power system comprising thermal and wind energy sources is considered.
Precise calculation of ED is based on accurate modelling of power system energy
sources, e.g. thermal units and wind. This is a great challenge and more attention
must be paid to it because of the stochastic nature of wind energy sources.
Therefore, SED is becoming more imperative along with the increased penetration
of wind energy sources. Modelling of thermal units and wind energy sources from
ED’s perspective is described in forthcoming subsections. The modelling is com-
posed of mathematical relations established from the physical behaviour of the
system’s energy sources. Consequently, the uncertainty associated with wind
power is considered in the stochastic dispatch. A pdf of the wind power output is
determined.
According to probability and statistics, a variable whose value is subject to
disparities owing to chance (i.e. randomness, in a mathematical sense) is called a
random variable or stochastic variable. As opposed to other mathematical variables,
a random variable theoretically does not have a single, fixed value (even if
unknown). Somewhat, it can take on a set of possible dissimilar values, each with
an associated probability [9]. Then, wind speed is considered a continuous random
variable which is affected by a number of factors such as geometric shape,
roughness and elevation of the ground surface. Solar radiation (also called clear-
ness index) is similarly considered as a continuous random variable [10].
15.2.1 Wind power modelling using proper pdf

The wind speed distribution primarily determines the performance and the feasi-
bility of wind power systems. Wind energy can be easily defined when the wind
speed distribution is identified. Modelling wind speed using a proper pdf provides a
few key parameters, which can illuminate the characteristics of a wide range of
wind speed data. Therefore, it is most desirable to use this modelling in long-term
planning problems.
Economic dispatch of hybrid thermal and wind plants 375
A pdf is used to model the wind velocity variation. It provides the probability
that an event will occur between two end points. For any two speeds larger than
zero, the area under the curve between those two speeds will provide the prob-
ability that wind will blow anywhere between those two speeds. It is important to
understand that the actual height and shape of a pdf curve are determined such that
the area under the curve from 0 to infinity is exactly 1. Physically, this means that
there is a 100% chance that the wind will blow at some speed in m/s between 0 and
infinite.
Many scientific studies in literature concerning wind energy have presented
the application of a variety of pdf (e.g. normal, lognormal, gamma, Rayleigh and
Weibull) to describe wind speed distribution [11,12]. These studies concluded that
the Weibull distribution with two parameters might be effectively applied to define
the principle wind speed variation. The Weibull probability density given by (15.1)
has been found to be a very precise model to define wind velocity variation. This
distribution is regularly used in wind energy engineering [13]. Its success originates
from the two adaptable parameters, c and k, which can offer great flexibility in
fitting the distribution function to the measured values with different performances
as shown in Figure 15.1. It is noted that the greater the value of k, the more the
curve looks like a reversed bell, though as the value of c increases, the curve
spreads out even more.

k v ðk1Þ ðv=cÞk
f v ðv Þ ¼ e (15.1)
c c
This expression is valid for k > 1, v 0 and c > 0. k is the shape factor and c is
the scale parameter.
Different methods can be used to calculate the Weibull parameters [13,14].
The shape factor, k, will typically range from 1 to 3. For given average wind speed,
a lower shape factor indicates a relatively wide distribution of wind speed around
0.12
k=1
k=2
0.10 k=3
0.08
fv(v)
0.06
0.04
0.02
0
0 5 10 15 20 25
m/s
Figure 15.1 Weibull probability distribution function with c ¼ 10 and k ¼ 1,

2 and 3
the average while a higher shape factor specifies a comparatively narrow dis-
tribution of wind speed around the average. A lower shape factor will normally lead
to higher energy production for a given average wind speed. Figure 15.1 shows the
Weibull pdf with c ¼ 10 and k ¼ 1, 2 and 3. Here, the parameter k is calculated,
approximately, using the mean wind speed, vm , and the standard deviation, s, as
follows:
1:086
s
k¼ (15.2)
vm
vm
c¼ (15.3)
Гð1 þ 1=k Þ
!
1 X n
vm ¼ vi (15.4)
n i¼1
" #0:5
1 X n
s¼ ðv i v m Þ2 (15.5)
n 1 i¼1
where n is the number of non-zero wind speeds.

A Weibull pdf with k ¼ 2 and c ¼ 6, 8, 10 and 12 is shown in Figure 15.2.
At k equal to 2, pdf is named Rayleigh pdf and is given by (15.6). In this case,
pdf imitates most wind speed profiles. In case the mean value of the wind speed for
a site is identified, the scale index, c, can be considered as in (15.7).
v h v i
f ðvÞ ¼ 2 exp (15.6)
c 2c2
c ¼ vm =1:253 (15.7)
Weibull and Rayleigh’s functions are generally used to define the random
behaviour of the wind speed in a specific location over a specific period.
0.25
c=6
c=8
0.20 c = 10
c = 12
0.15
fv(v)
0.10
0.05
0
0 5 10 15 20 25
m/s
Figure 15.2 Weibull probability distribution function with k ¼ 2 and c ¼ 6, 8,

10 and 12
Furthermore, these two functions can define wind speed distribution for any
archetypal hour of the year. The wind power output of wind turbine (WT) generator
has a constant zero value lower than the cut-in wind speed and beyond the cut-out
wind speed, and it is constant between rated wind speed and cut-out wind speed.
The wind power output is a mixed random variable, which is continuous between
values of zero and rated power and is discrete at values of zero and rated power
output. If it is supposed that the wind speed has a given distribution such as
Weibull, it is then essential to change that distribution to a wind power distribution.
This transformation may be accomplished in the following manner, where v and w
are the wind speed and wind power random variable, respectively, for a linear
transformation as follows [15]:
k–1 k
klvi 1 + ρl vi 1 + ρl vi (15.8)
c c exp – c for 0 <w <wr
v k vo k
fw(w) = 1–exp – ci +exp – c for w = 0
(15.9)
v k vo k
exp – cr –exp – c for w = wr
(15.10)
where r ¼ wwr , l ¼ vr vi vi , vr is rated wind speed, wr is rated wind power, vi is cut-in
speed and vo is cut-out speed.
15.3 Formulation of optimization problem

ED provides important functions, like unit commitment, load forecasting, secur-
ity analysis, scheduling of fuel purchase and such, in modern power systems. A
bibliographical survey of ED methods reveals that various numerical optimiza-
tion techniques have been employed to solve the ED problem. Traditionally, it is
solved using mathematical programming based on optimization techniques such
as lambda iteration, gradient method and so on [16–18]. ED with piecewise linear
cost functions is a highly heuristic, approximate and extremely fast form of
ED [19].
Lambda iteration, gradient method [20–22] can solve simple ED calculations.
Although they are fast, they are not enough for applications in the deregulated
market. Several heuristic methods, among them genetic algorithm, have been
applied to solve the real-time economic load dispatch problem [23,24]. An artificial
neural network is also used to solve the optimization problem [16]. Other research
papers have applied the swarm behaviour to the problem of optimum dispatch as
well as the unit commitment problem [25].
Heuristic algorithms are iterative techniques that can search not only locally
optimal solutions but also a globally optimal solution depending on problem
domain and execution time limit. There are general-purpose searching methods
based on principles inspired by the genetic and fruition mechanisms detected in
natural systems and populations of living creatures. These approaches have the
benefit of searching the solution space more methodically. The chief trouble is their
sensitivity to the selection of parameters. Amongst such techniques, particle swarm
optimization (PSO) is simple and promising. It necessitates less computation time
and memory. It has also standard values for its parameters. A flow chart illustrating
the main steps of PSO algorithm is depicted in Figure 15.3.
Start
Define objective function

to be minimized
Initialize the parameters of PSO,

c1, c2, itermax, w, error
Randomly allocate active power to the n

units satisfying the equality, inequality constraints
Iteration iter = 0
Calculate the objective function
Update the gbest and pbest values
Update the position and velocity of the particles
Iteration iter = iter + 1
YES
Check the stopping criteria
according to error limit
NO
gbest of PSO is the solution

of economic load dispatch problem
Figure 15.3 Main steps of PSO algorithm

15.4 Single-objective ED model

ED of a power system with thermal and wind units is considered. First, conven-
tional dispatch with only three non-convex thermal units is presented and second,
deterministic dispatch incorporating wind power is modelled as described in the
next two subsections. The objective is to minimize the combined operating cost
without considering imbalance cost due to the mismatch between the actual and
scheduled power output of wind and thermal units. Then, the uncertainty associated
with the wind power is considered in the stochastic dispatch and the pdf of the wind
power output is obtained.
15.4.1 Conventional ED with thermal units

In single-objective conventional ED (CED), the objective is to minimize the
combined operating cost of the non-convex thermal units. It can be formulated as
X
N
Minimize OC Pgi (15.11)
i¼1
Subject to
PgiðminÞ Pgi PgiðmaxÞ (15.12)
X
N
Pgi ¼ L (15.13)
i¼1
where
OC(Pgi): operating cost of non-convex thermal units.
N: number of non-convex thermal units.
Pgi: power of the ith generating thermal unit.
L: system load.
PgiðminÞ : minimum generation limit of the non-convex thermal unit.
PgiðmaxÞ : maximum generation limit of the non-convex thermal unit.
Taking a closer look at the objective function, to effectively solve the ED
problem, most algorithms require the incremental cost curves to be of mono-
tonically smooth increasing nature. The generating units with the multi-valve steam
turbines exhibit a greater variation by the fuel cost functions, where the valve point
results in the rippling form of the heat-rate curve and the cost function contains
higher-order non-linearity due to the valve-point effects. The more general fuel cost
function of each thermal generator, considering the valve point effect, is expressed
as the sum of a quadratic and a sinusoidal function. The total non-smooth fuel cost
function in terms of real power output can be expressed as
X N
OC Pgi ¼ Ci Pgi (15.14)
i¼1

Ci Pgi ¼ ai þ bi Pgi þ ci P2gi þ ei sin fi Pi min Pi (15.15)
where ai ; bi and ci are cost coefficients of ith generator, and ei, and fi are the coef-
ficients of that generator reflecting valve-point effects [26].
15.4.2 Deterministic ED incorporating wind power

In a single-objective deterministic ED (DED), the objective is to minimize the
combined operating cost of the non-convex thermal unit and wind plant without
considering the uncertainty of wind power, i.e. considering the cost coefficient of
wind power as constant values. Thus, its formulation can be written as
" #
XN X M
Minimize OC Pgi þ OC d ðWi Þ (15.16)
i¼1 i¼1
Subject to
0 Wi Wri (15.18)
X
N X
M
Pgi þ Wi ¼ L (15.19)
i¼1 i¼1
where
OCd(Wi): deterministic operating cost of wind farms.
Wri: rated wind power from the ith wind-powered generator.
M: number of wind-powered generators.
S: number of solar-powered modules.
Wi: scheduled wind power from the ith wind power generator.
The deterministic operating cost of wind power generators is formulated as
OC d ðWi Þ ¼ Cdwi ðWi Þ (15.20)
Cdwi ðWi Þ ¼ di (15.21)
where di is the cost coefficient for ith wind power plant.
15.4.3 SED incorporating wind power

Owing to the uncertainty in wind power output and imbalance cost due to the
mismatch between the actual and scheduled power outputs of wind units, SED is
suggested. The objective of SED is to minimize the stochastic operating cost sub-
ject to the system constraints. Weibull pdf is utilized to represent the stochastic
nature of wind power. Single-objective SED for the proposed system with wind
power and thermal units can be formulated as follows [27]:

Minimize OC Pgi ; Wi
that is,
" #
X
N X
M
Minimize OC Pgi þ OC s ðWi Þ (15.22)
i¼1 i¼1
Subject to
0 Wi Wri (15.24)
X
N X
M
Pgi þ Wi ¼ L (15.25)
i¼1 i¼1
where
OC s ðWi Þ: stochastic operating cost of wind farms.
X
M X
M XM
OC s ðWi Þ ¼ Cwi ðwi Þ þ Cp;wi Wi;av wi þ Cr;wi wi Wi;av
i¼1 i¼1 i¼1
(15.26)
On the RHS of (15.26), the first term, Cwi ðwi Þ, is the cost function based on
wind speed profile and formulated as in (15.27). The cost coefficient is multiplied
by Weibull pdf of wind power, f w ðwÞ, as in (15.8)–(15.10),
Cwi ðwi Þ ¼ di fw ðwÞ Wi (15.27)
This cost coefficient is considered in DED as a constant value, whereas in SED

it is weighted by Weibull pdf of wind speed. The second term in (15.26) is the
penalty cost for not using all the available wind power. It is linearly related to the
difference between available or actual wind power and scheduled wind power, and
can be written as
ð wri

Cp;wi Wi;av wi ¼ Kp;wi ðw wi Þfw ðwÞdw (15.28)
wi
where
K p;wi : the penalty cost coefficient of underestimation (over-generation) for the
ith wind generator.
W i;av : actual or available wind power from the ith wind-powered generator.
The third term is the reserve requirement cost, which is because the actual or
available wind power is less than the scheduled wind power, and can be given by
ð wi

Cr;wi wi Wi;av ¼ Kr;wi ðwi wÞfw ðwÞdw (15.29)
0
where K r;wi is the reserve cost coefficient of overestimation (under generation) for
the ith wind-powered generator.
15.5 Economic/emissions dispatch problem

Concentration of greenhouse gases (GHGs) such as CO2, SO and NO2 in the
atmosphere has been increasing. Due to concerns regarding global warming and air
pollution, there has been an international movement promoting renewable technol-
ogies for electricity generation and the development of national emissions limits.
Several directives controlling emissions are currently in place that have an impact
on the electricity industry. For instance, the Kyoto Protocol, which is an agreement
under which industrialized countries agreed to reduce their collective emissions of
GHGs in 2012 by 5.2% compared with that in 1990. In the Copenhagen Accord, the
convention agreed to reduce global emissions to maintain the increase in global
temperature below 2 C. The use of renewable energies can mitigate the GHG
effects to meet the strict requirements stated in the Kyoto Protocol [28].
Wind power generation plays an essential role in a grid. Several techniques to
handle the environmental ED (EED) problem have been reported [29]. There are three
approaches to solve EED problem. (i) The emission is considered as a constraint with
a permissible limit [30]. This formulation, however, has severe difficulty in getting the
trade-off relations between cost and emission. (ii) Treat the emission as another
objective in addition to the usual cost objective [30]. However, the EED problem is
converted to a single-objective problem either by a linear combination of both
objectives or by considering one objective at a time for optimization. (iii) Handle both
fuel cost and emission simultaneously as competing objectives. Stochastic search and
fuzzy-based multi-objective optimization techniques have been proposed for the EED
problem [31]. However, the algorithms do not provide a systematic framework for
directing the search towards Pareto-optimal front and the extension of these techni-
ques to include more objectives is a very involved question. In addition, these tech-
niques are computationally involved and time-consuming.
Unlike genetic algorithm and other heuristic techniques, PSO has a flexible
and well-balanced mechanism to enhance and adapt the global and local explora-
tion abilities. It usually results in faster convergence rates than the genetic algo-
rithm [32]. In recent years, PSO has been successfully implemented for different
power system optimization problems including the economic power dispatch pro-
blem with impressive success. The potential of PSO to handle non-smooth and non-
convex economic power dispatch problem has been demonstrated and reported in
[33]. However, the problem has been formulated as a conventional dispatch pro-
blem with the objective of minimizing only the fuel cost.
15.5.1 Multi-objective stochastic dispatch

Stochastic dispatch can be extended to include multiple objectives with a goal of
optimizing two conflicting functions simultaneously, the operating cost of thermal,
wind units and emissions caused by the thermal units.
However, some of the approaches that can be used for designing multi-
objective PSO (MO-PSO) algorithms consider each objective function separately,
whereas others convert the multi-objective functions into a single-objective func-
tion using weighted aggregation (WA) method [34]. Each of the objectives is
assigned a weight based on their importance level to solve the problem using PSO
as a single-objective function. Thus, the objective function is formulated by sum-
ming weighted terms, where each term represents an objective.
The key idea of WA method is that systematically changing the weights during
evolution will lead the population to the Pareto front. Two possible methods are
investigated. One method is to distribute the weights randomly within a population.
The other method is to periodically change the weights with the process of evolu-
tion. The two methods are briefly investigated in the following [35]:
1. Random distribution of weights within a population. For the sake of clarity,
two objective problems in the current discussion are considered. Their exten-
sion to problems with more than two objectives is straightforward. For a con-
ventional aggregation method, the fitness function is the weighted sum of the
two different objectives, f1 and f2, as
Fitness ¼ w1 f1 þ w2 f2 (15.30)
where w1 and w2 are two constants determined using prior knowledge about the
problem. Using a pair of fixed weights, only one Pareto solution can be
obtained.
The algorithm will run many times with a weight combination used for each
time. In this way, all Pareto solutions that the Pareto front consists of can be
obtained. If the different weight combinations can be distributed among the
individuals, the population may be able to approach the Pareto front during the
process of evolution. The weight combinations can be distributed uniformly
among the individuals in the offspring population. Let,
wi1 ðtÞ ¼ randð^Þ=^ (15.31)

wi2 ðtÞ ¼1 wi1 ðtÞ (15.32)
where i ¼ 1, 2, . . . , ^, and t is the index for generation number. The function
rand (^) generates a uniformly distributed random number between 0 and ^.
Thus, a uniformly distributed random weight combination (wi1 , wi2 ) among the
individuals can be obtained, where 0 wi1 wi2 1 and wi1 þ wi2 ¼ 1.
2. The idea of a uniformly distributed weight combination can straightforwardly
be extended to a generation-based approach. However, if a random weight
combination is still used, the convergence of the evolutionary algorithm will be
in question. Therefore, a weight combination that is changed gradually and

periodically with the process of evolution is used rather than using a distributed
weight combination randomly.
In this example, this is realized as follows:
w1 ðtÞ ¼ jsin ð2pt=F Þj (15.33)
w2 ðtÞ ¼ 1 w1 ðtÞ (15.34)
where t is the number of generations. We can see from (15.33) that w1(t) changes
from 0 to 1 periodically. The change frequency can be adjusted by F. For example,
in case of F ¼ 400, which means that in every 400 generation, w1 will change from
0 to 1 and then from 1 to 0 four times. Figure 15.4 shows an example of how the
weights change during evolution within 200 generations.
Accordingly, the economic/environmental dispatch for wind and thermal units
can, as an optimization problem, be formulated as an objective function subject to
the system constraints as follows:

9
Minimize OC Pgi ; Wi ; E Pgi >
>
" ! #=
X X
N M X
N
i:e:; Minimize w1 OC Pgi þ OC ðWi Þ þ w2 E Pgi > >
;
i¼1 i¼1 i¼1
(15.35)
Subject to
0 Wi Wri (15.37)
1 w1
w2
0.8
Weight
0.6
0.4
0.2
0 20 40 60 80 100 120 140 160 180 200

Generation
Figure 15.4 Determination of weights: generation-based periodical variation

Pi 0 Pi Di (15.38)
Pi Pi 0 Ui (15.39)
X
N X
M
Pgi þ Wi ¼ L (15.40)
i¼1 i¼1
where
E Pgi : emissions as a function of generator output.
w1, w2: non-negative weights, such that w1 þ w2 ¼ 1; w1 and w2 are used to
make a trade-off between emission and total cost.
Di, Ui: the down- and up-ramp rate limits of the ith generator, and Pi0 is the
former operating point (hour) of the ith generator, respectively [36].
The first part of the objective function, OC(Pgi, Wi), is the expected value of the
combined operating cost as given by (15.22)–(15.29) in addition to the ramp rate
constraints, (15.38) and (15.39), which limit the generated power of the thermal plants.
It is to be noted that the problem that has been framed as a minimization
problem of the fuel cost under load demand constraint and various other constraints
at a specific time of interest is known as the static ED problem. It may fail to deal
with the large fluctuations of the load demand because of the ramp rate limits of the
generators. Thus, due to this large oscillation of the electrical load demand as well
as the dynamic nature of the power system, it requires the inquiry of dynamic
dispatch (DD) problem. DD problem is a development of the static ED problem to
obtain the generation schedule of the committed units to meet the predicted load
demand over a certain period at a minimum operating cost under ramp rate con-
straints and other constraints. Ramp rates are dynamic constraints used to maintain
the life of the generators [37].
Although the environmental pollutant gases of fossil-fuel thermal units consist
of sulphur oxides SOx, nitrogen oxides NOx and dust particles, etc., only the NOx
particles are considered in this study as they are representative of the emissions
produced by the fossil-fuel units. Without loss of generality, the total (ton) emis-
sions E Pgi of the NOx pollutants are given as the sum of the quadratic and
exponential function of the thermal units output power [36]:
X N
E Pgi ¼ ai þ bi Pgi þ gi P2gi þ xi exp li Pgi (15.41)
i¼1
where ai ; bi ; gi ; xi and li are coefficients of the ith generator’s emission function.
The unit of E Pgi is the mass of nitrogen emitted per unit of energy liberated by
the oxidation of the fuel.
15.6 Role of ESS

Renewable energy sources, such as wind and solar, have a massive potential to
decrease the dependency on fossil fuels and GHG emissions in the electric area.
Climate change anxieties, state creativities including renewable portfolio standards

and consumer efforts are resulting in increased deployments of both technologies.
Wind energy has variable and uncertain (sometimes referred to as ‘intermittent’)
output, which is unlike the dispatchable sources used for most of the electricity
generation. Though there are clear benefits of using energy storage system (ESS) to
enable greater penetration of renewable energy, it is important to consider the
potential role of the ESS in relation to the needs of the electric power system.
The variability of these sources has led to concerns regarding the reliability of
an electric grid that derives a large fraction of its energy from these sources.
Because the wind does not blow all the time at any given location, there has been an
increased call for the deployment of ESS as an essential component of future
energy systems that use large amounts of variable renewable resources.
To consider wind power as a constant power source, an energy storage device
should be coupled with the wind generator. The charging and discharging process
of the connected ESS is dependent on the availability of wind power and the
connected load of the system. The periods at which the wind power exceeds the
connected load, the charging of the ESS is conceivable. However, when the load is
in excess of the available wind power, the discharging of the ESS is mandatory.
The capacity of the ESS should be selected carefully to be able to cope with the
load in the wind energy shortage periods in addition to the possibility to be fully
charged in the excess energy periods. A numeric example for further illustration is
introduced as in the following.
Example 15.1 A basic system consists of three thermal units (G1, G2 and G3). Two
wind plants (W1 and W2) and an ESS can be incorporated in different cases of ED
application (Figure 15.5). The daily-load curve is shown in Figure 15.6 and the data
of the non-convex thermal units is given in Table 15.1. The hourly wind speed is
PSO MATLAB toolbox is used to deal with the problem, and referring to
Figure 15.3, its data are [38] the following:
● The initial particles are randomly generated within the feasible range.
● The parameters c1, c2 and inertia weight are selected for best convergence
characteristic with values of c1 ¼ 2.1 and c2 ¼ 2.1.
● Here the maximum value of w is chosen as 0.9 and the minimum value is
chosen as 0.6.
● The velocity limits have a setting of 4 and 4.
● Hundred particles are selected in the population.
ED techniques (CED, DED and SED) are applied to various cases and the
results obtained are described as in the following.
Case #1: CED for the basic system
The results obtained, power scheduling and the corresponding cost, in this case, are
tabulated in Table 15.2 as well as the total cost is plotted in Figure 15.8.
G1
Three
thermal
G2
units
G3
Load
Power conditioning
system
Two W1
wind
plants Power conditioning
W2 system
ESS
Figure 15.5 The studied system: solid links for the basic system and dotted links
for incorporated sources
800
750
700
650
Load (MW)
600
550
500
450
400
350
300
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (h)
Figure 15.6 Daily-load curve
Table 15.1 Data of the non-convex thermal units [26]
Pmin Pmax a b c e f
1 100 600 561 7.92 0.00156 300 0.0315
2 100 400 310 7.85 0.00194 200 0.0420
3 50 200 78 7.97 0.00482 150 0.0630
12
11.5
Wind speed (m/s)
11
10.5
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (h)
Figure 15.7 Wind speed versus time
Table 15.2 Optimum solution obtained from CED
Hour Demand (MW) Pg1 (MW) Pg2 (MW) Pg3 (MW) Min cost 103 ($/h)
1 400 199.736 100.397 99.8665 4.5482
2 445 100.0000 174.799 170.200 5.0282
3 500 199.733 174.799 125.467 5.4003
4 450 199.9677 174.795 75.2388 5.200
5 480 100 249.599 130.400 5.5188
6 530 199.761 180.505 149.733 6.0227
7 750 399.199 201.067 149.733 8.4963
8 620 199.733 270.533 149.733 6.7143
9 643 299.466 242.667 99.8665 6.8145
10 500 200.1595 249.840 50 5.5257
11 520 100 320.133 99.8665 5.5481
12 530 100 330.131 99.8673 5.0013
13 420 100 170.266 149.733 4.4887
14 377 100.284 176.848 99.8674 4.4017
15 730 399.199 230.934 99.8665 8.3610
16 475 199.733 100.000 175.266 5.0924
17 375 100 175.133 99.8665 4.7377
18 563 299.466 113.800 149.733 6.3902
19 563 100 263.400 199.599 6.0512
20 542 199.733 242.400 99.8665 5.9135
21 433 100 233.133 99.8665 4.9637
22 342 100 174.799 67.2001 4.0779
23 542 299.466 142.667 99.8665 5.6994
24 400 199.756 100.374 99.8700 4.6483
9,000
8,000
7,000
Total cost ($)
6,000
5,000
4,000
3,000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (h)
Figure 15.8 Total cost obtained using CED
Case #2: DED for the basic system incorporating wind power plants
Two wind-based DG farms of 45 MW rating are used. Each WT has the following
characteristics:
● Cut-in speed ¼ 5 m/s.
● Nominal speed ¼ 15 m/s.
● Cut-off speed ¼ 45 m/s.
The Weibull function parameters of wind are k ¼ 2 and c ¼ 10 m/s. The results
obtained from the DED are given in Table 15.3. The total cost is drawn as in
Figure 15.9 and compared to that obtained using CED is shown in Figure 15.10.
Figure 15.8, Tables 15.2 and 15.3 and Figure 15.10 show that the total daily cost is
reduced when incorporating wind energy sources with the thermal units.
Case #3: SED for the basic system incorporating wind power plants
The penalty and reserve factors are set to kp,wi ¼ 2 and kr,wi ¼ 4 [13]. The deter-
ministic cost coefficients for ith wind power plant are d1 ¼ 1 and d2 ¼ 1.25 [27].
The results are presented in Table 15.4 and the total cost through the day is shown
in Figure 15.11.
Case #4: ED for the basic system incorporating wind power plants and ESS
The output power of the two wind plants will be constant at 75 MW as the ESS acts
for fixing this power. The total cost through day hours is presented in Figure 15.12.
The results of ED application in MW of each unit and the cost in ($/h) are listed in
Table 15.5. The charging and discharging state of the ESS through day hours is
presented in Figure 15.13.
Case #5: Multi-objective stochastic dispatch
Random distribution of weights within a population method is used to enable PSO
to solve environmental/economic MO problem. The diversity of the Pareto-optimal
Table 15.3 Optimum solution obtained from DED
Hour Load (MW) Pg1 (MW) Pg2 (MW) Pg3 (MW) W1 (MW) W2 (MW) Cost ($/h)
1 400 100 139.06 99.86 27.311 33.76 3,955.5
2 445 199.73 174.79 50 4.48 15.9 4,438.7
3 500 199.73 174.79 99.86 6.68 18.91 4,866.8
4 450 100 193.66 99.86 28.23 28.23 4,360.4
5 480 100 249.59 99.86 4.72 25.80 4,680.5
6 530 299.47 109.11 50 32.21 39.19 4,868.8
7 750 299.46 189.45 199.60 27.31 33.76 6,931.7
8 620 199.73 249.53 149.73 0 21 5,958.3
9 642 399.2 174.8 50 0 18 6,201.5
10 500 199.73 174.79 99.86 4.65 20.94 4,865.8
11 520 199.73 174.79 99.86 19.79 25.8 4,883.3
12 530 199.73 249.59 50 4.264 26.40 5,092.4
13 420 100 174.79 99.86 15.84 29.49 4,044.7
14 377 100 174.79 50 24.58 27.61 3,628.0
15 730 399.19 174.79 99.86 24.04 32.08 6,657.1
16 475 100 249.59 99.86 0 25.53 4,675.6
17 375 100.47 110.47 99.87 32.08 32.08 3,615.5
18 563 199.73 174.79 121.62 33.42 33.42 5,240.5
19 563 299.46 100 99.86 27.48 36.17 5,136.9
20 542 299.46 100 99.86 9.24 33.42 5,117.2
21 433 100 100 175.11 25.8 32 4,264.7
22 342 100 136.59 50 24.63 30.77 3,506.5
23 542 100.15 180.7 199.59 30.77 30.77 5,083.2
24 400 100 100 149.73 19.49 30.77 3,870.3
7,000
6,500
6,000
Cost ($/h)
5,500
5,000
4,500
4,000
3,500
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (h)
Figure 15.9 Total cost obtained using DED
set over the trade-off is shown in Figure 15.14, which shows that the operating cost
of the non-dominated solutions is inversely proportional to their emissions. The
data of the generators up and down ramp rates and emissions characteristics are
given in Table 15.6 [39] as well as the values of the weights of the two objective
functions are shown in Table 15.7. The weights in the proposed MO optimization
9,000
DED
8,000 CED
7,000
Cost ($/h)
6,000
5,000
4,000
3,000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (h)
Figure 15.10 Comparison between DED and CED results
Table 15.4 Optimum solution obtained by SED
Hour Load (MW) Pg1 (MW) Pg2 (MW) Pg3 (MW) W1 (MW) W2 (MW) Cost ($/h)
1 400 50 100 181 34.5 34.5 5,262.5
2 445 68.32 108.9 199.99 33.9 33.9 5,758.92
3 500 76.36 217.40 149.57 37.5 37.5 6,399.4
4 450 68.0 111.23 200 35.4 35.4 5,819.2
5 480 77.16 124.8 200 39 39 6,251.1
6 530 86.7 165.0 200 39.12 39.12 6,626.2
7 750 172.30 308.6 200 34.50 34.5 8,444.19
8 620 124.9 224.26 200 35.5 35.4 7,382.7
9 642 133.5 246.1 200 31.2 31.2 7,372.62
10 500 84.8 150.37 199.9 32.39 32.39 6,419.7
11 520 88.28 163.92 199.99 33.89 33.89 6,514.8
12 530 80.2 170.5 200 39.6 39.6 6,561.5
13 420 50 100.0 190.8 39.6 39.6 5,431.6
14 377 50 100 145.4 40.8 40.8 5,008.8
15 730 165.72 286.37 199.9 38.99 38.99 8,326.72
16 475 77.42 125.97 200 37.8 32.7 6,173.65
17 375 50 100.01 152.26 34 39 4,945.9
18 563 100.84 181.76 185 25.2 40.2 6,925
19 563 95.1 182.6 190 42.6 32.6 6,984.4
20 542 97 164.50 173 13.3 40.2 6,771.6
21 433 54.97 100 200 39 39 5,513
22 342 50 100 110.4 43.8 37.8 4,736
23 542 88.94 177.5 200 37.8 37.8 6,660.3
24 400 50 100 174.4 37.8 37.8 5,289.6
are considered as a uniformly distributed random number between 0 and 1. In each

run of the program, the weights assigned to each objective function are changed
randomly, and after several runs of the program the set of non-dominated solutions
can be obtained.
9,000
8,000
Total cost ($)
7,000
6,000
5,000
4,000
3,000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (h)
Figure 15.11 Total cost of SED
8,000
7,000
Cost ($/h)
6,000
5,000
4,000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (h)
Figure 15.12 Total cost of the power system incorporating wind power and ESS
To better understand the relationship between emissions and generators out-

puts, Figure 15.15 shows the NOx emissions versus the power outputs of the three
thermal units. Two concepts can be used to solve the EED problem:
1. Deterministic Dynamic Economic Emission Dispatch (DDEED) concept
where the imbalance cost due to the mismatch between the actual and sched-
uled power outputs of wind units is eliminated.
2. Stochastic Dynamic Economic Emission Dispatch (SDEED) concept.
Comparing the results of the two concepts as indicated by the Pareto front of
each, Figure 15.16, it is found that stochastic dispatch with wind cost and
emissions are usually higher than that of deterministic dispatch. This is because
the penalties associated with over generation and under generation decrease the
Table 15.5 ED incorporating wind power and ESS

P
Hour Load (MW) Pg1 (MW) Pg2 (MW) Pg3 (MW) w (MW) Cost ($/h)
1 400 100 100 125 75 4,769.4
2 445 100.000 170.13 99.87 75 5,013.3
3 500 100 249.83 75.4 75 5,470.5
4 450 100 175.13 99.86 75 5,194.4
5 480 100 255 50 75 5,443
6 530 100.65 254.46 99.87 75 5,889.3
7 750 100 399.19 175.8 75 7,658.4
8 620 299.98 195.01 50 75 6,629.8
9 642 100 317.27 149.7 75 6,716.6
10 500 199.73 100 125.27 75 5,619.4
11 520 199.73 195.27 50 75 5,660.8
12 530 100.9 254.21 99.87 75 5,763
13 420 100 100 145 75 4,902
14 377 100.0 152.00 50 75 4,423.1
15 730 299.47 174.78 180.73 75 7,527.6
16 475 100 250 50 75 5,406.3
17 375 100 150 50 75 4,389.2
18 563 199.7 100.00 188.27 75 6,103.2
19 563 299.47 100 88.53 75 6,141.2
20 542 100 267.13 99.87 75 5,952.4
21 433 100 174.8 83.2 75 4,935.4
22 342 100 117.0 50 75 4,236.2
23 542 199.73 167.4 99.87 75 5,965.5
24 400 100 100 125 75 4,760
90
Charging
85
Discharging
80
Power (MW)
75
70
65
60
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Time (h)
Figure 15.13 Charging and discharging state of EES with wind power
0.65
0.55
Emissions (ton/h)
0.45
0.35
0.25 ×103
5.1 5.3 5.5 5.7 5.9 6.1
Operating cost ($/h)
Figure 15.14 Trade-off in cost and emission
Table 15.6 Data of the generators up and down ramp rates and
emissions characteristics
UR DR a b g x l
1 90 90 0.0041 0.3276 13.8593 0 0
2 90 90 0.0068 0.5455 40.2669 0 0
3 90 90 0.0046 0.5111 42.8955 0 0
Table 15.7 Values of weights of the two objective functions through 50 runs
w1 w2 w1 w2 w1 w2 w1 w2
0.1190 0.8810 0.4733 0.5267 0.1386 0.8614 0.7792 0.2208
0.4984 0.5016 0.3517 0.6483 0.1493 0.8507 0.9340 0.0660
0.9597 0.0403 0.8308 0.1692 0.2575 0.7425 0.1299 0.8701
0.3404 0.6596 0.5853 0.4147 0.8407 0.1593 0.5688 0.4312
0.5853 0.4147 0.5497 0.4503 0.2543 0.7457 0.4694 0.5306
0.2238 0.7762 0.9172 0.0828 0.8143 0.1857 0.0119 0.9881
0.7513 0.2487 0.2858 0.7142 0.2435 0.7565 0.3371 0.6629
0.2551 0.7449 0.7572 0.2428 0.9293 0.0707 0.1622 0.8378
0.5060 0.4940 0.7537 0.2463 0.3500 0.6500 0.7943 0.2057
0.6991 0.3009 0.3804 0.6196 0.1966 0.8034 0.3112 0.6888
0.8909 0.1091 0.5678 0.4322 0.2511 0.7489 0.5285 0.4715
0.9593 0.0407 0.0759 0.9241 0.6160 0.3840
0.5472 0.4528 0.0540 0.9460 0.5308 0.4692
6
Pg1
Pg2
5
Pg3
Emissions (ton/h)
4
400 500 600 700 800

Generator outputs (MW)
Figure 15.15 Emissions versus the power output of the three thermal units
0.8
SDEED
0.7
DDEED
Emissions (ton/h)
0.6
0.5
0.4
0.3
5.2 5.4 5.6 5.8 6 6.2 6.4 6.6 6.8 3

×10
Operating cost ($/h)
Figure 15.16 Pareto front of DDEED and SDEED concepts
scheduled wind power causing an increase in imbalance charges at a rate

higher than the rate of decrease in fuel cost of thermal units. Over day-hours,
thermal generation schedule, wind power generation, optimum cost and
emissions are obtained as in Table 15.8. It may be noted that, although SDEED
has higher cost than that of DDEED, it is preferred from the reliability and
security point of view as the uncertainty of wind power is considered and
penalized. It is supposed that the total cost and emissions of SDEED could be
decreased if wind power forecast is taken into consideration in this model as
the mismatch between the actual and scheduled power is decreased.
Owing to the large variation of the customer’s load demand and the dynamic
nature of the power system, the investigation of SDEED problem is necessary.
Table 15.8 Optimum solution obtained from SDEED
Hour Load Pg1 Pg2 Pg3 W1 W2 Cost Emissions

(MW) (MW) (MW) (MW) (MW) (MW) 103($/h) (ton/h)
1 400 100.0000 100.0000 200.0000 0 0 4.9880 0.3389
2 445 121.3132 127.1361 175.4890 16.4031 4.6586 5.7882 0.3274
3 500 138.7271 150.6407 178.6715 8.7200 23.2408 6.3913 0.3727
4 450 131.5205 129.8715 172.5259 10.2949 5.7873 5.8927 0.3360
5 480 204.1537 143.5667 104.4820 13.6999 14.0977 5.8181 0.3719
6 530 145.0540 213.0200 149.9610 8.5397 13.4253 6.5277 0.4053
7 750 232.843 301.1052 199.7375 13.1569 3.1574 9.0323 0.8522
8 620 150.8029 304.0147 146.3775 14.2972 4.5079 7.6652 0.6022
9 642 139.5146 366.7067 122.9687 6.0960 6.7140 7.9765 0.7611
10 500 120.1499 300.3758 79.4743 9.7881 0 6.0142 0.4854
11 520 136.1991 266.1519 91.3968 15.8314 10.4207 6.6084 0.40134
12 530 166.2866 244.84497 116.842667 0.0018 2.0239 6.8907 0.47538
13 420 111.2687 162.9554 118.4672 12.6375 14.6712 5.0132 0.3251
14 377 133.9896 138.7291 95.6955 1.9287 6.6573 4.9878 0.2651
15 730 236.77743 241.51693 198.48333 25.5000 27.7223 7.9890 0.6853
16 475 132.27633 166.17633 147.47633 12.28 16.63 5.956 0.393
17 375 185.22267 102.62267 63.372667 7.052 16.64 4.875 0.287
18 563 232.9 171.8 125.9 7.475 24.81 6.847 0.496
19 563 238.1 174.7 130.7 12.37 6.864 6.827 0.520
20 542 328.1 136.5 55.13 3.727 18.37 6.913 0.706
21 433 238.7 110.1 61.34 17.52 5.190 5.628 0.391
22 342 161.5 116.3 54.50 7.597 1.913 4.664 0.220
23 542 215.2 199.5 102.7 8.895 15.53 6.490 0.455
24 400 126.9 137.7 104.0 1.643 29.51 5.293 0.240
References
[1] Huang S., Sun Y. and Wu Q. ‘Stochastic economic dispatch with wind using
versatile probability distribution and L-BFGS-B based dual decomposition’.
IEEE Transactions on Power Systems. 2018;33(6):6254–63.
[2] Wang L. and Singh C. ‘Stochastic economic emission load dispatch through
a modified particle swarm optimization algorithm’. Electric Power Systems
Research. 2008;78(8):1466–76.
[3] Hetzer J., Yu D. C. and Bhattaria K. ‘An economic dispatch model incor-
porating wind power’. IEEE Transactions on Energy Conversion. 2008;
23(2):603–11.
[4] Liao G. C. ‘A novel evolutionary algorithm for dynamic economic dispatch
with energy saving and emission reduction in power system integrated wind
power’. Energy. 2011;36(2):1018–29.
[5] Dhillon J. S., Parti S. C. and Kothari D. P. ‘Stochastic economic emission
load dispatch’. Electric Power Systems Research. 1993;26(3):179–86.
[6] Liang R. H. and Liao J. H. ‘A fuzzy-optimization approach for generation
scheduling with wind and solar energy systems’. IEEE Transactions on
Power Systems. 2007;22(4):1665–74.
[7] Miranda V. and Hang P. S. ‘Economic dispatch model with fuzzy wind
constraints and attitudes of dispatchers’. IEEE Transactions on Power
Systems. 2005;20(4):2143–5.
[8] Wu L., Shahidehpour M. and Li T. ‘Stochastic security-constraint unit
commitment’. IEEE Transactions on Power Systems. 2000;22(2):800–11.
[9] Papoulis A. and Pillai U. Probability, random variables and stochastic
processes. 4th edn. London: McGraw Hill Europe; 2002.
[10] Rice J. A. Mathematical statistics and data analysis. 3rd edn. Belmont, CA,
USA: Duxbury Press; 2007.
[11] Atwa Y. M., El-Saadany E. F. and Salama M. M. A. ‘Adequacy evaluation
of distribution system including wind/solar DG during different modes of
operation’. IEEE Transactions on Power Systems. 2011;26(4):1945–52.
[12] Atwa Y. M. and El-Saadany E. F. ‘Reliability evaluation for distribution
system with renewable distributed generation during islanded mode of
operation’. IEEE Transactions on Power Systems. 2009;24(2):572–81.
[13] Jangamshetti S. H. and Rau V. G. ‘Site matching of wind turbine generators:
A case study’. Energy Conversion, IEEE Transactions on Energy
Conversion. 1999;14(4):1537–43.
[14] Janbaz G. H., Gholizadeh B. and Soltani B. ‘Statistical evaluation of wind
speed and energy potential for the construction of a power plant in Baladeh,
Nur, Northern Iran’. International Journal of the Physical Sciences. 2011;
6(19):4621–8.
[15] Al-Awami A. T., Sortomme E. and El-Sharkawi M. A. ‘Optimizing eco-
nomic environmental dispatch with wind and thermal units’. IEEE/PES
General Meeting; Calgary, AB, Canada, July 2009. IEEE; 2009, pp. 1–6.
[16] Wood A. J. and Wollenberg B. F. Power generation, operation, and control.
2nd edn. New York: John Wiley & Sons; 1996.
[17] Aravindhababu P. and Nayar K. R. ‘Economic dispatch based on optimal
lambda using radial basis function network’. International Journal of
Electrical Power & Energy Systems. 2002;24(7):551–6.
[18] IEEE Committee Report. ‘Present practices in the economic operation
of power systems’. IEEE Trans. Power Apparatus and Systems. 1971;
PAS-90(4):1768–75.
[19] Chowdhury B. H. and Rahman S. ‘A review of recent advances in economic
dispatch’. IEEE Transactions on Power Systems.1990;5(4):1248–59.
[20] Momoh J. A., Adapa R. and El-Hawary M. E. ‘A review of selected
optimal power flow literature to 1993, Part I: Nonlinear and quadratic
programming approaches’. IEEE Transactions on Power Systems. 1999;14
(1):96–104.
[21] Walters D. C. and Sheble G. B. ‘Genetic algorithm solution of economic
dispatch with valve point loading’. IEEE Transactions on Power Systems.
1993;8(3):1325–32.
[22] Tippayachai J., Ongsakul W. and Ngamroo I. ‘Parallel micro genetic algo-
rithm for constrained economic dispatch’. IEEE Transactions on Power
Systems. 2003;17(3):790–7.
[23] Sinha N., Chakrabarti R. and Chattopadhyay P. K. ‘Evolutionary program-

ming techniques for economic load dispatch’. IEEE Transactions on
Evolutionary Computation. 2003;7(1):83–94.
[24] Yang H. T., Yang P. C. and Huang C. L. ‘Evolutionary programming based
economic dispatch for units with non-smooth fuel cost functions’. IEEE
[25] Lin W. M., Cheng F. S. and Tsay M. T. ‘An improved tabu search for eco-
nomic dispatch with multiple minima’. IEEE Transactions on Power
Systems. 2002;17(1):108–12.
[26] Lim S. Y., Montakhab M. and Nouri H. ‘Economic dispatch of power
system using particle swarm optimization with constriction factor’.
International Journal of Innovations in Energy Systems and Power.
2009;4(2):29–34.
[27] Hetzer J., Yu D. C. and Bhattaria K. ‘An economic dispatch model incor-
porating wind power’. IEEE Transactions on Energy Conversion. 2008;
23(2):603–11.
[28] Hong Y. Y. and Lian R. C. ‘Optimal sizing of hybrid wind/PV/diesel gen-
eration in a stand-alone power system using Markov-based genetic algo-
rithm’. IEEE Transactions on Power Delivery. 2012;27(2):640–7.
[29] Granelli G. P., Montagna M., Pasini G. L. and Marannino P. ‘Emission
constrained dynamic dispatch’. Electric Power System Research. 1002;
24(1):55–64.
[30] Farag A., Al-Baiyat S. and Cheng T. C. ‘Economic load dispatch multi-
objective optimization procedures using linear programming techniques’.
[31] Srinivasan D., Chang C. S. and Liew A. C. ‘Multi-objective generation
schedule using fuzzy optimal search technique’. IEE Proceedings of
Generation, Transmission and Distribution. 1995;141(3):233–42.
[32] Poli R. ‘Analysis of the publications on the applications of particle swarm
optimisation’. Journal of Artificial Evolution and Applications. 2008;
2008:1–10.
[33] Park J. B., Lee K. S., Shin J. R. and Lee K. Y. ‘A particle swarm optimi-
sation for economic dispatch with non-smooth cost functions’. IEEE
[34] Berizzi A., Innorta M. and Marannino P. (eds.). ‘Multi-objective optimiza-
tion techniques applied to modern power systems’. IEEE PES Winter
Meeting; Columbus, OH, USA, Jan/Feb 2001. IEEE; 2001, Vol. III,
pp. 1503–8.
[35] Jin Y., Okabe T. and Sendho B. (eds.). ‘Adapting weighted aggregation for
multi-objective evolution strategies’. Proceedings of International Conference
on Evolutionary Multi-Criterion Optimisation; Zurich, Switzerland, Mar 2001.
Springer, pp. 96–110.
[36] Abido M. A. ‘Multi-objective evolutionary algorithms for electric power
dispatch problem’. IEEE Transactions on Evolutionary Computation.
2006;10(3):315–29.
[37] Xia1 X. and Elaiw A. M. ‘Dynamic economic dispatch: A review’. The

Online Journal on Electronics and Electrical Engineering (OJEEE). 2010;
2(2):234–45.
[38] MATLAB/Optimization Toolbox. Available from http://www.mathworks.
com/matlabcentral/fileexchange/7506-particle-swarm-optimization-toolbox
[Accessed Sep 2018].
[39] Cai J., Ma X., Li Q. and Peng H. ‘A multi-objective chaotic particle swarm
optimization for environmental/economic dispatch’. Energy Conversion and
Management. 2009;50(5):1318–25.
Chapter 16
Generation scheduling for power grids
with renewables
16.1 Introduction
Electricity demand can exhibit large variations from weekdays to weekends, one
day to another, peak to off-peak hours as well as from one season to another. As
explained in Chapter 15, optimal loading of each generating unit in a power system
encountering conventional thermal units integrated with renewable wind sources
can be determined using economic dispatch (ED) techniques. All available gen-
erating units can be turned on to meet the demand, but it is usually not economical
to continuously run all the units available all the time. Hence, it is desirable to look
for an efficient economic short-term generation scheduling that plays an important
role in the economic operation of a power system. The principal objective of the
short-term generation scheduling is to efficiently determine the commitment of
generating units as well as the economic loading of each committed unit so that the
forecasted load demand and spinning reserve can be met over a short period ran-
ging from 1 day to 1 week. The resultant schedule should minimize the system
production cost during a specified period while simultaneously satisfying system
and operational unit constraints. Thus, scheduling power system operation should
involve two basic economic decisions:
1. A unit commitment (UC) decision that determines which units should be
brought online to meet the expected load demand, power losses and reserve
requirements.
2. An embedded ED decision that determines the most economic generation level
for each of the committed (synchronized) units.
These two decisions must be taken to minimize the total start-up, shutdown
and running costs subject to various constraints such as power balance, spinning
reserve and minimum up and downtime constraints [1].
Based on this aspect, the problem of generation scheduling of a conventional
power system without and with renewables (wind and solar photovoltaic (PV)) is
studied in this chapter. Several critical constraints that govern the operation of
synchronized generators and affect the generation scheduling are crucial. The most
widely solutions to the generation scheduling are discussed. Emphasis is placed on
the advantages and limitations of each method.
Moreover, security-constrained generation scheduling (SCGS) problem for a

grid with renewables is formulated. The formulation considers the stochastic nature
of both the wind and PV power output, and imbalance charges due to mismatch
between the actual and scheduled wind and PV power outputs. A hybrid technique
in which the basic elements are a genetic algorithm (GA) with artificial neural
network (ANN), and a priority list (PL) can be used to minimize the total operating
costs while satisfying all operational constraints considering both conventional and
renewable energy generators.
16.2 Thermal generating units

The main generating units in a thermal power system [2] can be categorized under
the following three groups:
● Nuclear steam thermal units: These units are characterized by having high
capital investment but low online generation costs. Normally, such units are
operated as base generation.
● Fossil-fuelled steam thermal units: These units offer the flexibility to follow
the changes in load and are normally considered for cycling and peak duties.
One characteristic of these units as well as the nuclear units is that a minimum
period should elapse before they can be started up (or shut down).
● Gas turbine: These units are another option for dealing with peaks in demand.
The reason is that although they have comparatively high online generation
costs, they can be started up quickly and shut down after operation for a short
period. The start-up costs are comparatively insignificant.
● Combined cycle gas turbine: Modern generating units are increasingly having
to cope with complex operation regimes to increase efficiency and reduce
costs. A combined cycle unit can combine several subunits such as gas turbine
and steam turbine subunits, resulting in several different operating modes. The
plant efficiency depends on the number of gas turbines in operation.
16.2.1 Characteristic of thermal units

From an economic point of view, the set of input–output characteristics of a ther-
mal power generation unit is of most interest. The input in this case may be either
the heat energy requirements, H, or in terms of total fuel cost, F, measured in
MBtu/h or $/h, respectively, whereas the net output is the electrical power output,
P, measured in MW, available to the electric utility system. The total operating cost
includes the fuel cost, and the cost of labour, supplies and maintenance. However, it
is common to assume that the cost of labour, supplies and maintenance is a fixed
percentage of the fuel costs.
The input–output characteristic of a steam unit in idealized form is presented
as a smooth, convex curve (Figure 16.1(a)). The incremental heat rate character-
istic, which is the derivative of the input–output characteristic, for a unit of this
type is shown in Figure 16.1(b). The data in this curve are in terms of Btu/kWh
Generation scheduling for power grids with renewables 403
Incremental heat rate ∆H (Btu/kWh)

∆P
Incremental fuel cost ∆F ($/kWh)
∆P
Input H (MBtu/h) or F ($/h)
Approximation
Output Output
Pmin Pmax P (MW) Pmin Pmax P (MW)
(a) (b)
Unit heat rate H (Btu/kWh)

P
Output
Pmin Prated Pmax P (MW)
(c)
Figure 16.1 Typical steam unit characteristics: (a) input–output characteristic,

(b) incremental characteristic and (c) net heat rate characteristic
(or $/kWh) versus the net power output of the unit in MW. Very often, this char-
acteristic is approximated by a sequence of straight-line segments.
In addition, an important characteristic of a steam unit, which is the unit (net)
heat rate characteristic, is represented by H/P versus P as shown in Figure 16.1(c).
It is proportional to the reciprocal of the usual efficiency characteristic developed
for machinery [3]. However, a series of straight-line segments may be also used to
represent the input–output characteristics. The different representations will of
course result in different incremental heat rate characteristics.
16.3 Formulation of the thermal scheduling problem
The thermal scheduling problem involves the determination of start-up and shut-
down time as well as the power output levels of all system generation units at each
time step, over a specified scheduling period T, so that the total start-up, shutdown
and running costs are minimized subject to several system and unit constraints.
Obtaining an optimal schedule of generation involves the solution of a mixed-
integer non-linear optimization problem with many constraints.
16.3.1 Unit commitment

UC represents a challenging problem in the operation of a power system. It belongs
to the class of complex, non-linear, mixed-integer programming whose objective
function is the system production cost over the entire period. The solution of the
UC problem involves non-linear optimization, comprising of both integer and
continuous variables, with many equality and inequality constraints.
The UC problem in a thermal system arises because of the following [3]:
● The variation of system load demand during the day. Although some gen-
erators can be operated at a fixed output level throughout the day as ‘base
generation’, it is necessary to start up and shut down some generators during
the day as ‘cycling units’, to match the system total generation to the demand.
● The variety in the generating units, in terms of operational characteristics and
economic efficiency. Also, the fact that it usually takes some time to start up
and shut down a thermal generating unit raises the question as to which com-
bination of generators is the most economical and feasible to meet the system
load demand at various times.
● There are several critical constraints that govern the operation of a synchro-
nized generator and affect the UC. It is imperative that these constraints are
taken into consideration, to achieve a feasible and realistic solution, while
solving the problem of UC.
16.3.2 Problem objective function

The main objective of scheduling in thermal systems is to minimize system
operation costs. The total production cost, Fcost, for the scheduling period is the
sum of the running, start-up and shutdown costs for all the units. Shutdown cost is
defined as the amount of the fixed cost for each unit per shutdown and is applied
in the hour the unit is taken offline. The total production cost to be minimized is
given by
" #
XT X
Fcost ¼ Fi Pij þ SC i;j (16.1)
j¼1 i2G
where
Fi ðPij Þ: the production cost of the ith thermal unit in the jth period, and Pij is
the power output of ith unit in the period j;
SCi;j : unit start-up cost;
T: the time horizon, which is taken as 24 h;
G: the number of generating units.
16.3.3 Fuel costs

Fuel costs of thermal units are usually represented by a quadratic heat rate curve as
a function of power output multiplied by the price of the selected fuel. If Fi is the
function that relates the generator power output, Pi, to fuel cost, an example of the
function often used is
Fi ¼ ai P2i þ bi Pi þ ci (16.2)
where ai, bi and ci represent the unit cost coefficients.
Also, a suitable higher-order polynomial may be used to express the unit cost
function. In some studies, fuel costs are represented by a fixed linear function of
generator output level. However, if more precise and better results are needed,
several linear segments (a piecewise approximation) may be implemented.
16.3.4 Transition costs

The start-up cost of a unit is a mixture of fixed and variable downtime-dependent
costs. The generator start-up cost, SCi;j , depends on the time the unit has been off
prior to start-up and can be represented by an exponential cost function [4]:
h h ii
SC i;j ¼ uij ai þ bi 1 exp Ti;joff =ti (16.3)
where
ai : integrates labour start-up cost and equipment maintenance cost of unit i;
bi : start-up cost of unit i from cold condition;
ti : time constant that characterizes unit i cooling speed;
Ti;joff : the time a unit has been off prior to start up;
uij : the status index of ith unit in the jth period (1 for up and 0 for down).
However, there are two approaches to treating a thermal unit during its down
period. The first allows the unit’s boiler to cool down and then heat, back up to
operating temperature in time for a schedule turn on, and in this case, the start-up
cost is governed by (16.3). The second is called banking and requires that enough
energy be input to the boiler to just maintain operating temperature [3].
The objective function Fcost is minimized, subject to several system and unit con-
straints. Each individual power system may specify its own set of constraints depending
on certain factors such as network topology, reliability and security load curve char-
acteristic and so on. These constraints are described in the forthcoming sections.
16.3.5 System constraints

System constraints are applied in the power system operating objective function to
keep the system operating point within acceptable limits of stability and to meet
security requirements. The most common system constraints are discussed in the
following subsections.
16.3.5.1 Active power balance

The load demand variation over the scheduling interval provides one of the biggest
challenges for the scheduling algorithm. The total power generated by the online
units must supply the load demand, Pd , and system losses, PL, that is,
X
ui;j Pij ¼ Pdj þ PL (16.4)
i2G
16.3.5.2 Load forecast error and system reserve requirements

One of the primary objectives of UC is to ensure the security of the system to
withstand predicted load uncertainty and generation forced outage. Spinning and
supplemented reserves must be allocated in the UC to meet the security criteria set
down by the management. Spinning reserve can be defined as the extra generation
available on demand from the synchronized generators, while the supplemented
reserve specifies an amount of additional reserve that must be available to the
system and is composed of hot standby units and rapid start-up units [5].
The hourly spinning reserve requirement, R, that must be met over the whole
scheduling period is as follows:
X
ui;j Pmax
ij Pdj þ PL þ R (16.5)
i2G
16.3.5.3 Transmission network constraints

The transmission line loading limits must not be violated by the power flow
resulting from a given schedule. These constraints are especially crucial for sys-
tems that are loaded closer to their thermal limits, that is,
Pmax max
m Pmj Pm ; m ¼ 1; . . . ; M (16.6)
where
Pmax
m : the maximum transmission capacity of line m;
Pmj : the load flow on the line m at period j;
M: the total number of transmission lines.
Transmission capacity constraints present a challenge to researchers in the UC
field. Usually, linear DC transmission constraints are considered in UC formulation
for system security purposes. However, Pmj can be formulated as
X X
NB
Pmj ¼ kmi Pij ui;j kmj Pd s;j (16.7)
i2j s¼1
where NB, Pds,j are the total number of system buses and load at Bus s, respec-
tively. The sensitivity coefficients kmi and kmj can be determined, based on DC or
AC power flow corresponding to normal or contingency conditions [6].
16.3.5.4 Emission constraint

With the increased demand for a cleaner environment, the thermal schedules
must not violate allowable emission limits, and if possible, the nitrogen dioxide
(NO2), sulphur dioxide (SO2) and carbon dioxide (CO2) emissions should be
minimized [7].
16.3.6 Unit constraints

The generating unit can be subjected to several constraints that must be accounted
for in the process; these include the following.
16.3.6.1 Unit minimum up and downtime

Unit minimum up and downtime limit the thermal stresses that the machines would
be subjected to if they were started or shut down. Thus, a unit can only be shut
down (started) once it has been staying on (off) for a minimum period,
called minimum up (down)time. Thus, the limits of minimum up/downtime (MUT/
MDT) of units must not be violated, that is,

on
Ti;j1 MUT ui;j1 ui;j 0 (16.8)

off
Ti;j1 MDT ui;j1 ui;j 0 (16.9)
where T off =T on is the unit off/on time.
16.3.6.2 Real power operating limits

It is important that unit rated minimum capacity must not be violated. The max-
imum and minimum real power limits are generally expressed as
Pmin
i Pij Pmax
i ; i 2 G; j 2 T (16.10)
16.3.6.3 Unit ramp rate limits

In the boiler fuel control, there is a limiting factor for increasing the rate of fuel in
feed from the nulls and the time lag to convert the chemical energy of coal to heat
energy. Similarly, when a unit is in the shutdown process, it will take a while for
the turbine to cool down. Before the unit generating capability decreases to its
lower limit, the residual energy must be used to meet the load demand [8]. These
ramp limits may be expressed as
Pij Pij1 URi (16.11)
Pij1 Pij DRi (16.12)
where URi , and DRi are the ramp up and ramp down rate limits, respectively.
In addition to the above-mentioned constraints, there are other unit constraints,
which include the following:
● The initial unit states at the start of the scheduling period.
● Crew constraints that limit the number of units that can be started at the same
time in a specific plant, and unit operating status or mode restrictions.
● Fixed generation units are prescheduled units. Any unit may be prescheduled
to ‘must be on’, ‘must be off’ or fixed generation for certain intervals of the
study period. Specification of such requirements is fed to the UC program as
Constraints
– System constraints Estimated
– Unit constraints production
cost
– Must be on/off and
fixed generation
Unit commitment Economic dispatch

(UC) main program Commitment schedule (ED) subprogram
Data
– Generators data
– Forecasted load Generation
– Spinning reserve schedule
– Maintenance schedule
and forced outage
– De-rated capacity
Figure 16.2 UC with ED processes
input data. Units scheduled for maintenance or in forced outage can, therefore,
be treated as ‘must be off’ units.
● De-rated capacity: Partial outages of the units leading to de-rated capacity or
changes from de-rated state to full capacity state or to another de-rated state.
A typical block diagram for the input/output of UC with ED program is
16.4 Economic dispatch

Once the generating units in the power system are scheduled and the committed
units are identified, it is necessary to determine the optimal allocation of the system
power demand among the committed units, while satisfying different constraints.
ED problem can be treated as a non-linear optimization problem, which consists of
several equality and non-equality constraints. The problem is formulated as follows:
X
Minimize Fcost ¼ Fi Pij (16.13)
i2G
subject to (16.4)–(16.6) and (16.10)–(16.12).

The ED problem can be handled from the perspective of a sub-problem of
generation schedule.
16.5 Review of thermal scheduling techniques

Generation scheduling in practical power systems poses difficult analytical chal-
lenges, as it involves the solution of a large-scale mixed-integer non-linear, opti-
mization problem, with several constraints. The ideal solution to the problem can
be obtained by complete enumeration, which cannot be applied to realistic power

systems due to its excessive computation time requirement. In the past, a few
attempts using both empirical analysis and rigorous mathematical programming
techniques have been used for solving the generation scheduling problem [9]. A
closer review of some solution methods to the generation scheduling problem is
given in the next subsections.
16.5.1 Priority list

The units in this method are committed according to a PL based on full-load average
production and transition costs [3]. The PL schemes could be obtained after an
exhaustive enumeration of all unit combinations at each load level. The pre-
determined order is then used to commit units such that the system load is satisfied.
The merit order dispatch is the simplest dispatch method and it relies on the
availability of linear or piecewise linear cost function. The committed generators
are indexed in order of increasing incremental cost and initialized at their minimum
power output levels. The generators are then considered for loading to their max-
imum capacities according to their rank in the PL, until the load demand is satis-
fied. The system incremental cost is then determined by the partly loaded generator
in the system.
This method provides reasonably accurate results only if linear cost functions
are used. If a more accurate modelling of the unit performance is required, then the
merit order scheme fails to provide an acceptable solution. However, its simplicity,
speed and ability to guarantee feasible solutions make it one of the most widely
used method by electric utilities.
16.5.2 Dynamic programming

Dynamic programming (DP) is one of the techniques that have been most widely
researched because of its ability to handle any problem, which can be formulated as
a set of separable state transitions [10]. It works by determining the optimal set of
state transitions, which will bring the system from an initial state to a final state. It
decomposes the problem in time, beginning at the first hour of the scheduling
period, committing the units progressively 1 h at a time and storing the unit com-
binations at each hour and their associated costs. At the end of the schedule, the
steps are traced backwards to obtain the optimal combinations that results in the
least cost schedule. However, only small systems can be optimized, as the number
of states to be stored and evaluated increases exponentially as the number of
variables increases. To avoid this curse of dimensionality problem in UC, DP is
used to solve the commitment of a single unit, in combination with other techniques
such as Lagrangian relaxation (LR). The truncated or variable window DP method
is another attempt at reducing the DP storage difficulties [11].
DP method has also been employed for ED [12]. This method, however, suf-
fers from the problem of large dimensionality as it requires exploration of the entire
solution space. The ability to find the global optimum solution is dependent on the
size of the discrete load step assigned to each generator in the solution process.
16.5.3 Lagrangian relaxation

LR is a mathematical decomposition technique, which is rapidly growing in
importance as a solution method for the generation scheduling problem, especially
for large power systems. The LR decomposition technique based on duality theory
generates a separable problem by integrating some constraints into the objective
function through penalty factors, which are functions of the constraint violations [11].
The penalty factors, called Lagrangian multipliers, are determined iteratively and
they determine the solution quality. The approach involves decomposition of the
problem into a sequence of master problem and easy sub-problems. Each sub-
problem consists of determining the commitment schedule for a single unit over
the planning horizon and is solved independently using any suitable technique such
as dynamic or mixed-integer programming. The only link between these sub-
problems is the Lagrangian multipliers, which are adjusted in the master problem
by a sub-gradient method or heuristically to ensure that the system constraints are
met. An iterative approach usually ensures convergence to a near-optimal solution
and generates bounds, which give an indication on the proximity of the solution to
the true optimal.
The drawback of this technique is that the quality of the final LR solution
depends on the sensitivity of the commitment to Lagrangian multipliers. Slow and
unsteady convergence of LR has always been a problem in finding the global
optimum solution. The difficulties are often explained by the non-convexity of this
type of optimization problem. The augmented Lagrangian method has been applied
to overcome the problems of the LR method [12]. In this method, quadratic penalty
terms associated with power demand are added to the objective function to improve
the convexity of the problem.
16.5.4 Branch and bound

The branch and bound method is a powerful enumeration strategy that helps to
reduce the number of combinations of integer variables considered in a mixed-
integer non-linear programming problem [13]. It is a technique to solve a discrete
variable problem by solving a sequence of simpler problems derived from the
original problem. As reported in [14], the UC problem can be formulated as a linear
mixed-integer programming and then standard integer programming algorithms are
used to solve for the commitment schedule. Several problem approximations and
extensions to the basic branch and bound algorithm have been described to make
the solution approach tractable for this problem. The advantage of branch and
bound technique is that it can provide a sequence of solutions with estimates of
their suboptimality. However, it suffers from the curse of dimensionality as the
execution time exponentially grows with the size of the UC.
16.5.5 Linear programming

Linear programming [11] is usually combined with other techniques such as DP
and mixed-integer programming for solving the continuous ED sub-problem. The
advantage of linear programming lies in its fast and reliable solution, whereas its
main disadvantage is that all the objective functions and constraints must be line-
arized, which often leads to suboptimal solutions.
16.5.6 Artificial intelligence

Recently, a new direction in research for reducing the computation time while
preserving the system characteristics has pointed out the potentials of using artifi-
cial intelligence (AI) techniques. The AI techniques have incorporated the practical
operational policies in the mathematical techniques to improve the system models
considerably. They, therefore, accommodate more complicated constraints and are
claimed to have better solution quality. The AI techniques, unlike strict mathe-
matical methods, have apparent ability to adapt to non-linearities and dis-
continuities found commonly in physical systems. Although they look promising,
they are still evolving.
16.5.6.1 Heuristic and expert systems

Heuristic and expert system methods are based around rules that are derived from
the system characteristics and the knowledge of experienced power system opera-
tors. The expert system improves the solution of the UC problem by adjusting the
program’s parameters through interaction with the system operator. Several expert
system-based techniques with specified rules for systems have also found wide-
spread usage in utility scheduling [15]. Also, the fuzzy set theory has been applied
to solve the UC problem [16]. With this approach, the security constraints, gen-
eration costs and load demands have all been expressed in fuzzy set notation. The
fuzzy DP has then been performed to reach an optimal commitment schedule [17].
However, the rule-based reasoning method has its own drawbacks, primarily
because the human knowledge may not be easily expressed and classified by rules,
and the solution of a problem based on many rules may be costly and require a long
processing time.
16.5.6.2 Simulated annealing

Simulated annealing (SA), a heuristic technique, takes the analogy of the physical
annealing of melted particles [18]. Starting from a very high temperature, a melted
particle is cooled gradually until it is solidified at a low temperature. This physical
annealing process is simulated in the SA technique for the determination of near-
global or global optimum solutions for combinatorial optimization problems.
Different researchers have used SA method to solve the UC problem. The main
drawback of the SA method is the long computation time as well as the inability to
guarantee feasible solutions.
An SA-based ED algorithm has been reported in [19]. It has been shown that
the algorithm can determine the global or near-global optimum solution for ED.
However, the convergence of this algorithm is slow.
16.5.6.3 Artificial neural network

The mechanism of ANN is based on simulating the learning process of the human
brain. For implementing ANN, the necessity of the mathematical representation of
a physical problem, which on some occasions may create additional problems in

achieving a reasonable solution, can be relaxed. The Hopfield neural network
mimics the computational capability of biological organisms by utilizing simple
computing neurones with high interconnectivity and parallelism. The Hopfield
neural network to solve the UC problem has been studied as in [20]. The average
execution time was approximately 2.5 times that of the Lagrangian method for
30 generators. In addition, ANN has a great potential in facilitating an economic
solution within a reasonable time and a combination of the back-propagation ANN
combined with DP have been used to solve the UC problem [21]. The training data
set used for back-propagation contains past unit schedules, which are assumed to be
optimal. Once trained, the network is presented with a new load profile for which a
schedule is desired. Although the ANN can produce fast schedules, its main
drawback is the difficulty in providing feasible and near-optimal solutions.
The Hopfield neural network has also been applied to optimal economic/
environmental dispatch of thermal generating units [4]. A simulator has been
developed and criteria for selecting its parameters have been explored to obtain
optimal dispatch in a minimum number of iterations. However, many iterations and
oscillation are the major concern in solving the ED problem using the Hopfield
neural networks.
16.5.6.4 Tabu search

Tabu search (TS) is an iterative improvement procedure that starts from some
initial feasible solution and attempts to determine a better solution in the manner of
a greatest-descent algorithm [5]. However, TS is characterized by an ability to
escape local optima using a short-term memory of recent solutions. Moreover, TS
permits backtracking to previous solutions, which may ultimately lead, via a dif-
ferent direction, to better solutions. An application of the TS method to solve the
UC problem is presented in [6]. Numerical results show an improvement in the
solution cost, but it suffers from a long execution time.
16.5.6.5 Evolutionary algorithms

All evolutionary algorithm models are based on a learning process using a popu-
lation of potential solutions to a problem, each representing a search point in the
Euclidean space of all the possible solution. The initial population is usually ran-
domly selected and evolves towards the optimal solution according to rules of
selection and other operators such as recombination and mutation. Everyone in the
population is awarded a fitness measure based on the problem objective function,
which determines its ability to propagate its elements in the search process. The
recombination allows the mixing of parental information while the mutation
operator introduces diversity in the population [22]. There are three main approa-
ches: GA, evolution strategies (ES) and evolutionary programming (EP). These
methods differ in terms of representation, operators and selection process. The goal
of the EP is to achieve intelligent behaviour through simulated evolution.
GA is one of the most widely studied evolutionary techniques. The GA has
been successfully applied to solve the short-term generation scheduling [23].
The results are encouraging, but unrealistic for real-time operation because of long
execution time.
GA-based method has also been reported recently for solving the ED problem [8].
On the other hand, extended evolutionary algorithms for solving the optimal reactive
power dispatch has been explored [9]. Good and reliable performance has been
achieved and validation tests using the standard IEEE57 system have been reported.
However, none of the basic evolutionary computational methods can guarantee
an optimal solution to the scheduling problem. They only provide near-optimal
solutions, and the quality of each solution provided is affected by either solution
time limitations or by premature convergence. Various enhancements and mod-
ifications are required to enable the basic evolutionary computational models to be
applied successfully in the solution of the scheduling problem.
16.6 Problem representation and solution strategy

The objective is to efficiently determine the effective commitment of the generat-
ing units as well as the output level of each committed unit to meet the forecasted
demand and spinning reserve at the minimum possible production cost over a
specific period, for example, one day. Moreover, this should satisfy system and unit
constraints. In practice, short-term generation scheduling often requires a method
that is fast to meet system changes and reduce the scheduling error. With the advent
of the stochastic global search algorithms, the GA has been successfully applied to
solve the short-term generation scheduling problem [23]. The results are promising,
but unrealistic for real-time operation because of long execution time. To surmount
this problem, a hybrid technique using GA, PL and ANN can be applied accounting
for load forecast uncertainty to accelerate and enhance the performance of the
short-term generating scheduling solution.
In the UC problem, with ANN computing, most of the time is spent on offline
training the network as ANN accumulates knowledge during offline training from
the given input/output data pairs. Once the network is completely trained, the
online operating time would be very short as compared to that attained for analy-
tical programming techniques [21]. The ANN-supported GA (ANN/GA/PL) is a
hybrid of intelligent system and traditional heuristic technique. First, a pre-
scheduling of the generating units using the ANN is presented. The GA with PL is
then used to search for the best combination of the hesitant units that result from the
ANN and satisfying both the system and unit constraints. With its two-stage pro-
cessing, the algorithm can benefit from the advantage of this technique. The ED
sub-problem is then solved for each member of the population satisfying all con-
straints. The flow chart of the (ANN/GA/PL) hybrid technique is shown in
Figure 16.3.
16.6.1 GA implementation
The two main actions that link the GA to the problem it is solving are the means of
translating the problem solution to a chromosome representation and an evaluation
Start
Read in load
profile
Use of ANN for

prescheduling all
generating units
Examine ANNs’ output

Categorize the state of generating units:
on/off or hesitant
No Hesitant
Certain state unit
state unit?
Yes
h=1
Run PL heuristic algorithm
Run GA with a copy of PL

solution in the initial population
Dispatching load among committed

units by using merit order
No
Last hour? h=h+1
Yes
Stop
Figure 16.3 Flow chart of (ANN/GA/PL) hybrid technique
function that returns a measurement of worth for any chromosome in the context of
the problem. The success of the GA approach is heavily dependent on the problem
encoding and the choice of the evaluation function.
The thermal scheduling problem is a mixed-integer non-linear optimization
problem that lends itself to a suitable binary representation. For the application of
GA to the UC problem, a binary string of [‘0’, ‘1’] denoting the off/on state of the
unit is randomly generated. The total solution string length is a product of the
scheduling period T and the number of generation units G. A typical UC problem
representation for scheduling G units over T hours is shown in Figure 16.4.
Hour #1 Hour #2 Hour #T

μ11 μ12 μ1G μ21 μ22 μ2G μT1 μT2 μTG
Figure 16.4 An example of binary UC problem encoding (mt,i ¼ 0 or 1, where t is

the time index and i is the unit index)
16.6.2 Decomposition methods

It is expected that as the number of units increases, the string length increases
proportionately resulting in a long computational time. To surmount the problem, a
sequential GA is used to solve the scheduling problem, and this limits the selection,
mutation and crossover of the GA to a single time interval instead of the whole-
time interval. However, this process does not strictly partition the problem into a
single time span, but involves cumulative time-span partitioning, in that the linking
constraints’ parameters, such as unit minimum up/downtime, are continuously up-
dated and re-evaluated as the time steps increase. Also, an elitist GA search is used
to guarantee that the best solution at each generation is retained and used in the
subsequent generation, thereby ensuring that no good solution is being lost during a
GA search.
Since a GA works by searching both the feasible and infeasible problem
solution search space, it is necessary because of its population-based nature and its
selection mechanism to differentiate between the performances of two non-feasible
strings. Thus, a string that violates a constraint more than another should receive a
higher penalty and expect a corresponding diminished chance of selection. The
discrimination between feasible and non-feasible solutions is achieved by making
sure that any non-feasible solution is assigned a fitness lower than that of a feasible
solution through the application of the penalty algorithm.
The constraints of the problem are handled by including penalty factors within
the GA.
The fitness function of the scheduling problem is a linear combination of
the total operating cost and the penalty factors and can be represented by (16.14).
The minimization cost function is transformed into a maximization problem by the
reciprocal of the cost function.
X
T
FGA ¼ OC Pij þ a1 FSC ðjÞ þ a2 FDC ðjÞ þ a3 FTRðjÞ
j¼1 (16.14)
þ a4 FTDðjÞ þ a5 VTLðjÞ
and
" #
X
N X
N
OC Pij ¼ FC i Pij þ SC ij þ SDC ij (16.15)
i¼1 i¼1

where FC i Pij , Pij , SC ij and SDC ij are the production cost of ith thermal unit, its
power output, start-up and shutdown costs in jth hour. a1 ; . . . ; a5 are penalty
factors, and FSCðjÞ, FDCðjÞ, FTRðjÞ, FTDðjÞ and VTLðjÞ are the penalty terms,
which are utilized when a unit does not comply with the minimum up/minimum
downtime, fails to meet the reserve and the load demand, respectively, and when
the transmission line constraints of the system are violated.
Choosing a penalty factor that combines well with the cost function is a dif-
ficult task as it affects the performance of the algorithm. Different values of
a1 ; :::; a5 are tried to calculate the effective values that satisfy the violated
constraints.
16.6.3 PL implementation
To start a typical run of a GA, an initial population of strings is generated randomly
subject to the initial conditions of the studied system. However, an extensive
research has indicated the importance of the choice of the initial population [24].
By including the domain knowledge into a hybrid GA technique, the hybrid tech-
nique would outperform both the traditional GA and the conventional methods.
The PL scheme is the most popular technique of solving the UC problem, as it
is a fast and reliable method. To use a hybrid technique comprising the GA with the
PL at each hour, the PL algorithm is run followed by the GA. The latter is run with
the inclusion of a copy of the PL solution in its initial population.
16.6.4 ANN implementation

ANN is a high-speed online computational technique, which once trained through
an offline algorithm using example patterns, can provide an output corresponding
to a new pattern without any iteration in real time. Three-layered feed-forward
ANN with a sigmoid function is used in this study to preschedule the generating
units and the back-propagation algorithm is used for training the ANN.
Like the human brain, as the dimension of patterns increases, the ANN may
find it difficult to remember and recognize each pattern. Thus, each generating unit
is prescheduled by a separate ANN as this has proved to be more efficient than
training all the units in advance [25]. Hence, the number of ANNs is determined by
the number of the generating units. Each ANN has 24 input neurons corresponding
to the 24-hourly loads, while the output layer has 24 neurons. The optimum number
of hidden neurons is determined by trial and error. In the technique presented here,
this is equal to nine.
Load-forecast uncertainty imposes a threat to a firm UC decision.
Underestimation or overestimation of load can lead to a failure to provide enough
reserve or lead to an unnecessary large amount of spinning reserve, which in turn
leads to higher cost [26]. In practice, load demand is very much dependent on
weekdays, weekends including holidays and in many countries is dependent on
different seasons in a year. An extensive series of studies has shown that to affect a
single ANN-based technique to cater for all the foregoing different loads is not very
satisfactory. The approach adopted herein is thus based on dividing the load
demands into various categories, and in practice, there will thus be several ANN-
based algorithms each to deal with a category. Regarding any one category, the
practical error margin is a maximum of 5% [27] and hence, the training and
testing data used to illustrate the technique described in this study is realistic in
terms of evaluating the performance of the method. After training and testing the
ANN using different load curve samples, the ANN produces an output feature value
(i.e. the preschedule), which is classified as follows:
● Certain state unit: The output value of this is greater than or equal to 0.9 (with
certain ‘on’ state), or an output feature value of less than 0.1 (with certain ‘off’
state).
● Hesitant unit: 0.9 > the output feature value 0.1. Here, the GA or the GA/PL
is used to solve the problem of scheduling of hesitant units.
16.6.5 ED implementation
In the scheduling process, once a decision is made on whether a unit is to be on or
off, a further optimization process must be carried out to optimally allocate the load
to the online units, a process which is referred to as ED. The main objective of the
ED is to minimize fuel costs whilst allocating the load among the online units. This
should also satisfy the physical limitations of units as well as those on the power
system. Owing to the ability of the GA to look for the global optimal solution, it is
applied to the ED problem. The GA is thus further extended for solving the ED
problem concurrently with the UC problem. At each time interval, individual
generators are represented by ten bits, where the first bit represents the unit on/off
status, whilst the latter nine bits encode the power output level of that unit. These
nine bits could easily be decoded and mapped into each unit’s own operating range
according to the following equation [28]:
Dcd
Mapped value ¼ L
Pmax
i Pmin
i þ Pmin
i (16.16)
2 1
where
Dcd : the decoded value;
L: string length;
Pmax
i ; Pmin
i : maximum and minimum capacities of the corresponding generating
unit, respectively.
By modifying ten bits for every generator, the GA effectively searches for the
optimum solution in the combined UC and ED search space. Also, to a certain
extent, its computational time is considerably higher than the techniques currently
being used. However, one major advantage of the GA approach to solve the ED
problem lies in the fact that it can solve highly non-linear and discontinuous pro-
blems without the approximation needed in the classical problem. In addition, with
the use of multiprocessors, its computational time can be very significantly reduced
and becomes comparable to the time it takes to solve the problem using classical
approaches.
One approach to reduce the computational time of the GA is to classify the
generating units into three categories: base, cycling and peaking units.
Numerical example 16.1

The hybrid technique based on ANN with GA and PL (ANN/GA/PL) is
examined on the 26-unit IEEE reliability system taking into consideration unit and
system constraints. The transmission network of the IEEE reliability system con-
sists of 24-bus bars connected by 38 lines and transformers, Appendix B [12,29,30].
A quadratic cost function and exponential start-up cost are considered to meet the
system and unit constraints. To show the effect of line flow limits on generation
scheduling, flow limits on all lines are reduced to 300 MW.
In both problem formulations, a 5% load forecast error is aggregated into the
ANN training data. Thus, during training the ANN, 11 samples of load curves and
its corresponding commitment schedule are used. These samples are the original
load curve signified as ‘S’ p.u. plus five other derived samples above and below the
original load curve. These derived samples are denoted by ‘S x %’, where x ¼ 1,
2, 3, 4 and 5. The UC schedules used as training data are obtained by GA.
The basic design of a GA solution method involves making an optimal choice
of the various GA parameters, once the appropriate GA model has been chosen.
This GA model is composed of the basic sequential processes of fitness evaluation,
selection, recombination and mutation. Different GA control variables are tested,
and a satisfactory performance is found with the following parameters:
Population size: 100.

Maximum number of generations: 200.
Two-point crossover is used in this study with a probability, Pcrossover ¼ 0:9.
The mutation probability is not fixed; it begins at 0.04, and then it is changed
according to an exponential function as
Pmutation ¼ 1 þ ðexpð0:5 generation=max generationÞ þ 0:04Þ (16.17)
The GA and PL can be programmed in C language.
Units 17–19 and 24–26 of the IEEE system constitute base-load units, which
are responsible for a large portion of system generation; Units 1–9 constitute
peaking units; and lastly the other units represent cycling units. The values of
penalty factors a1 ; . . . ; a5 are taken as 8,000, 8,000, 5,000, 5,000 and 8,000,
respectively.
Generally, optimal power flow (OPF) is a tool to regulate the generator active
power outputs, voltage, shunt capacitors/reactors, transformer tap-setting and other
controllable variables to minimize the fuel cost and the network active power los-
ses. This should also keep the load bus voltage, generator reactive power outputs,
network power flows and other state variables in power system within their
operational and secure limits. The generators active power is considered as the only
control variable, which is used to minimize the total fuel cost while keeping system
and units constraints within their limits. However, other control variables such as
voltage magnitude and reactive power generation at each bus are ignored for sim-
plification purposes.
The system operational cost with transmission constraints and transmission
line losses using AC load flow program is $763,931. By including the ramp rate
Table 16.1 Unit scheduling generated by the ANN/GA/PL with transmission line,
ramp rate constraints and transmission line losses
Unit Hour (1–24)

1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 1 1 1 1 1 0
2 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 1 1 1 1 0
3 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 1
4 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0
5 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0
6 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 1 0 0
7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0
8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
10 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
11 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
12 1 1 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
13 1 1 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
14 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
15 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 1 1 1 1 1 0 0 0
16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1
17 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
18 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
19 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
20 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
21 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
22 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0
23 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 0 0
24 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
25 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
26 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
constraints into the generation scheduling problem, the system operational cost
goes up to $791,332 and the committed schedule is shown in Table 16.1. As the
transmission line losses are imposed into the generation scheduling program, a
solution which satisfies all required system constraints is reached. In this case, the
system operational cost increases even further to $791,332.
16.7 SCGS for smart grid incorporating wind and

solar thermal energy
With increase of the penetration of wind and PV power generation into power
systems, more innovative and sophisticated approaches to system operation need to
be adopted due to the intermittency and unpredictability of wind and PV power
generations.
In the SCGS problem, the main objective is to efficiently determine the com-
mitment of generating units as well as the output level of each committed unit to
meet the forecasted demand at the minimum possible production cost. Furthermore,
this should satisfy generating units and system constraints in a scheduling horizon
over one day. The stochastic nature effect of wind power generation and PV is
considered. The objective function is to minimize the total system generation cost
of thermal, wind and PV units across the schedule horizon. It can be formulated as
" !#
X
T X N X M X S
Minimize OC Pij þ OC wij þ OCðpvij Þ (16.18)
j¼1 i¼1 i¼1 i¼1

where T is the time horizon, which is 24 h in this study; OC Pij , OC wij and
OCðpvij Þ are the operating cost of the ith thermal unit, wind farm and PV modules
in jth hour, respectively; N, M and S are the number of thermal units, wind farm and
PV modules, respectively.
It is to be noted that (16.18) comprises three terms to, respectively, represent
(i) the operating cost of thermal units, (ii) the combined operating cost of wind farm
and (iii) the combined operating cost of PV modules.
P
Calculation of thermal units operating cost, Ni¼1 OC Pij : It can be calcu-
lated using (16.15).
PM
Calculation of wind farm operating cost, i¼1 OC wij : The wind farm
generation cost includes the following:

● A cost based on wind speed profile, Cwi wij , which equals the product of a
cost coefficient, di, and Weibull pdf of wind power, fw ðwÞ. Thus,

Cwi wij ¼ di fw ðwÞwij (16.19)
where di is the cost coefficient for ith wind farm and wij is the scheduled output
of ith wind farm in jth hour.
● Penalty cost for not using all the available wind power, Cp;wi . It is assumed to
be linearly proportional to the difference between the actual and scheduled
wind powers and can be expressed as follows [31]:
ð wri

Cp;wi ¼ kp;wi Wij;av wij ¼ kp;wi w wij fw ðwÞdw (16.20)
wi
where kp;wi is the penalty cost coefficient for over generation of ith wind farm,
Wij;av is the actual or available wind power from ith wind farm and wri is the
rated output of ith wind farm.
● Penalty reserve requirement cost, which is due to that the actual or available
wind power, is less than the scheduled wind power and can be given by
ð wi

Cr;wi ¼ kr;wi wij Wij;av ¼ kr;wi wij w fw ðwÞdw (16.21)
0
where kr;wi is the reserve cost coefficient for under generation of ith wind
farm [18].
From (16.19), (16.20) and (16.21), the total wind farm cost can be deduced as
X
M X M X M
OC wij ¼ Cwi wij þ Cp;wi Wij;av wg
i¼1 i¼1 i¼1
X
M
þ Cr;wi wij Wij;av (16.22)
i¼1
PS
Calculation of total operating cost of PV modules, i¼1 OCðpvij Þ: Like to
what is followed for wind farm costs, the total operating cost of the PV
modules includes the following:
• A weighted cost function, which represents the cost based on solar
irradiance profile. The weight is the Weibull pdf of PV power and can
be expressed as
Cpvi ðpvij Þ ¼ hi fpv ðpvÞðpvi Þ (16.23)
where hi and pvi are the cost coefficient and the scheduled output from
ith PV module in jth hour.
• The penalty cost for not using all the available solar power, which is
linearly related to the difference between available or actual solar power
and scheduled solar power. It can be formulated as
Cp;pv ¼ kp;pvi ðPV ij;av pvij Þ
ð pvðktmax Þ (16.24)
¼ kp;pvi ðpv pvij Þfpv ðpvÞdpv
pvi
where kp;pvi is the penalty cost coefficient for over generation of ith PV
module, PVij;av is the actual or available PV power from the ith PV
module, which is a random variable, and pvðktmax Þ is the maximum
output of ith PV module.
• The penalty reserve requirement cost, which is due to that the actual or
available PV power, is less than the scheduled PV power and can be
given by
Cr;pvi ¼ kr;pvi ðpvij PV ij;av Þ
ð pvi (16.25)
¼ kr;pvi ðpvij pvÞfpv ðPV Þdpv
0
where kr;pvi is the reserve cost coefficient for under generation of the ith
PV module.
From (16.23) to (16.25), the total operating cost of PV modules can be

obtained as follows:
X
S X
S X
S
OCðpvij Þ ¼ Cpvi ðpvij Þ þ Cp;pv ðPV ij;av pvij Þ
i¼1 i¼1 i¼1
X
S
þ Cr;pvi ðpvij PV ij;av Þ (16.26)
i¼1
Therefore, the objective function given by (16.18) can be determined in terms

of (16.15), (16.22) and (16.26) by which the total operating cost of thermal, wind
and PV generation, respectively, are computed.
The following are the system and unit constraints that are considered for sol-
ving the optimization problem.
1. Real power balance constraint: The total thermal, wind and PV power must
match the total load demand (PDj), the power losses (Plossj) in transmission
lines and spinning reserve (Rj Þ. Thus,
X
N X
M X
S
uij Pij þ uij wij þ uij pvij ¼ PDj þ Plossj þ Rj (16.27)
i¼1 i¼1 i¼1
where uij is the status index of ith unit in the jth period (1 for up and 0 for
down). Plossj can be computed using Newton–Rapson AC power flow.
2. In addition to the constraints given by (16.5), (16.7)–(16.9), (16.11) and
(16.12), the real power operating limits of thermal, wind and PV generating
units are
9
Pmin
i Pi Pmax
i =
0 wi wr (16.28)
;
0 pvi pvðktmax Þ
Numerical example 16.2

As a supplementary part to Example16.1, two scenarios are considered. In the
first, the stochastic nature of both the wind and PV power generators are not con-
sidered in the SCGS problem. In this case, the operating cost of the wind and PV
power generators are represented by (16.19) and (16.23), respectively. The second
scenario considers the stochastic nature of wind and PV power outputs and the
imbalance charges due to the mismatch between the actual and scheduled wind and
PV power outputs. Thus, the operating cost of wind and PV power generators are
calculated by (16.22) and (16.26), respectively.
In both scenarios, 26 thermal units, 2 wind farms and 2 PV power plants are
considered. The wind speed parameters vi , vr and vo are 5, 15 and 45 m/s, respec-
tively. The Weibull function parameters of wind are k ¼ 2 and c ¼ 10 m/s [32,33].
The penalty and reserve factors are set to kp;wi ¼ 2 and kr;wi ¼ 4. The clearness
index parameter ktmax ¼ 0:755 and the inclination b ¼ 0.
The Weibull function parameters of PV are c0 ¼ 0.6251 and l ¼ 3.5507 [34].
The penalty and reserve factors are set to kp;pvi ¼ 3 and kr;pvi ¼ 4. The cost coeffi-
cients of the two wind farms and two PV modules are d1 ¼ 1:0; d2 ¼ 1:25 and
h1 ¼ 1:3; h2 ¼ 1:4, respectively [32].
To test the effectiveness of both the wind and PV on the SCGS problem, two
wind farms and two PV modules are placed at Bus 15, as this bus contains most of
peaking units, shown in Figure 16.5. The total installed conventional capacity on
the system is approximately 3,225 MW and the assumed peak and minimum loads
for the test system are 2,850 and 1,842 MW, respectively. By applying the hybrid
technique comprising of ANN/GA/PL and taking into consideration system and
unit constraints including the ramp rate and transmission constraints, the total
operating costs for both scenarios are $785,660.44 and $784,953.31, respectively.
The solution times for both scenarios are 53 and 75 s, respectively. The overall
saving for daily power system operation for both scenarios, compared to the cost of
the same problem without the consideration of WP and PV power generators, are
$37,013.56 and $37,720.69, respectively. This amounts to a net saving over the
year of approximately $13.3 million and $13.6 million, respectively. A comparison
Bus 17 Bus 18
Bus 21 Bus 22
Bus 23
Bus 16
Bus 19 Bus 20
Bus 15
Bus 14 Bus 13
Bus 24 Bus 11 Bus 12
Bus 3 Bus 9 Bus 10

Bus 6
Bus 4
Bus 5
Bus 8
Bus 7
Bus 1 Bus 2
Figure 16.5 Modified IEEE 24-bus system

of flow on the line from Bus 14 to Bus 16, with and without transmission con-
straints, is given in Table 16.2. A comparison between the hourly cost without
embedding wind and PV in the SCGS problem and Scenarios 1 and 2 is demon-
strated in Figure 16.6. The second scenario which considers the stochastic nature of
both the wind and PV power generators has the smallest cost. It takes more
Table 16.2 Power flows on the line from Bus 14 to Bus 16 (MW)
Hour Without With line limits

line limits
Without wind With wind and With wind and PV
and PV PV first scenario second scenario
1 314.0 300.0 175.3 12.9
2 335.2 300.0 225.7 117.0
3 323.3 298.0 167.9 271.3
4 322.4 300.0 229.6 112.8
5 300.2 297.8 65.0 240.8
6 328.8 299.4 252.4 248.2
7 323.5 299.8 203.4 215.7
8 339.0 300.0 263.4 254.7
9 307.1 298.8 4.7 112.2
24 317.4 299.1 78.8 5.4
× 104
5
Without wind and PV
Scenario 1
4.5 Scenario 2
4
Hourly cost
3.5
2.5
2
2 4 6 8 10 12 14 16 18 20 22 24
Time
Figure 16.6 Comparison between the hourly cost of the SCGS problem with and
without embedding wind and PV generators
× 104
5
Case 1
Case 2
4.5
4
Hourly cost
3.5
2.5
2
2 4 6 8 10 12 14 16 18 20 22 24
Time
Figure 16.7 Comparison between the hourly cost of the SCGS problem for
Cases 1 and 2
Table 16.3 UC with wind and PV (Scenario 2), Case 2
Unit Hour (1–24)

1 1 0 0 1 1 1 1 1 1 1 0 0 0 1 1 1 0 1 0 1 0 0 0 1
2 1 0 0 0 1 0 1 0 1 0 0 0 0 0 1 0 1 0 0 1 0 0 0 1
3 1 0 0 1 1 1 0 0 0 0 0 0 0 0 1 1 0 0 0 1 0 0 0 1
4 1 0 0 1 1 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0
5 1 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
6 1 1 1 1 0 0 0 1 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
7 1 1 0 1 1 1 0 1 0 0 0 1 1 0 0 1 1 1 1 1 0 1 1 0
8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
10 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
11 1 1 1 1 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
12 1 1 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
13 1 0 0 0 0 0 0 0 1 1 0 0 1 1 1 1 1 1 1 1 1 1 1 1
14 1 1 1 1 1 1 0 0 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1
15 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 1 1 1 1 0 0 0
16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0
17 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
18 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
19 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
(Continues)
Table 16.3 (Continued)
Unit Hour (1–24)

20 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
21 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
22 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0
23 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0
24 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
25 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
26 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
27 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
28 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
29 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0
30 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0
3,200
Load demand
Generation (Case 1)
3,000 Generation (Case 2)
2,800
2,600
Power
2,400
2,200
2,000
1,800
2 4 6 8 10 12 14 16 18 20 22 24
Time
Figure 16.8 Daily load demand and scheduled generation (Cases 1 and 2)
scheduling time than that of the first scenario but still is acceptable for real-time
operation.
To ensure that the system security is indeed maintained and to test the feasi-
bility of the hybrid technique, the second scenario is applied to the same system
with different locations of wind and PV power generators taking into consideration
all the system and unit constraints. First, the same two wind farms and the two PV
modules are inserted on Bus 14 (instead of Bus 15) and the SCGS problem is solved
using the hybrid technique with the influence of both the ramp rate and transmis-
sion constraints. The system operational cost of $784,732.63 with no violation is
reported.
Second, after inserting the renewable generators on Bus 16, the system
operational cost is $784,076.81, with also no violation reported. The hourly cost of
both cases is illustrated in Figure 16.7. The committed schedule of the second case
is given in Table 16.3. The numbers (1 or 0) in this table represent on/off states of
units at different hours. Units 24–26 constitute base load units, which are respon-
sible for a large portion of system generation, whereas Units 27 and 28 represent
the wind farm units and 29 and 30 represent the PV modules generating units.
Figure 16.8 shows load demand and generating capability for 24 h, Cases 1 and 2.
References
[1] Marwali M. K. C. and Shahidehpour S. M. ‘Coordination between long-term
and short-term generation scheduling with network constraints’. IEEE
[2] El-Hawary M. E. and Christensen G. S. Optimal economic operation of
electric power systems. Mathematics in Science and Engineering series, no.
142. New York: Academic Press; 1979.
[3] Wood A. J. and Wollenberg B. F. Power generation operation and control.
New York: John Wiley & Sons; 1996.
[4] Yalcinoz T. and Short M. J. ‘Large-scale economic dispatch using an
improved Hopfield neural network’. IEE Proceedings of Generation,
Transmission and Distribution. 1997;144(2):181–5.
[5] Glover F. ‘Tabu search – Part I’. ORSA Journal on Computing. 1989;1(3):
190–206.
[6] Mantawy A. H., Abdel-Magid Y. L. and Selim S. Z. ‘Unit commitment by
tabu search’. IEE Proceedings of Generation, Transmission and Distribution.
1998;145(1):56–64.
[7] IEEE Report. ‘Potential impacts of clean air regulations on system opera-
tions’. IEEE Transactions on Power Systems. 1995;10(2):647–53.
[8] Sheble G. B. and Brittig K. ‘Refined genetic algorithm – Economic dispatch
example’. IEEE Transactions on Power Systems. 1995;10(1):117–24.
[9] Saavedra O. R. and Gomes J. R. ‘Optimal reactive power dispatch using
evolutionary computation: Extended algorithms’. IEE Proceedings of
[10] Liang Z. X. and Glover J. D. ‘A zoom feature for a dynamic programming
solution to economic dispatch including transmission losses’. IEEE
[11] Tang S. K. and Shahidehpour S. M. ‘A combination of Lagrangian relaxa-
tion and linear programming approaches for fuel constrained unit commit-
ment problem’. IEE Proceedings Part C. 1989;136(3):162–74.
[12] Wang S. J., Shahidehpour S. M., Kirschen D. S., Mokhtari S. and Irisarri G. D.
‘Short-term generation scheduling with transmission and environmental con-
straints using an augmented Lagrangian relaxation’. IEEE Transactions on
Power Systems. 1995;10(3):1294–301.
[13] Cohen A. I. and Yoshimura M. ‘A branch-and-bound algorithm for unit
commitment’. IEEE Transactions on Power Apparatus and Systems. 1983;
PAS-102(2):444–51.
[14] Dillon T. S., Edwin K. W., Kochs H. D. and Taud R. J. ‘Integer program-
ming approach to the problem of optimal unit commitment with probabilistic
reserve determination’. IEEE Transactions on Power Apparatus and
Systems. 1978;PAS-97(6):2154–66.
[15] Mokhtari S., Sing J. and Wollenberg B. ‘A unit commitment expert system’.
[16] Bansal R. C. ‘Bibliography on the fuzzy set theory applications in power
systems (1994–2001)’. IEEE Transactions on Power Systems. 2003;
18(4):1291–9.
[17] Su M. C. and Hs Y. Y. ‘Fuzzy dynamic programming: An application to unit
commitment’. Transactions on Power Systems. 1991;6(3):1231–7.
[18] Mantawy A. H., Abdel-Magid Y. L. and Selim S. Z. ‘A simulated annealing
algorithm for unit commitment’. IEEE Transactions on Power Systems.
1998;13(1):197–204.
[19] Wong K. P. and Fung C. C. ‘Simulated annealing based economic dispatch
algorithm’. IEE Proceedings Part C. 1993;140(6):509–15.
[20] Sasaki H., Watanabe M. and Yokoyama R. ‘A solution method of unit
commitment by artificial neural networks’. IEEE Transactions on Power
Systems. 1992;7(3):974–81.
[21] Ouyang Z. and Shahidehpour S. M. ‘A hybrid artificial neural network-
dynamic programming approach to unit commitment’. IEEE Transactions
on Power Systems. 1992;7(1):236–42.
[22] Lai L. L. Intelligent system applications in power engineering evolutionary
programming and neural network. New York: John Wiley & Sons; 1999.
[23] Kazarlis S. A., Bakirtzis A. G. and Petridis V. ‘A genetic algorithm solution
to the unit commitment problem’. IEEE Transactions on Power Systems.
1996:11(1):83–92.
[24] Orero S. O. and Iriving M. R. ‘Scheduling of generators with a hybrid
genetic algorithm’. Proceedings of 1st International Conference on Genetic
Algorithms in Engineering Systems: Innovations and Applications
(GALESIA); Sheffield, UK, Sep 1995, IET; 1995, pp. 200–6.
[25] El Desouky A. A., Aggarwal R., Elkateb M. M. and Li F. ‘Advanced hybrid
genetic algorithm for short-term generation scheduling’. IEE Proceedings of
[26] El Desouky A. A. and ElKateb M. M. ‘Hybrid adaptive techniques for
electric-load forecast using ANN and ARIMA’. IEE Proceedings of
[27] Bunn D. H. and Farmer E. D. Comparative models for electrical load fore-
casting. New York: John Wiley & Sons; 1985.
[28] Li F. and Aggarwal R. K. ‘Fast and accurate power dispatch using a relaxed
genetic algorithm and a Local gradient technique’. Expert Systems with
Applications. 2000;19(3):159–65.
[29] Reliability Test System Task Force. ‘IEEE reliability test system’. IEEE
Transactions on Power Apparatus and Systems 1979;PAS-98(6):2047–54.
[30] Tong S. K., Shahidehpour S. M. and Ouyang Z. ‘A heuristic short-term unit
commitment’. IEEE Transactions on Power Systems. 1991;6(3):1210–16.
[31] El Desouky A. A. ‘Security and stochastic economic dispatch of power sys-
tem including wind and solar resources with environmental consideration’.
International Journal of Renewable Energy Research. 2013;3(4):951–58.
[32] Orgill J. F. and Hollands K. T. G. ‘Correlation equation for hourly diffuse
radiation on horizontal surface’. Solar Energy. 1977;19(4):357–9.
[33] Eladany M. M., El-Desouky A. A., Daoud A. A. and Sallam A. A.
‘Optimizing stochastic economic dispatch of wind/solar DG grid’.
International Journal of Distribution Energy Resources. 2012;8(4):100–12.
[34] Tina G. and Gagliano S. ‘Probabilistic analysis of weather data for a hybrid
solar/wind energy system’. International Journal of Energy Research.
2011;35(3):221–32.
Chapter 17
Load flow analysis for power systems
with renewables
17.1 Introduction
The techniques of load flow analysis are of utmost importance to analyse, plan and
design the power system. Power flow analysis defines the power system parameters
at normal and abnormal operating conditions through the calculation of current or
power flow in the lines of the interconnected transmission network, voltages at all
nodes of the power system and the power delivered from generating sources.
Several load flow methods, such as Newton–Raphson, fast-decoupled Newton–
Raphson, Gauss–Seidel, etc., have been developed for conventional power systems.
Based on load flow analysis, protection system and capacity of the power system,
assets can be specified, as well as reliability, stability and security can be analysed
and enhanced.
Renewable energy sources (RESs) have become an important part of electrical
generation and its importance is continuing to increase. But incorporating such
sources (e.g. wind farms and photovoltaic (PV) plants) in electrical power system
causes some problems because of the fluctuation of power generated by such
sources. Forecasting the power generated from renewable sources to predict power
output in advance through methods based on numerical weather prediction models
and statistical approaches is a recently developed tool for load forecasting, but with
a level of accuracy not as great as for the conventional load forecasting techniques.
The overall shape of renewable production can be predicted most of the time but
with significant errors that can occur in both the level and timing. However,
renewable energies forecast for shorter time horizons tend to be more accurate than
forecasts for longer time; that is, forecasts for 1–2 h ahead can achieve an accuracy
much better than that of one day ahead [1].
Another method, probability distribution, rather than the prediction approach,
can be used. For example, the wind speed or solar radiation for PV may be difficult
to predict, but the probability of a specific wind speed or specific solar radiation
occurring can be estimated. This can be done if the probability distribution is
known by assuming it to be a Weibull or Rayleigh distribution. Then, once the wind
speed or solar radiation is known, the injected power from the wind farm or PV plant
into the grid can be calculated by means of the corresponding power curves [2].
In addition, the renewable sources usually are associated with energy storage
systems (ESSs). So, the methods of load flow analysis should be developed in a
way that can meet such situations.
17.2 Steady-state load flow calculations

To calculate the load flow in a power system incorporating RESs and ESS and
operating in normal conditions, some features must be taken into consideration.
These include the following:
● The conventional generating units have a constant power rating. The power
delivered from each unit to the power system depends on the system topology,
system losses and load demand, up to the rating of the generating unit, Pg.
● The available capacity of RESs is climate dependent, and consequently it is
time dependent, Pren(t). At a prespecified time, the power injected by the RES
into the transmission network corresponds to the load demand, system losses
and state of ESS, but is restricted by the available capacity at that time.
● The ESS can inject power into the transmission system during its discharging
state and receive power from the system during its charge state, PS, depending
on the load demand and the generation over the entire power system.
● The total load is, of course, time variant and its profile is given, PL(t).
A schematic diagram of the power system configuration as well as the direc-
tion of power flow regarding the generating units (conventional and renewable),
ESS and loads are illustrated in Figure 17.1.
The charging/discharging power of the ESS can be calculated as the difference,
DP; between the total power generation (conventional and renewable), PGen, and
the power consumption including the system losses, Plosses. Thus,
RESs ESS
Pren(t) Discharging
Ps Charging
PL(t)
generating units
Conventional
Loads
Pg
Transmission
network
Figure 17.1 Schematic diagram of a power system with RESs and ESS
Load flow analysis for power systems with renewables 433
DP ¼ PGen fPL ðtÞ þ Plosses g (17.1)

and
X
N X
M
PGen ¼ Pg þ Pren;i ðtÞ (17.2)
g¼1 i¼1
where N and M are the number of conventional generating units and RESs,
respectively, and Pg represents the available capacity of the gth generator.
Hence,
9
Ps DP for DP > 0 for charging state ðESS as a loadÞ =
¼ DP for DP < 0 for discharging state ðESS as a sourceÞ (17.3)
;
¼ 0 for DP ¼ 0 for system balance state
According to (17.3), the power balance condition can be expressed as follows:
PGen ¼ PL ðtÞ þ Plosses þ PS ffor ESS charging stateg (17.4)
and
PGen þ PS ¼ PL ðtÞ þ Plosses ffor ESS discharging stateg (17.5)
As the ESS acts as a load or as a source of power depending on its state that is
defined by the sign of DP, it can be represented by a bus (ESS bus) connected to the
power system through a very low reactance not to affect the voltages of different
buses in the system. The ESS bus is defined as a load bus in the charging state or
PV bus in the discharging state in the load flow studies.
17.2.1 Load flow algorithm

The algorithm of load flow program for a time span of T and starting at time t0
includes the following steps:
(i) Input the system data incorporating renewables (wind and/or solar PV).
(ii) Specify the time in hour, t1 .
(iii) Read the wind speed and/or solar radiation at the given time.
(iv) Compute the corresponding power generated from wind and/or PV.
(v) Compute the value of DP to decide the state of the ESS (negative value for
ESS acting as a source, ‘discharge state’, and positive value for ESS acting
as a load, ‘charging state’, or zero value at which the power is already
balanced, i.e. no need to charge/discharge the ESS).
(vi) Proceed with the load flow calculations.
(vii) Output the results: Line flows, bus voltages, power generated from each
source, system losses and the role of ESS.
(viii) Advance the time incrementally to t1 þ Dt and repeat Steps (iii)–(vii) until
reaching the end of time span T.
Start
Input system
data
Specify the time, t = t1
R Load P (t)
L
Wind speed and/or
irradiation
Compute the wind and/or

PV-generated power
Compute ∆P
No and < 0 Is Yes

∆P > 0?
No and = 0
ESS in discharge state ESS in charge state

(acting as a source) (acting as a load)
System in balance
state
Proceed the load

flow calculations
Output the results
t = t + ∆t
Is No
t > T?
Yes
End
Figure 17.2 Flow chart for load flow calculations
Accuracy of the results increases by decreasing both the incremental time Dt

and prediction error of renewables generation [3]. A flow chart indicating the
sequence of these steps is shown in Figure 17.2.
Example 17.1 IEEE 14-bus test system (Figure 17.3) and its data are given in
Tables 17.1–17.3. It is desired to perform load flow calculations and determine the
Bus 13 9
Bus 12 3 Bus 14
8
4
Bus 11 6 7
2 5 Bus 9
Bus 10
ESS 10 Bus 8
Bus 6 19 21
G1
Bus 15 Bus 7 G5
17 G4 20
Bus 1 18 Bus 4
14
15
Bus 5
1 13
11 16
Bus 2 12 Bus 3
G2 G3
Figure 17.3 Single-line-diagram of IEEE 14-bus system (the dotted lines

represent the connections of energy sources that can replace the
conventional units)
Table 17.1 Generation data
Generator 1 (G1) 2 (G2) 3 (G3) 6 (G4) 8 (G5)

bus #
MVA 615 60 60 25 25
PF 0.989 (lead) 0 0.89 (lag) 0 0
Under-excited Overexcited Overexcited Overexcited Overexcited
XL (p.u.) 0.2396 0.00 0.00 0.134 0.134
Ra (p.u.) 0.00 0.0031 0.0031 0.0014 0.0041
Xd (p.u.) 0.8979 1.05 1.05 1.25 1.25
X 0d (p.u.) 0.2995 0.1850 0.1850 0.232 0.232
X d00 (p.u.) 0.23 0.13 0.13 0.12 0.12
0
Tdo (p.u.) 7.4 6.1 6.1 4.75 4.75
00
T do (p.u.) 0.03 0.04 0.04 0.06 0.06
Xq (p.u.) 0.646 0.98 0.98 1.22 1.22
Xq0 (p.u.) 0.646 0.36 0.36 0.715 0.715
X q00 (p.u.) 0.4 0.13 0.13 0.12 0.12
0
Tqo (p.u.) 0.00 0.099 0.099 0.21 0.21
00
T qo (p.u.) 0.033 0.099 0.099 0.21 0.21
H 5.148 6.54 6.54 5.06 5.06
D 2 2 2 2 2
Bus type: (1) swing bus, (2) generator bus (PV bus) and (3) load bus (PQ bus).
charging/discharging modes of the ESS through 24 h for three cases: (i) the con-
ventional generator at Bus #1 is replaced by a renewable wind farm and ESS; (ii)
the conventional generator at Bus #2 is replaced by a solar PV plant and ESS; and
(iii) the generator at Bus #1 is replaced by a wind farm and ESS as well as the
Table 17.2 Bus data
Bus # Pg Qg PL QL Bus type Qg max. Qg min.

(p.u.) (p.u.) (p.u.) (p.u.) (p.u.) (p.u.)
1 2.32 0.00 0.00 0.00 1 10.0 10.0
2 0.4 0.424 0.2170 0.1270 2 0.5 0.4
3 0.00 0.2737 0.9420 0.1900 2 0.4 0.00
4 0.00 0.00 0.4780 0.00 3 0.00 0.00
5 0.00 0.00 0.0760 0.0160 3 0.00 0.00
6 0.00 0.2251 0.1120 0.0750 2 0.24 0.06
7 0.00 0.00 0.00 0.00 3 0.00 0.00
8 0.00 0.2516 0.00 0.00 2 0.24 0.06
9 0.00 0.00 0.2950 0.1660 3 0.00 0.00
10 0.00 0.00 0.0900 0.0580 3 0.00 0.00
11 0.00 0.00 0.0350 0.0180 3 0.00 0.00
12 0.00 0.00 0.0610 0.0160 3 0.00 0.00
13 0.00 0.00 0.1350 0.0580 3 0.00 0.00
14 0.00 0.00 0.1490 0.0500 3 0.00 0.00
Table 17.3 Line data
From To Line # Resistance Reactance Line Tap

bus bus (p.u.) (p.u.) charging ratio
(p.u.)
1 2 11 0.01938 0.05917 0.0528 1
1 5 14 0.05403 0.22304 0.0492 1
2 3 12 0.04699 0.19797 0.0438 1
2 4 16 0.05811 0.17632 0.0374 1
2 5 1 0.05695 0.17388 0.034 1
3 4 13 0.06701 0.17103 0.0346 1
4 5 15 0.01335 0.04211 0.0128 1
4 7 20 0.00 0.20912 0.00 0.978
4 9 19 0.00 0.55618 0.00 0.969
5 6 18 0.00 0.25202 0.00 0.932
6 11 5 0.09498 0.1989 0.00 1
6 12 2 0.12291 0.25581 0.00 1
6 13 4 0.06615 0.13027 0.00 1
7 8 21 0.00 0.17615 0.00 1
7 9 10 0.00 0.11001 0.00 1
9 10 7 0.03181 0.08450 0.00 1
9 14 8 0.12711 0.27038 0.00 1
10 11 6 0.08205 0.19207 0.00 1
12 13 3 0.22092 0.19988 0.00 1
13 14 9 0.17093 0.34802 0.00 1
7.5
7
6.5
Load (p.u.)
6
5.5
5
4.5
4
3.5
0 5 10 15 20 25
Time (h)
Figure 17.4 Load profile through 24 h
Table 17.4 Load profile data
Time (s) Load (p.u.) Time (s) Load (p.u.)

base 100 MVA base 100 MVA
1 4.3513 13 5.4390
2 4.2172 14 5.8016
3 4.0612 15 5.9467
4 3.9159 16 6.5267
5 3.6260 17 6.8894
6 3.9159 18 7.2520
7 4.0612 19 6.8894
8 4.2172 20 6.5267
9 4.3513 21 5.9467
10 4.6413 22 5.8016
11 4.7863 23 5.0764
12 5.0764 24 4.7863
generator at Bus #2 is replaced by a solar PV plant. The load profile is drawn in

Figure 17.4 and its data are given in Table 17.4. The time span is 1 day starting at
t ¼ 1 am, and incrementally increases by 1 h. The base power for per unit values is
100 MVA.
Solution
Case (i): Replacing G1 by a wind farm and an ESS at Bus #1
Data of wind energy farm are given in Table 17.5. Wind speed and its respective
power, computed as explained in Chapter 2, through the 24 h are plotted in
Figures 17.5 and 17.6, respectively. The variation of DP and Plosses through 24 h is
computed and depicted in Figures 17.7 and 17.8, respectively.
By using the load flow algorithm, the calculations are implemented 24 times
over 24 h, that is, one each hour. As a sample of the results obtained by load flow
calculations, the results of line flows at which DP is maximum positive and
maximum negative, t6 and t18, respectively, are selected and tabulated in
Table 17.5 Wind farm data
Power, voltage and frequency rating (MVA, kV, Hz) [610 69 60]
Stator resistance Rs and reactance Xs (p.u., p.u.) [0.01 0.10]
Rotor resistance Rr and reactance Xr (p.u., p.u.) [0.01 0.08]
Magnetization reactance Xm (p.u.) 3.00
Number of poles P and gear box ratio [4 1/89]
Blade length and number (m integer) [75 3]
Number of wind generators that compose the farm 30
Pitch control gain and time constant Kp, Tp (p.u., s) [10 3]
17.5
17
16.5
Wind speed (m/s)
16
15.5
15
14.5
14
13.5
0 5 10 15 20 25
Time (h)
Figure 17.5 Wind speed versus day hours
6.5
Wind-generated power (p.u.)
6
5.5
5
4.5
4
3.5
3
0 5 10 15 20 25
Time (h)
Figure 17.6 Wind-generated power with respect to the speed variation

versus hours
Tables 17.6–17.9. The ESS bus is denoted by Bus #15, which is connected to
Bus #1 through a reactance of 0.00271 p.u.
Case (ii): Power system incorporating a solar PV plant and an ESS at Bus #2
The procedure of load flow calculations is the same as that used in Case
(i) incorporating a wind farm. The solar PV plant is connected to Bus #2, with the
following data:
2.5
ESS during charging state
2
1.5
1
∆P (p.u.)
0.5
0
–0.5
–1
–1.5
–2
ESS during discharging state
–2.5
0 5 10 15 20 25
Time (h)
Figure 17.7 DP versus day hours
1.6
1.2
Plosses (p.u.)
0.8
0.4
0
0 5 10 15 20 25
Time (h)
Figure 17.8 Power losses along 24 h
● The output power of each unit in the plant is 10 MW.

● Number of units used six, constituting three models.
● Active power of each model is 20 MW at a rated voltage of 1.045 p.u.
● Daily solar radiation curve is shown in Figure 17.9, and the corresponding
generated power is depicted in Figure 17.10.
Changes of DP and Plosses for 24 h, resulting from load flow calculations, are
shown in Figures 17.11 and 17.12, respectively. DP is maximum positive and
maximum negative at t5 and t18, respectively, and the corresponding line flows, and
bus voltage and power are tabulated in Tables 17.10–17.13.
Case (iii): The power system incorporates a wind farm and an ESS at Bus #1 and
a solar PV at Bus #2 without a conventional generating unit at these buses
With the same data as in the first and second cases, load flow calculations result in the
changes of DP and Plosses for 24 has shown in Figures 17.13 and 17.14. DP is max-
imum positive and maximum negative at t6 and t18, respectively, and the corre-
sponding line flows and bus voltage and power are tabulated in Tables 17.14–17.17.
Table 17.6 Line flows at t6
From bus To bus Line # P (p.u.) Q (p.u.) Plosses (p.u.) Qlosses (p.u.)
02 05 1 0.62426 0.08109 0.02084 0.02809
06 12 2 0.12309 0.04667 0.00186 0.00387
12 13 3 0.02903 0.02039 0.00025 0.00023
06 13 4 0.28102 0.14775 0.00582 0.01147
06 11 5 0.12787 0.13218 0.00281 0.00588
11 10 6 0.07217 0.1011 0.00118 0.00277
09 10 7 0.06526 0.01675 0.00014 0.00038
09 14 8 0.13003 0.00432 0.00211 0.00448
14 13 9 0.09738 0.07016 0.0025 0.00508
07 09 10 0.40945 0.22868 0 0.02264
01 02 11 2.6641 0.42397 0.1251 0.32346
03 02 12 1.0842 0.17793 0.05599 0.18964
03 04 13 0.3401 0.22551 0.01148 0.00543
01 05 14 1.2024 0.11556 0.07051 0.23888
05 04 15 0.91916 0.15627 0.0116 0.02386
02 04 16 0.84653 0.05945 0.03847 0.07783
15 01 17 1e–05 0.00841 0 0.0083
05 06 18 0.70128 0.06356 0 0.1087
04 09 19 0.23184 0.04332 0 0.02939
04 07 20 0.40945 0.0669 0 0.03484
08 07 21 0 0.34842 0 0.018
Table 17.7 Bus voltage and power at t6
Bus # Bus voltage Pg (p.u.) Qg (p.u.) PL (p.u.) QL (p.u.)
jVj (p.u.) Phase (rad)

01 1.06 0 6.03588 0.27831 2.4922 0
02 1.045 0.14977 0.4 1.0775 0.3281 0.1778
03 1.01 0.36367 0 0.67076 1.4243 0.266
04 0.99413 0.28943 0 0 0.7227 0.056
05 0.99926 0.24844 0 0 0.1149 0.0224
06 1.07 0.40311 0 0.4789 0.1693 0.105
07 1.0337 0.37101 0 0 0 0
08 1.09 0.37101 0 0.34915 0 0
09 1.0103 0.41416 0 0 0.446 0.2324
10 1.0097 0.42008 0 0 0.1361 0.0812
11 1.0342 0.41475 0 0 0.0529 0.0252
12 1.045 0.42615 0 0 0.0922 0.0224
13 1.0349 0.42735 0 0 0.2041 0.0812
14 0.99337 0.44865 0 0 0.2253 0.07
15 1.045 0 1e–05 0.000841 0 0
From bus To bus Line # P (p.u.) Q (p.u.) Plosses Qlosses

02 05 1 1.2068 0.35734 0.08333 0.22101
06 12 2 0.23178 0.06236 0.00618 0.01287
12 13 3 0.05479 0.02709 0.00078 0.0007
06 13 4 0.53615 0.21035 0.01916 0.03774
06 11 5 0.24485 0.20546 0.00848 0.01775
11 10 6 0.13837 0.16251 0.00366 0.00857
09 10 7 0.11793 0.07102 0.00065 0.00172
09 14 8 0.24064 0.01361 0.0079 0.01683
14 13 9 0.18447 0.10044 0.00852 0.01735
07 09 10 0.76122 0.28364 0 0.07352
01 02 11 5.9947 0.66673 0.62684 1.8554
03 02 12 2.0725 0.79697 0.22877 0.91755
03 04 13 0.56515 0.73375 0.05807 0.11555
01 05 14 2.4045 0.60112 0.29703 1.177
05 04 15 1.6918 0.40512 0.04622 0.13465
02 04 16 1.6522 0.32518 0.15162 0.4234
15 01 17 2.261 0.00083 0 0.13347
05 06 18 1.3264 0.05686 0 0.44205
04 09 19 0.42336 0.04679 0 0.10914
04 07 20 0.76122 0.12257 0 0.13698
08 07 21 0 0.59583 0 0.05264
Bus # Bus voltage Psupply (p.u.) Qsupply (p.u.) Pload (p.u.) Qload (p.u.)

01 1.06 0 6.1382 0.2248 0 0
02 1.045 0.33775 0.4 3.503 0.6076 0.1778
03 1.01 0.77694 5.85e–13 0.67076 2.6376 0.266
04 0.9317 0.62007 3.54e–12 4.15e–13 1.3384 0.056
05 0.93425 0.53198 1.2e–12 4.83e–12 0.2128 0.0224
06 1.07 0.84891 2.46e–12 0.4789 0.3136 0.105
07 0.99371 0.78903 7.89e–13 1.42e–13 0 0
08 1.09 0.78903 1.4e–15 0.34915 0 0
09 0.96599 0.87638 8.63e–13 1.75e–13 0.826 0.2324
10 0.9684 0.88945 1.15e–13 2.1e–14 0.252 0.0812
11 1.0104 0.87591 3.67e–14 1.1e–14 0.098 0.0252
12 1.0296 0.89579 1.05e–15 1.8e–16 0.1708 0.0224
13 1.0126 0.90056 1.2e–15 2.5e–15 0.378 0.0812
14 0.94068 0.94995 1.17e–13 3.1e–14 0.4172 0.07
15 1.045 0.05534 2.261 0.51583 0 0
1,000
900
800
Solar radiation (W/m2)
700
600
500
400
300
200
100
05 7 9 11 13 15 17 19
Time (h)
Figure 17.9 Solar radiation through day hours
12
10
Power generation (p.u.)
0
0 5 10 15 20 25
Time (h)
Figure 17.10 PV power versus day hours
3
2
1
∆P (p.u.)
–1
–2
–3
0 5 10 15 20 25
Time (h)

1.8
1.6
1.4
1.2
Plosses (p.u.)
1.0
0.8
0.6
0.4
0.2
0
0 5 10 15 20 25
Time (h)
2 5 1 0.37021 0.14556 0.00855 0.009419
6 12 2 0.11521 0.0457 0.00165 0.00343
12 13 3 0.02816 0.01987 0.00024 0.00022
6 13 4 0.26445 0.14438 0.00525 0.01033
6 11 5 0.12743 0.12796 0.00271 0.005567
11 10 6 0.07573 0.09709 0.00116 0.00272
9 10 7 0.05152 0.01294 9e–05 0.00023
9 14 8 0.11447 0.00579 0.00163 0.00347
14 13 9 0.09576 0.06768 0.00237 0.00482
7 9 10 0.36961 0.22715 0 0.01933
1 2 11 4.9378 0.65777 0.42735 1.2463
3 2 12 0.93624 0.10987 0.04118 0.12725
3 4 13 0.38256 0.23647 0.01386 0.00057
1 5 14 1.583 0.13704 0.12181 0.45065
5 4 15 1.0526 0.23731 0.01554 0.03628
2 4 16 0.62934 0.09407 0.02177 0.02711
15 1 17 2.2297 0.00618 0 0.01234
5 6 18 0.6639 0.05628 0 0.09739
4 9 19 0.20938 0.04143 0 0.024
4 7 20 0.36961 0.06814 0 0.0285
8 7 21 0 0.34104 0 0.01724
17.2.2 Role of the ESS

Distribution of DP through the 24 h for the three cases collected in one graph as
shown in Figure 17.15 as well as the corresponding Plosses of the system, which are
almost equal, is depicted in Figure 17.16. The time at which the value of DP in both
charging and discharging state of ESS is maximum is given in Table 17.18.

1 1.06 0.27833 6.15208 0.2248 0 0
2 1.045 0 0 2.345 0.3038 0.1778
3 1.01 0.27833 0 0.67076 1.3188 0.266
4 0.9957 0.46084 0 0 0.6692 0.056
5 0.9989 0.37875 0 0 0.1064 0.0224
6 1.07 0.47749 0 0.46416 0.1568 0.105
7 1.0349 0.45218 0 0 0 0
8 1.09 0.45218 0 0.34104 0 0
9 1.0115 0.49103 0 0 0.1568 0.2324
10 1.011 0.49569 0 0 0.126 0.0812
11 1.035 0.4894 0 0 0.049 0.0252
12 1.0461 0.4988 0 0 0.0854 0.0224
13 1.0353 0.49995 0 0 0.189 0.0812
14 0.99602 0.52103 0 0 0.2086 0.07
15 1.045 0.28387 1e–05 0.00618 2.2297 0
2 5 1 1.375 0.29915 0.10387 0.2835
6 12 2 0.23073 0.06221 0.00613 0.01276
12 13 3 0.05379 0.02705 0.00076 0.00068
6 13 4 0.53221 0.20914 0.01889 0.03721
6 11 5 0.23751 0.20379 0.00813 0.01702
11 10 6 0.13139 0.16157 0.00348 0.00814
9 10 7 0.12479 0.07038 0.0007 0.00185
9 14 8 0.24522 0.01302 0.00818 0.0174
14 13 9 0.18016 0.10042 0.00819 0.01668
7 9 10 0.76833 0.28316 0 0.07443
1 2 11 3.7014 0.56731 0.24129 0.67821
3 2 12 2.1156 0.83171 0.23978 0.96392
3 4 13 0.52198 0.6845 0.05028 0.09555
1 5 14 2.0061 0.46776 0.20533 0.79818
5 4 15 1.5451 0.30279 0.0373 0.1064
2 4 16 1.7713 0.30071 0.17244 0.48645
15 2 17 2.4692 0.00758 0 0.01513
5 6 18 1.3141 0.03438 0 0.42694
4 9 19 0.42767 0.05097 0 0.11069
4 7 20 0.76833 0.11294 0 0.13783
8 7 21 0 0.58459 0 0.0529

1 1.06 0 5.7076 0.2248 0 0
2 1.045 0.20864 0.18 2.163 0.6076 0.1778
3 1.01 0.65851 0 0.67076 2.6376 0.266
4 0.93552 0.51372 0 0 1.3384 0.056
5 0.94126 0.43525 0 0 0.2128 0.0224
6 1.07 0.74672 0 0.4789 0.3136 0.105
7 0.99553 0.68326 0 0 0 0
8 1.09 0.68326 0 0.34915 0 0
9 0.96797 0.77108 0 0 0.826 0.2324
10 0.9701 0.7847 0 0 0.252 0.0812
11 1.0114 0.77249 0 0 0.098 0.0252
12 1.0297 0.79336 0 0 0.1708 0.0224
13 1.013 0.79794 0 0 0.378 0.0812
14 0.94202 0.84568 0 0 0.4172 0.07
15 1.045 0.20251 2.4692 0.00758 0 0
3
2
1
∆P (p.u.)
–1
–2
–3
0 5 10 15 20 25
Time (h)
1.6
1.2
Plosses (p.u.)
0.8
0.4
0
0 5 10 15 20 25
Time (h)

2 5 1 0.59363 0.09061 0.019 0.02249
6 12 2 0.12325 0.04663 0.00186 0.00388
12 13 3 0.02919 0.02035 0.00026 0.00023
6 13 4 0.28166 0.14769 0.00584 0.01151
6 11 5 0.12913 0.13202 0.00283 0.00592
11 10 6 0.0734 0.10089 0.00119 0.0028
9 10 7 0.06403 0.01654 0.00014 0.00036
9 14 8 0.12925 0.00446 0.00208 0.00443
14 13 9 0.09813 0.06997 0.00252 0.00512
7 9 10 0.40818 0.22898 0 0.02256
1 2 11 3.0245 0.48083 0.16129 0.43394
3 2 12 1.0745 0.17331 0.05495 0.18523
3 4 13 0.34981 0.23145 0.01211 0.00382
1 5 14 1.2712 0.11891 0.07873 0.27283
5 4 15 0.94881 0.17004 0.01241 0.02642
2 4 16 0.82401 0.06477 0.03652 0.07191
15 1 17 1e–05 0.00841 0 0.0083
5 6 18 0.70334 0.06183 0 0.10939
4 9 19 0.2311 0.04312 0 0.02922
4 7 20 0.40818 0.06745 0 0.03466
8 7 21 0 0.34915 0 0.01807

1 1.06 0 6.0998 0.2248 2.1042 0
2 1.045 0.17027 0.012 1.2599 0.3281 0.1778
3 1.01 0.38211 0 0.67076 1.4243 0.266
4 0.99392 0.30581 0 0 0.7227 0.056
5 0.99881 0.26334 0 0 0.1149 0.0224
6 1.07 0.41855 0 0.4789 0.1693 0.105
7 1.0336 0.38716 0 0 0 0
8 1.09 0.38716 0 0.34915 0 0
9 1.0101 0.43018 0 0 0.446 0.2324
10 1.0095 0.436 0 0 0.1361 0.0812
11 1.0341 0.43043 0 0 0.0529 0.0252
12 1.045 0.44162 0 0 0.0922 0.0224
13 1.0349 0.44286 0 0 0.2041 0.0812
14 0.99326 0.46445 0 0 0.2253 0.07
15 1.045 0 1e–05 0.0005 0 0
2 5 1 1.1905 0.36448 0.08157 0.21566
6 12 2 0.23189 0.06239 0.00619 0.01288
12 13 3 0.0549 0.02711 0.00078 0.00071
6 13 4 0.53656 0.21054 0.0192 0.0378
6 11 5 0.24561 0.20575 0.00852 0.01783
11 10 6 0.13909 0.16271 0.00368 0.00862
9 10 7 0.11723 0.07119 0.00064 0.0017
9 14 8 0.24016 0.01373 0.00789 0.01678
14 13 9 0.18492 0.10051 0.00856 0.01743
. 9 10 0.76049 0.28366 0 0.07343
1 2 11 6.2318 0.66011 0.67669 2.0076
3 2 12 2.0685 0.7938 0.22777 0.91334
3 4 13 0.56914 0.73936 0.05892 0.11774
1 5 14 2.445 0.61791 0.30751 1.2204
5 4 15 1.706 0.41627 0.04719 0.13771
2 4 16 1.6408 0.32849 0.14973 0.41769
15 1 17 2.481 0.00305 0 0.15903
5 6 18 1.3277 0.05975 0 0.44383
4 9 19 0.4229 0.04625 0 0.109
4 7 20 0.76049 0.12382 0 0.13695
8 7 21 0 0.59733 0 0.0529

1 1.06 0 6.1958 0.2248 0 0
2 1.045 0.35108 0.18 2.1658 0.6076 0.1778
3 1.01 0.78929 0 0.67076 2.6376 0.266
4 0.93118 0.6312 0 0 1.3384 0.056
5 0.93336 0.54211 0 0 0.2128 0.0224
6 1.07 0.85967 0 0.4789 0.3136 0.105
7 0.99347 0.80013 0 0 0 0
8 1.09 0.80013 0 0.34915 0 0
9 0.96574 0.88744 0 0 0.826 0.2324
10 0.96818 0.90045 0 0 0.252 0.0812
11 1.0103 0.88679 0 0 0.098 0.0252
12 1.0296 0.90658 0 0 0.1708 0.0224
13 1.0126 0.91136 0 0 0.378 0.0812
14 0.94051 0.96092 0 0 0.4172 0.07
15 1.045 0.06074 2.481 0.00385 0 0
Wind + pv
2 Wind
pv
1
DP (p.u.)
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
–1
–2
–3
Time (h)
1.8
1.6
1.4
Wind + pv
1.2
Wind
Plosses (p.u.)
1
pv
0.8
0.6
0.4
0.2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (h)
Figure 17.16 Power losses versus day hours
Table 17.18 Maximum DP values and time for the different cases
Case # First case Second case Third case

DPþ
max ; ti
þ 2.5, t6 2.25, t5 2.1, t6
DP ;
max i t
2.3, t18 2.6, t18 2.6, t18
þ
max is the maximum positive value at the ith hour of the day, ti .
DPþ
DP
max is the maximum negative value at the ith hour of the day, ti .

It is to be noted that the positive value of DP represents the available power

that can be supplied from the system to charge the ESS, whereas the negative value
is the power that must be discharged into the system. Therefore, DP can be con-
sidered as an important parameter by which the size of the ESS in terms of its state
of charge and depth of charge can be specified [4–6].
17.3 Load flow calculations for transient stability

It is important to apply the load flow calculations to study the transient stability
of power systems with renewables during contingencies such as fault occurrence
and system congestion. The following examples illustrate how to identify
the rotor angle stability of conventional synchronous generators in the power
system that is integrated with RESs, for example, wind and/or solar PV and
ESSs.
Example 17.2 As in Example 17.1, with the same data of the IEEE-14 bus system,
load profile, wind farm and solar PV plant, investigate the transient stability of the
system for the three cases (merging wind, solar PV and combination of wind and
PV) when a three-phase fault occurs at Bus #2 at 5 am, and is cleared by isolating
line #16.
Solution
The original power system (i.e. without any renewable sources) has a total load of
3.626 p.u. according to the load profile (Table 17.6). Applying the load flow
technique to the system by using MATLAB/Power System Analyse Toolbox
(PSAT), version 2.1.10 programs [7], it is found that the initial steady-state rotor
angle of each generating unit, d, is as given in Table 17.19. Running the program at
different clearing times, it is noted that the critical clearing time (CCT) that keeps
the system stability is 290 ms and at clearing time greater than that the system is
unstable (Figure 17.17).
Case (i): Replacing G1 by a wind farm and ESS at Bus #1
Considering a wing generating plant with a doubly fed induction generator (DFIG)
operating at the sub-synchronous mode, its equivalent circuit is shown in
Figure 17.18(a). The power injected into the grid and equations of motion for the
Table 17.19 Rotor angle of generating units at steady state
Generating unit # G1 G2 G3 G4 G5
d (rad) 0.53341 0.10487 0.33367 0.37167 0.34132
100 100
δG1
0 δG2 50 δG1
δG4 δG5 δG2
–100 0
δG4
δ (°)
δ (°)
–200 –50
δG5
–300 δG3 –100
δG3
–400 –150
–500 –200
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time (s) Time (s)
(a) (b)
1.05
1.04 ωG4 ωG3
1.03
ωG2
ω (p.u.)
1.02
ωG5
ωG1
1.01
All generators
1
0.99
0.98
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time (s)
(c)
Figure 17.17 Rotor angles response of generating units in the conventional

system: (a) angle at clearing time 300 ms, (b) angle at CCT 290 ms
and (c) speed variation at CCT 290 ms
*
is is ir ir Q*s – + + Vqr 1 idr Lm –Σ Qs
+ V*s
Σ PI Σ PI Σ Σ Rr + SσLr Ls Vs
+ Rs Xσs Xσr Rr + – – + +
vs Xm s vr Qs/Vs SωsσLr SωsσLr 1
idr Ls
s
– –
iqr iqr ψds
(a) (b)
Figure 17.18 DFIG at sub-synchronous speed mode: (a) DFIG equivalent circuit
and (b) stator reactive power control via RSC
DFIG dynamic model in d-q frame, as explained in Chapter 2, Section 2.4.1.2, can
be rewritten as
g
P ¼ vds ids þ vqs iqs þ vdr idr þ vqr iqr
xm vidr v2 (17.6)
Q¼
xs þ xm xm
g
ðTm Te Þ
w_ m ¼
2Hm (17.7)
Te ¼ yds iqs yqs ids
Table 17.20 Wind farm data
Inertia constants Hm (kWs/kVA) 3

Voltage control gain Kv (p.u.) 10
Power control time constant Te (s) 0.01
Pmax and Pmin (p.u., p.u.) [1 0.00]
Number of wind generators that composes the park 30
500 500
400 400
300 300 δG2
200 δG2 200
δ (°)
δ (°)
100 100
0 0 δG3
δG3 δG4
–100 –100
δG5 δG4
–200 –200 δG5
–300 –300
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Time (s) Time (s)
(a) (b)
20 1.004
10 δG3 1.002
ωG4
All generators
0 1
ω (p.u.)
δ (°)
δ
–10 0.998 ωG3G2
δG4 ωG5
–20 δG2 0.996
δG5
–30 0.994
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 2 4 6 8 10 12 14 16 18 20
Time (s) Time (s)
(c) (d)
Figure 17.19 Performance of synchronous generators for the system with wind
energy source at Bus #1: (a) angle at clearing time 290 ms,
(b) angle at clearing time 76.5 ms, (c) angle at CCT 76.4 ms and
(d) speed variation at CCT 76.4 ms
The grid-side converter is generally operated at unity power factor mode. The
reactive power injected/absorbed via the stator terminals is controlled by the rotor-
side converter (RSC) which is considered as a lossless converter as shown in
Figure 17.18(b) [7].
The wind farm data given in Table 17.20 in addition to the data tabulated in
Table 17.5 are the input parameters to run PSAT program. At the instant of fault
occurrence, PL ¼ 3.626, DP ¼ 2.6297 p.u., the initial rotor angles, d, for G2, G3, G4
and G5 are given in Table 17.19. As DP is positive, the ESS is in the charging
mode; that is, it acts as a load and consequently Bus #15 is a load bus.
Applying MATLAB/PSAT programs to solve the stability problem, it is found
that the CCT is 76.4 ms and for any time more than that the system is
unstable (Figure 17.19).
pref idref 1 id
v Current 1 + sTd
– set point
vref qref controller iqref 1 iq
PI
+ 1 + sTq
v,ϑ Inverter
Figure 17.20 Block diagram of PV interface model
60 40
40 δG1 30
δG1
20 20
0 10
δ (°)
δ (°)
–20 0
δG5
–40 δG4 –10
δG3, δG4
–60 –20
δG3 δG5
–80 –30
0 0.2 0.4 0.6 0.8 1 1.2 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
(a) Time (s) (b) Time (s)
40
30
δG1
20
10
δ (°)
–10 δG4
–20 δG3
δG5
–30
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
(c) Time (s)
Figure 17.21 Rotor angles of synchronous generators for the system with solar
PV energy source at Bus #2: (a) at clearing time 290 ms, (b) at
clearing time 76.4 ms and (c) at clearing time 3.24 ms
Case (ii): Replacing the conventional generating unit, G2 by a solar PV plant and
an ESS at Bus #2
At the instant of fault occurrence, PL ¼ 3.626, DP ¼ 2.2297 p.u., rated active
power ¼ 60 MW, the rotor angles of G1, G3, G4 and G5 are given in Table 17.19.
The solar PV system is interfaced with the grid through an inverter as shown in the
block diagram (Figure 17.20). The inverter response times, Td and Tq, are equal to
0.015 s. The voltage PI controller gains, Kv and Ki, are 0.0868 and 50.9005,
respectively, and the voltage reference is 1.045 p.u. [7]. The CCT that keeps the
system stability is 3.24 ms, whereas at a clearing time of 290 or 76.4 ms, the system
is unstable (Figure 17.21).
The injected active and reactive power to the AC bus can be computed as
P ¼ vd i d þ vq i q
Q ¼ vq i d þ vd i q g (17.8)
Case (iii): The power system incorporates a wind farm and an ESS at Bus #1 and
a solar PV at Bus #2 rather than a conventional generating unit at these buses
Combining the configuration of Cases (i) and (ii) as a configuration of Case (iii)
and applying the load flow algorithm with MATLAB/PSAT programs to study the
rotor angle stability of the rotating generators, the variation of rotor angle against
time at different clearing times is obtained as shown in Figure 17.22.
–18
–19 δG4
δG3
–20
δ (°)
–21
–22 δG5
–23
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
(a) Time (s)
–18
–19
δG3
–20
δ (°)
–21
δG4
–22 δG5
–23
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
(b) Time (s)
Figure 17.22 Rotor angles of synchronous generators for the system with
wind þ solar PV energy sources at Buses #1 and #2, respectively:
(a) unstable at clearing time 76.4 ms and (b) stable at clearing time
0.324 ms
Table 17.21 Maximum CCT and status of different system configurations
System Original Case (i) Case (ii) Case (iii)

configuration
CCT (ms) 290 76.4 3.24 0.324
Status Stable Stable Critically stable, Critically stable,
not secured not secured
The maximum values of CCT that can keep the system stability for the original
and three cases of system configurations are summarized in Table 17.21. It is to be
noted that the configurations of both Cases (ii) and (iii) are critically stable and may
have no stability margin, that is, may not be secure as the value of CCT is too small.
In other words, with the addition of renewable energy generation the effective
system inertia is reduced significantly and should be enhanced by introducing
virtual inertia into the system as explained in Chapter 14, Section 14.4.4, for
example, by using hybrid ESS (flywheel þ batteries) [8–14].
References
[1] Ackerman T. Wind power in power systems. 2nd edn. New York: John Wiley &
Sons Ltd.; 2012. pp. 6–20.
[2] Feijóo A. L. and Cidrás J. ‘Modelling of wind farms in the load flow ana-
lysis’. IEEE Transactions on Power Systems. 2000;15(1):110–15.
[3] Su H. I. and El-Gamal A. ‘Modelling and analysis of the role of fast-
response energy storage in the smart grid’. Proceedings of Forty-Ninth
Annual Allerton Conference; Allerton House, UIUC, Illinois, USA, Sep
2011, pp. 719–26.
[4] Luba-Rubio R., Trejo-Perea M., Vargas-Vázquez D. and Rı́os-Moreno G. J.
‘Optimal sizing of renewable energy systems: A review of methodologies’.
Journal of Solar Energy, Elsevier. 2012;86:1077–88.
[5] Oimm A. J., Cockerill T. T. and Taylor P. G. ‘The potential of peak shaving
distribution networks using electricity storage’. Journal of Energy Storage,
Elsevier. 2018;16:231–42.
[6] Reihani E., Sepasi S., Roose L. R. and Matsuura M. ‘Energy management at
the distribution grid using a battery energy storage system (BESS)’. Electrical
Power and Energy Systems, Elsevier. 2016;77:337–44.
[7] Federico M., Vanfretti L. and Morataya J. C. ‘An open source power system
virtual laboratory: The PSAT case and experience’. IEEE Transactions on
Education. 2008;51(1):17–23.
[8] Taczi I. and Vokony I. ‘Possible role and effects of synthetic inertia on the
future’s power systems’. Engineering and Industry Series. Volume Power
Systems. Energy Markets and Renewable Energy Sources in South-Eastern
Europe. Budapest, Ungaria: Trivent Publishing; 2016, pp. 32–47.
[9] Gonzalez-Longatt F., Chikuni E. and Rashayi E. ‘Effects of the synthetic

inertia from wind power on the total system inertia after a frequency dis-
turbance’. IEEE International Conference on Industrial Technology (ICIT);
Cape Town, South Africa, Feb 2013, pp. 25–8.
[10] Tamrakar U., Shrestha D., Maharjan M., Bhattarai B. P., Hansen T. M. and
Tonkoski R. ‘Virtual inertia: Current trends and future directions’. Journal
of Applied Sciences. 2017;7(7):1–19.
[11] Xia S., Zhang Q., Hussain S. T., Hong B. and Zou W. ‘Impacts of integration
of wind farms on power system transient stability’. Journal of Applied
Sciences. 2018;8(8):1–16.
[12] Eriksson R., Modig N. and Elkington K. ‘Synthetic inertia versus fast fre-
quency response: a definition’. IET Journal of Renewable Power Generation.
2018;12(5):507–14.
[13] Waffenschmidt E. and Hui R. S. Y. ‘Virtual inertia with PV inverters using
DC-link capacitors’. IEEE 18th European Conference on Power Electronics
and Applications (ERE’16 ECCE Europe); Karlsruhe, Germany, Sep 2016,
pp. 1–10.
[14] Shi L., Dai S., Yao Y., Ni L. and Bazargan M. ‘Transient stability of power
systems with high penetration of DFIG based wind farm’. IEEE Power
Engineering Society, General Meeting; Calgary, AB, July 2009, pp. 1–6.
Chapter 18
Integration of electric vehicles with renewables
into power grids
18.1 Electrification of transportation

Internal combustion engine vehicles using fossil fuels are at present the dominant
energy sources for transportation sector. In the future, some problems may occur
because of the fossil fuel availability and the continuous increase in fuel price. In
addition, burning the fossil fuels produces greenhouse gases (GHGs) that highly
affect the global climate change.
Electrification of transportation sector seems to be one of the feasible solutions
to these problems. Electric vehicles (EVs) such as plug-in hybrid electric vehicles
(PHEVs) and plug-in electric vehicles (PEVs) will transfer energy demand from
fossil fuels to electricity for transportation sector. Thus, reducing the pollution and
alleviating issues that may compromise security can be achieved [1].
EVs can be envisioned as distributed battery energy storage. Hence, EVs can
help enable the power system to accommodate large-scale intermittent renewable
sources (e.g. wind and solar photovoltaics (PVs)) through utilizing the EVs to
balance the power generated from these sources.
Moreover, integration of EVs into the power grid in addition to the develop-
ment of battery technology has a potential to provide grid support by delivering
ancillary services (peak power shaving, spinning reserve, voltage and frequency
regulations) whenever needed under the concept of vehicle to grid (V2G) presented
in Section 18.3.
Penetration of large-scale renewable sources into power system entails large
battery energy storage system (BESS) to smoothly support electric grids so that the
electrical demand and operating standards are always met [2,3].
For all modes of EVs’ operation (delivering ancillary services, supporting
the grids and charging the batteries), the vehicles can be clustered in a way that
constitute several vehicle fleets based on the similarity of both performance and
charging patterns of vehicles in each fleet. Then, the EV batteries of each fleet
are aggregated for the network management response as they need to be managed
carefully in order to avoid interruption when several EVs are introduced into
the system, whatever the state of charge (SoC) of the batteries. So, the aggregator
clusters EVs in the form of a virtual battery that can be controlled to provide
ancillary services such as primary, secondary or tertiary frequency control.
It knows exactly the state in which the EVs are currently in, that is, the vehicles’
SoC and power capacity. Consequently, the aggregator can determine the virtual
storage characteristics. Then, a sophisticated controller is required to regulate
multiple EV batteries for each fleet based on signals received from the aggregators.
The controller also receives signals indicating the power flow in each vehicle fleet.
Consequently, it can decide which fleets are able to afford the power needed by the
grid, or the power required from the grid to charge the batteries (Figure 18.1).
Charging/discharging of EVs should be controllable to satisfy both utility and
vehicle owners’ technical, environmental, economic and social requirements. For
instance, technically, emerging large number of EVs to be connected to the grid
simultaneously may compromise the quality and stability of the overall power
system. Regarding economic and social aspects, it is useful for designing the
charger control to understand how people charge their EVs: once per day, charging
early or at night, always expecting a full battery or charging on the same day as
other neighbours [4]. More details about charging schemes are described in
Chapter 12, Section 12.6.2 [5–10].
Substation MV/LV
To
controller
Fee
der
Vehicle
fleet #1
Controller
Ag
gre
To gat
From or
Power controller
aggregators Solar
signal power
Distribution
board Charge
Discharge
EV
Vehicle
fleet #i
To Po
we
r fl
controller ow
Co
ntr
ol f
low
Vehicle
fleet #i
Figure 18.1 Vehicle fleets connected to the substation for energy management
Integration of electric vehicles with renewables into power grids 459
18.2 EVs integration with RESs

Penetration of renewable energy sources (RESs) into the electric power systems
(EPSs) causes challenges. The power generated from these sources, especially PV
and wind, is fluctuating and unpredictable because of its dependency on the solar
radiation and wind speed. This nature of RESs necessitates that it be associated
with an energy storage system (ESS) for being dispatchable and enabling energy
management. EV batteries can be aggregated and act as dynamic energy storage
devices. They can be charged by the excess power generated by the RES or dis-
charged to deliver power to the grid during low power generation. Thus, EVs will
act like energy buffer for generation–demand balance and introducing ancillary
services. In addition, with adopting EVs capabilities and integrating the RESs
specifically, optimal GHG emissions and cost reduction can be achieved by
applying multi-objective optimization techniques that maximize customer benefits
and minimize disturbances to the grid.
As a perspective of the application of EVs with RESs (PV and wind) integrated
into electric distribution grid, an envisioned schematic diagram is depicted in
Figure 18.2. There are two configurations of PV solar energy sources: centralized
PV solar plant and distributed public solar carport charging stations. EVs are also
either in parking lots or in a cluster. EV fleets are integrated to the PV solar plant.
PV solar plant
PV arrays
LV/MV
Inverter transformer
Controller
PV controller
Power distribution substation
EV fleets
MV/LV
transformer
Parking lot Parking lot

Wind energy Commercial and/or
(public charging station) (public charging station)
source residential load
Figure 18.2 EVs with renewable sources integrated into power grid
They act like a load when charging the batteries by surplus energy provided by the
PV plant or like a source in discharging state when the PV delivered power is low.
The PV plant with EV fleets is integrated into the power grid through PV controller
and an inverter to introduce the services required by the grid. However, for a spe-
cific area, the residential and/or commercial loads are supplied by power delivered
from wind energy source with EV batteries presented in the parking lots as ESS.
Meanwhile, these batteries are charged by the solar carport charging stations and
utility grid that act as backup source. It is to be noted that the EVs are connected to
bi-directional DC/AC power converters depending on the battery state, charging or
discharging. The utility power distribution substation is equipped with a controller
for energy management [11].
18.3 EVs for ancillary services in power systems

Ancillary service in power systems such as voltage control, frequency regulation,
spinning and non-spinning reserve can be provided by aggregated EVs (PEVs and
PHEVs). Provision of these services can be local for isolated systems or wide for
integrated power systems. To achieve that especially for widespread systems, need
of communication infrastructure and information exchange technologies as well as
advanced controls and measuring devices is indispensable. The advanced control
envisions the EVs as a storage that can be scheduled to either consume power
(charging) or feed power to the grid (discharging). This means that the EVs remain
connected to the grid and are ready to deliver the energy stored in their batteries
when needed. This mode is referred to as V2G.
18.3.1 V2G concept

EVs integrated into power systems can operate as dynamic load drawing power
from the grid during charging or dynamic ESS feeding power to the grid during
discharging, which is referred to as V2G. Spatial location and low individual sto-
rage capacity of EVs entail the aggregation of a large number of vehicles to enable
them for V2G service. The aggregation of EVs depends on the control scheme and
the service required under V2G concept. The aggregator coordinates all required
operational activities like communicating with distribution system operator (DSO),
transmission system operator (TSO) and energy system providers, and maintains
the link between energy market players and the EV owners. Thus, it helps parti-
cipation of supporting electric grid regulation and energy management as a single
controllable distributed energy source (DES) in a virtual power plant (VPP). The
VPP can be understood as a coalition composed of multiple energy producers (e.g.
wind and PV solar) and energy storage providers (e.g. EVs) that come together to
sell electricity as an aggregate [11,12]. So, within the VPP architecture, the EVs are
visible to the DSO, TSO or grid operator through the aggregator and can easily
participate in the energy market.
On the other hand, the individual EV owners can play a central role by parti-
cipating in the energy market through managing the queries from the DSO, TSO
and/or energy market players with the aid of the two-way communication and
control systems [13].
A schematic diagram to depict the role of VPP control and implementation
under the V2G concept is shown in Figure 18.3. The EV aggregator operates as a
VPP and the clustered EV fleet at charging station provides status (e.g. SoC,
available power) to the charging management system (CMS), which communicates
with the aggregator control centre (ACC). The aggregated battery power can be
dispatched by the VPP control centre to provide ancillary services according to the
request of DSO or TSO. The VPP is responsible for communications and power
flow management between market players and grid operators. Communication and
power lines are illustrated by dotted and solid lines, respectively. The individual
EV owners communicate with ACC through EV management system (EVMS)
using smart meters (SMs) as measuring devices. It is noted that the operation of EV
fleet as demand-side management unit, dynamic load and ESS can be optimized by
aiming to minimize the operation cost.
Accommodating RESs and ESSs, or in general DESs, and satisfying the
requirements of secure operation make the energy management of the traditional
power grid to be more complex and challenging. A grid that is more flexible,
efficient, reliable and secure can be configured with a goal of enhancing power
generation and distribution. Such a grid encompasses advanced information and
communication technologies (ICTs), smart energy metering and advanced control.
18.3.2 EVs for power grid frequency regulation

Electric power grids require that the power generation and loads are in balance to
keep the grid frequency constant. In the case of imbalance between generation and
TSO DSO
DSM response
Ancillary
services
VPP control Aggregator

centre control centre
PV arrays Power flow
Ancillary services
request
Energy demand
CMS EVMS
EV status
Wind Individual EV
energy sources
SM
DG/
ESSs
Energy market players: EV fleet in charging station Distribution
power consumers and box
producers From/to Energy demand
SM SM energy market Charger
players
Figure 18.3 Perspective of VPP framework under V2G concept

load, the grid frequency will change. Grid frequency regulation is usually a service
performed by the power plants that are controlled to vary their generation output up
and down from a nominal value. EVs can provide a similar function by varying
their charging rate based on locally measured grid frequency. As EVs reach large-
scale adoption, they have the potential for frequency regulation through charge/
discharge process.
Concerning the environment, climate change, fossil energy reserve, energy
cost and energy security, EVs may have a promising role and are getting great
attention of governments and automotive companies in the world. Because of their
‘zero emissions’ and potential application, the market size of EVs will be enor-
mous. V2G concept has been proposed to make full use of EVs in supporting power
grids [14]. It means that EVs can represent a reasonable source of generation and
storage capacity for the power system. Several investigations have shown the fea-
sibility of this service [15] and new niche markets [16]. Owing to the fast charging
rate of the battery, frequency regulation is regarded as one of the most promising
and practical services under V2G concept [17].
By adopting adequate control strategies of different regulation modes (primary/
second/tertiary frequency regulation), benefits for the stability of frequency may be
achieved. However, frequency regulation with EVs is still in its infancy. A series of
adaptive control strategies (effective smart management, coordination between the
two parties – ‘utility grid and vehicle owners’ – and the incentive policy to be in a
win-win situation) is needed.
EVs can be used as mobile storage devices, which can provide regulation
services [18]. According to the scale, EVs are divided into two types as follows:
● Small-scale EV: The charging points are located in private areas with public or
private access, for example, domestic charging. These charging points allow
slower charging rate.
● Large-scale EV: The charging points are located in public areas with public
access such as charging stations, battery exchanging/swapping stations and fast
charging station. Large-scale EV is dedicated to EV fleets and allows for fast
charging.
18.3.2.1 Large-scale EV participation in frequency regulation

Nowadays, the development of large-scale RES has brought up new challenges to
power system secure operation [19]. With high penetration of renewable energy
connected to grid, the proportion of traditional generating units in operation is
gradually dropping, and the frequency fluctuation caused by intermittent power
generation will exceed the capability of traditional generators for suppressing such
fluctuations. Therefore, traditional frequency regulation may be hardly able to meet
the balance of supply and demand in the grid integrated with RES.
Meanwhile, management of a ‘large-scale EV’ requires extra facilities
including (i) a smart interface with the grid, (ii) communication with the grid
operator and (iii) metering the exchange of power flows information of the grid
[20]. These requirements bring more challenges to power grids in the process of
frequency regulation such as the coordination between the RES and the EVs, bat-
tery charging strategies for different system operating conditions, and technical and
economic challenges of adaptive control algorithms in frequency regulation. To
accommodate these problems, evaluation of the impact of charging/discharging
strategies on power systems, active management solution and control strategies
during the batteries’ charging periods should be accounted for.
EV fleets more easily accommodate the existing electric market rules [21]. In
power markets, the demand response and management are required to achieve the
active management target. A scheme for integration into power system can be
modelled by energy hubs and demand management [22]. To support large-scale
RES and stabilize the grid, the use of EVs under V2G concept to provide storage
for large-scale renewable energy generation has been studied [23]. The storage
capacity can attenuate the intermittence of RES, and still provide an asymmetrical
regulation, such as ‘regulation up’ or ‘regulation down’ [24].
When large disturbances such as faults and line trips occur, the power system
is subjected to speed and frequency variation, and voltage fluctuations. These
fluctuations affect the power quality that customers experience, and if they are
strong enough, they could even damage sensitive appliances. Moreover, if the
disturbance is not cleared within a specific time (critical clearing time), the systems
may loose synchronism and are unable to resume stable operation.
In frequency control, there are three layers of control: primary, secondary and
tertiary. EVs can simulate the generator droop characteristic to perform primary
frequency regulation and achieve secondary frequency control based on area con-
trol error (ACE). Tertiary frequency modulation can be achieved based on eco-
nomic dispatch [25].
Primary frequency control. In microgrids, a power disturbance would result in
a significant oscillation of frequency due to lack of enough generation resources
and spinning reserve. EVs connected to the grid can inject power into the grid,
which not only benefits the frequency stability, but also increases robustness of
operation, especially in isolated system. EVs can be considered as new resources in
primary frequency control owing to their fast response to disturbances [26]. Studies
such as in [26] make the V2G more noticeable than ever in primary frequency
control.
In primary frequency control, there are two loops: first, the droop control that
reduces the output of traditional generators, which can cut the fossils consumption,
and second, the inertial control that simulates the behaviour of conventional gen-
erator. So, these two control schemes autonomously respond to frequency devia-
tion. To alleviate the effect of intermittence of RES, a series of V2G control
approaches are adopted in some literature [27]. The effect of EVs as controllable
load in primary frequency control and battery SoC should be investigated [28] and
to satisfy the expected battery SoC, an automotive distributed control should be
applied [29]. The scheduled charging power can be estimated based on the historic
data of frequency deviation, which is dominated by normal distribution [27].
Secondary frequency control. As a centrepiece of secondary frequency con-
trol, automatic generation control (AGC) is an important control technology for
keeping frequency stability. If EVs participate in dispatching load, AGC would

respond to frequency deviation on both generation and load sides simultaneously to
offset the insufficient response of traditional generating unit. The feasibility and
potential AGC operation with EVs are analysed in [30] and the methodology to
make EV participate in AGC in bulk power system is designed [28]. These studies
show that EVs can assist AGC. Additionally, some scholars have discussed cen-
tralized V2G control method to participate in the load frequency control. In [25], a
method for tracking signal is designed. But these studies have not considered how
to send signal to EVs.
At present, the study of AGC with EVs is still in initial stages, and the focus is
on EV energy storage to reduce the cost of BESS. EV as a distributed controllable
load is discussed in [31]. Coordination control method is applied to integrate large-
scale wind power and PV power into a large power system [32]. These results show
that EVs can effectively reduce the ACE, reduce the tie-line flow deviation and
lower the frequency modulation capacity [28].
Tertiary control. EVs participate in tertiary control based on economic dis-
patch. Stochastic dynamic programming method is used in [33] to optimize EV
charging and frequency regulation decisions, taking into consideration the uncer-
tainty of EV and wind power. The economic dispatch model was established in
[34]. The results indicate that the EVs plug-in grid can reduce the gap between
demand and supply in different periods of time with optimal dispatch schedule.
Nevertheless, the economic dispatch with EVs does not include several constraints
such as reduction of GHG emissions and loss reduction in power grid. The eco-
nomic dispatch models containing many uncertain factors and calculating the
objective function of the volatility of constraints, using various mathematical pro-
gramming techniques and intelligent optimization, has received little attention
recently.
However, more advanced controllers can be utilized to stabilize the frequency
deviation in case of fault in the system. In these cases, fleets of EVs are aggregated
into one actuator, which either draws or feeds power from or into the network. The
impact of these services on the battery energy content is neglected due to its
duration of some seconds only. The dynamic model of batteries is assumed to be
homogenous; that is, battery-ageing differences between individual EVs are
neglected.
In addition, batteries are equipped with smart chargers that introduce intelli-
gent charging schemes to control the EV charging process. Controllable EV char-
ging surmounts or alleviates the undesired impacts produced by uncontrollable EV
charging such as the overloading of the power system facility and increasing power
demand resulting in a less efficient electricity supply. SMs can pursue various
objectives such as minimization of charging costs in the electricity market, GHG
emissions or power system losses. It is crucial to achieve these objectives in an
optimal manner through applying optimization techniques and considering the
system constraints especially in case of large-scale EV penetration. Several con-
straints coexist in the real-world implementation as well as their variation
depending on the objectives of the deployed EV system.
More description of charge/discharge management and frequency regulation as

ancillary services under V2G concept is presented in the following example.
Example 18.1 Scheduling charging/discharging processes of EVs’ batteries as well

as use of EVs as a source of distributed frequency regulation are investigated
through the application to the nine-bus system with the data shown in Figure 18.4
and Tables 18.1 and 18.2. The load curve is given as depicted in Figure 18.5.
Five different car user profiles are included in each fleet:

● Profile #1: People going to work with a possibility to charge their cars at work
(100 EVs).
● Profile #2: People going to work with a possibility to charge their cars at work,
but with a longer ride (100 EVs).
● Profile #3: People going to work with no possibility to charge their cars at
work (100 EVs).
● Profile #4: People are staying home (250 EVs).
● Profile #5: People working on a night shift (250 EVs).
Load #3
18 kV 230 kV 230 kV 13.8 kV
j0.0625 j0.0586
0.0085 + j0.072 0.0119 + j0.1008
B/2 = j0.0745 B/2 = j0.1045
G2 18/230 kV 230/13.8 kV G3
2 8 3
0.039 + j0.170
7 9
0.032 + j0.161
B/2 = j0.179
B/2 = j0.153
5 6
0.010 + j0.085
0.017 + j0.092
B/2 = j0.088
B/2 = j0.079
Load #2
Load #1
230 kV
4
16.5/230 kV
j0.0576
16.5 kV 1
G1
Figure 18.4 Nine-bus system impedance diagram: all impedances are in p.u. on a
100-MVA base
Table 18.1 Data bank of the nine-bus system
Bus no. Bus type Generation (p.u.) Load (p.u.) Voltage magnitude
PG QG PL QL
1 Swing – – 0 0 1.04
2 PV 1.63 – 0 0 1.025
3 PV 0.85 – 0 0 1.025
4 PQ 0 0 0 0 –
5 PQ 0 0 1.25 0 –
6 PQ 0 0 0.9 0 –
7 PQ 0 0 0 0 –
8 PQ 0 0 1 0 –
9 PQ 0 0 0 0 –
Table 18.2 Nine-bus generation data
Generator G1 G2 G3
Rated MVA 247.5 192 128
Power factor 0.85 0.85 0.85
Line–line voltage (kV rms) 16.5 18 13.8
Frequency 60 60 60
xd 0.8979 1.05 1.305
x0d 0.2996 0.235 0.252
xq 0.474 0.646 0.889
0
xq 0.243 0.23 0.22
x‘ (leakage) 0.186 0.134 0.134
T 0d0 7 601 4.75
T 00d0 0.03 0.04 0.068
T 00q0 0.0513 0.099 0.21
Stator resistance (Rs) 0.0031 0.0031 0.0031
Inertia coefficient (H) 4.7 5.14 6.54
Friction factor (F) 0 0 0
Pole pairs (P) 1 1 1
18.3.2.2 Control architecture

The multi-agent system used to perform distributed frequency regulation has a
hierarchical structure. The vehicle local controllers (LCs) and ACC are working
together to determine the regulation needs and control local dispatch. The LC will
reside within each EV to monitor and control battery charging and discharging
while a vehicle is plugged-in. The objective of the LC is to control charging based
on the owner’s preference and minimizing the impact on the life of the battery
system. Although the LC relies on the ACC to inform its control decisions, it will
necessarily make the decisions without regard to what neighbouring LCs are doing.
Therefore, the LC acts a decentralized control (Figure 18.6).
350
300
250
Power (MW)
200
150
100
50
0
0 5 10 15 20
Time (h)
Figure 18.5 Daily load curve
Po
AG
we
AC
3
rc
C
alc
ine
AG
nl
ula
tio
2
ica
tor
un
AG
mm
Co
1
Parking lot (EV charging station #3) Parking lot (EV charging station #2) Parking lot (EV charging station #1)
LC
LC
LC
TSO
LC
LC
LC
et et
fle fle et
EV EV fle
EV DSO
SM SM SM
SM SM SM
Power line
Figure 18.6 Hierarchical structure of control system
● LC: It controls the battery behaviour and makes the decision as to how much
power and energy the battery system can provide for frequency regulation.
Determining how much energy a vehicle will supply or absorb for frequency
regulation depends on the battery system state and the supply recommendation
made by the ACC. As a simplifying assumption in this study, a vehicle will
remain plugged-in for the full expected charge duration once it returns home,
also minimizing the impact on vehicle drivability. In the case that supplying
regulation extends the expected charge duration and charging is interrupted,
the vehicle could still be driven.
The general approach of the LC is to begin charging at a mid-level rate, e.g.

1 kW, such that there is a margin to increase or decrease charging to supply
frequency regulation.
The duration of regulation supply from an EV could be several hours
depending on the current battery state and how long the vehicle remains
parked. However, as the vehicle may be unplugged or LC overridden at any
point, a shorter supply duration may make more sense. Therefore, several
supply durations ranging between 5 min and 1 h may be considered to deter-
mine the effects on the aggregate supply curve.
The functionality of the LC may also be integrated into a home energy
management system (EMS), which monitors household energy consumption
in real time and acts as a communications gateway between devices within
the home and beyond the point of utility metering. The home EMS may
control several loads in the home to smooth or shift the load profile. Here,
the EV could aid in managing the energy profile of a home without external
system needs. The same LC functionality will be useful in either situation:
controlling the charge rate or discharging to supply part of the load over
short periods in a smooth manner. The vehicle LC will interact with the
ACC, which provides a supply recommendation for frequency regulation,
and tracks the aggregate local EV regulation supply based on the control
actions taken by the LC.
● ACC: Frequency measurements are not reliable at the point of customer ser-
vice where the EV is connected to the grid. Therefore, the LCs are not able to
perform frequency control independently based on local information. Reliable
frequency measurements can be made beginning at the high side of distribution
substation transformers. The ACC located at a higher voltage level where
reliable frequency measurements are available is used in this study to oversee
and coordinate local EVs participating in frequency regulation under V2G
concept. Here, local EV refers to all vehicles that are below the ACC in the
distribution system, which are participating in V2G. The ACC will act as a link
between the area in EPS and the local EV. It would also be the natural choice
to act as an aggregator to perform market functions for the local EV.
The aggregator is a central controller in-charge of coordinating all required
operational activities, like communicating with the DSO, TSO, grid operator
and energy service providers. It also maintains a link between energy market
players and the EV owners. The aggregator knows exactly the state in which
the EVs are currently in, that is, the vehicles, SoC and power capacity. Based
on this information, the aggregator can determine the virtual storage char-
acteristics and can directly control their charging/discharging according to
TSO demands. It then submits bids to the TSO for ancillary services.
Because each level of the power system has different operational needs,
‘local energy needs’ will be defined differently throughout the system. For
instance, determining the amount of local energy needed to contribute to
improving system frequency performance will depend on the load fluctuations
and nearby generation at that point in the system. In addition, the market price
for energy and ancillary services will behave differently at different nodes in
the transmission system.
The ACC will necessarily interact with local sensing and measurement
devices to perform its functions. However, it will also need to interact with the
system operator in EPS and of course the LCs.
18.3.2.3 System simulation

In the context of the bulk power system, the basic objective is to make sure that
there is enough power being generated to supply the power drawn by the system
load, providing customers with a secure and reliable electrical energy supply. There
are several types of measures to determine how well the system is operating,
including power balance, voltage performance, and financial returns. Each con-
sumer and stakeholder will have a unique set of expectations and desired outcomes
with respect to power system performance.
To simulate the use of EV as a source of distributed frequency regulation, a model
including power system behaviour, EV loads, and additional controllers is necessary.
The ACC acts as an interconnection for energy transfer to and from EV participating in
V2G. However, all communications necessary between the EV and the system
operator to facilitate distributed frequency regulation are handled by the ACC.
Communication systems are not considered in this study, neither the physical
implementation nor the protocols that may be used. Standards for communication
systems on the user end are currently being developed and will continue to be refined
as the smart grid is implemented. Future studies may consider the communication
systems to get a more realistic picture of how well the distributed control works.
The simulator is designed to study the speed response of generating units to
changes in system loading, and to facilitate testing of control strategies for balan-
cing load and generation. The ACC tracks the system frequency of the bus and
provides the LCs with the regulation control signal, which combines the elements
of primary and secondary frequency regulation.
The single-line-diagram and the simulation model of the nine-bus system with
EVs are illustrated in Figures 18.7 and 18.8, respectively. The system includes
three generators, three load buses with a load curve shown in Figure 18.5 for each.
Three EV fleets are embedded into the system at buses 4, 7 and 9, with one fleet in
a parking lot (a charging station) for each.
The five load profiles are simulated as shown in Figure 18.9(a). Initialization
of SoC and the state of cars (plug-in or plug-out) can be identified by the simulation
model as depicted in Figure 18.9(b). Once a mileage range is identified, the average
distance driven in that range during a specific hour is used to determine the bat-
tery’s remaining SoC assuming that each vehicle consumes 0.2 kWh/mi in its all-
electric range with a minimum SoC of 20%.
Consideration of a maximum and a minimum SoC is necessary for two rea-
sons: first, battery protections often prevent the battery from reaching too extreme
values to avoid cell degradations, and second, because the battery cannot recharge/
discharge at its maximum rate when approaching extreme SoC values. Maximum
SoC of 0.9 is assumed as for SoC > 0.9, the charging rate decreases and a regular
V2G – Parking lot #2 V2G – Parking lot #3
100.0
Load #3
(35.0)
18 kV 230 kV 230 kV 13.8 kV
G2 G3
2 8 3
1.025 1.016 1.025
9.273° 0.6306° 4.558°
7 9
1.026 1.032
3.62° 1.867°
5 0.996 6 1.013
(50.0)
125.0
90.0
–4.002° 1.867°
Load #2
Load #1
230 kV
4 1.026
V2G – Parking lot #1 –2.226°
16.5 kV 1 1.040
0.0°
G1
Figure 18.7 Nine-bus system with three parking lots showing the bus voltages
(in kV and p.u.)
contribution to frequency regulation cannot be provided. Minimum SoC of 0.4 is

also taken because of the battery protection tripping at SoC below 0.2, and 20%
more for home-to-work journeys. When the SoC drops below 0.4 because of a
journey, the EV will only be allowed to take part in regulation ‘down’, in a con-
dition that it will be recharged when necessary.
The generator speed (Hz) is monitored and controlled to be kept at a desired
value (reference value) by varying the load and thus deciding which of regulation,
up/down, will be done? Accordingly, the ACC estimates the number of EVs to
participate in V2G and gives decision of charging or discharging based on the state
of plugging and regulation up/down. The simulation models of speed control of
both the generators and the aggregator are shown in Figure 18.10(a) and (b).
18.3.2.4 Scheduling the charging/discharging processes

for EVs fleet
The financial transactions recorded at a parking lot for 24 h are presented in
Figure 18.11. Most electricity purchases should be shifted to the off-peak periods as
Bus_8 Bus_3
Bus_3 230 V 13.8 kV
18 kV 1.016 p.u. 1.025 p.u.
1.025 p.u. 18 kV/230 V 0.6306º 13.8 kV/230 V 4.558°
9.173° 6.25% Z1 V2G 3
V2G 2 5.86% Z1
A a a A
B b b B
CYgYg c c YgYgC
Bus_2 Bus_8
G2 230 V
Bus_3 G3
18 kV Load measurements 13.8 kV
(35 MVAR)
100 MW
Bus_7 Bus_9
Bus_7 Bus_9
230 V 230 V 230 V
230 V
1.26 p.u. 1.032 p.u.
3.62° Load 2 1.867°
Bus_5 Bus_5 Bus_6

230 V 230 V Bus_6
230 V
230 V
0.996 p.u.
1.013 p.u.
–4.002°
1.867°
Load measurements Load measurements
125 MW 90 MW
(50 MVAR) (30 MVAR)
Bus_4
Load 1 230 V Load 3
Bus_4
230 V
1.026 p.u.
a
b
YgYgc
–2.226° 16.5 kV/230 V V2G 1
5.76% Z1
C
A
B
Bus_1
16.5 kV Bus_1
1.04 p.u. 16.5 kV
0.0°
G1
Figure 18.8 Simulation model of nine-bus system with EVs
the price is low, whereas electricity sales are centred around the peak periods as the
price is high. Therefore, more connected EVs are scheduled for charging during the
periods when electricity tariffs are relatively low. The available power supplied by
V2G operation of EVs at these periods is insufficient to meet the huge amount of
charging load, and so, the CMS needs to purchase electricity from the grid to
substitute the power deficits. For scheduling the periods with most expensive
electricity prices, the CMS tends to discharge the energy stored in most connected
EVs. Since the EV charging demands at these periods are relatively small, the CMS
can sell the excess electricity to the power grid for more revenue [35].
The variations in average SoC for existing EV profiles in a parking lot and the
rated MW in response to the dynamic electricity tariff are given in Table 18.3. In
this table, EV plug-in state indicator is represented by ‘1’ during charging/dis-
charging or standby states and by ‘0’ for EVs in driven or out of service states. The
power that can be supplied to/delivered from EVs in charging/discharging state is
positive and negative, respectively, whereas in standby state zero power is assumed
as the EV is only participating when needed (like spinning reserve), e.g. Profile 4 at
hours 21 and 22.
The average SoC of EVs starts to grow when the electricity tariff is relatively
low (Figure 18.12). When the electricity price reaches its peak, the majority of
connected EVs are scheduled by the CMS to discharge their stored energy to the
grid, leading to decrement of SoC [36,37].
Applying the load flow technique (e.g. Newton–Raphson method) and
MATLAB/Simulink programs to the nine-bus system without EVs, the values of
dispatch power of each generator, system losses and the load over day hours are
1 Cars in regulation 2 Cars in charge
Nbreg Nbcgarg
Rate limiter Goto Rate limiter Goto 1
× × × × × × × × × ×
Goto 2 Goto 5 Goto 6

Goto 3 Goto 4
Nbcars 1 Nbcars 4 Nbcars 5
Nbcars 2 Nbcars 3
Cars Number Cars Number Cars Number Cars Number Cars Number
1 of cars 1 2 of cars 2 3 of cars 3 4 of cars 4 5 of cars 5
Charge mode
Charge mode
Charge mode
Charge mode
Charge mode
Regulation
Regulation
Regulation
Regulation
Regulation
mode
mode
mode
mode
mode
Car profile 1 Car profile 2 Car profile 3 Car profile 4 Car profile 5
(a)
1 – D T(u)
Second → Hour
Rate limiter
– 1/(60*60) + SoC initialization Charge
Clock Charge
1 + SoC SE% SE% 2
mode
1 – D T(u)
Plug Regulation
Rate limiter 1 Plug mode 1
State of charge
Plug state
Charger control
(b)
Figure 18.9 Simulation model of (a) car profiles and (b) SoC initialization and
plug-in state of each profile
obtained as given in Table 18.4. In the case of the system incorporating EVs, these
values in addition to the net power (the difference between the load power and EVs
powers) are obtained by using the simulation models of the system, aggregator and
load profiles as tabulated in Table 18.5.
The EVs participation in feeding the load and the corresponding power
delivered from the conventional synchronous generators (net power) through day
hours are drawn in Figure 18.13.
It is to be noted that, in Figure 18.13, the EV acts as a load during out of peak
times as it is in charging process, and discharges energy at peak times to operate as
a source of electricity. Therefore, there is a reduction in the load during the peak
wref
+ 1
Switch
Integrator
P max reg 1 Compare
limited
to constant wm1
Gain 1 Derivative ≥0.0005
OR
du/
Saturation + K >= AND
dt Compare to
dynamic
up
1 constant 1
Y
S Gain
u 0 ≤–0.0005
1 lo K
PQ Gain 2 Q1
Q Compare to
–1 constant 2
0 Clock
>1
(a)
P_charge ×
÷ ×
Form 1 Divide 1
Nbcharge
Form 4 K
1 >= Discharge
Nbcars 1 >=
Charge Switch 1 Terminator
Form 2 Switch 2 Rate –1
limiter 2
P_EV × 0 Terminator 1
÷ Scope 1
Form
Divide 2 Constant 1 1 K
10%
K
S 100 1
Nbreg Charge Switch
÷ SE%
Form 3
2 K Capacity*
X X0 1e3
SoC
Product 1 Integrator X SoC cars 1
Nbcars 1 Goto
Nbcars 1 Product 2
Form 5
3 Form 6
Plug
(b)
Figure 18.10 MATLAB/Simulink implementation of (a) speed control and

(b) aggregator model
times, which yields a reduction of the electricity consumed from the conventional
synchronous generators.
18.3.2.5 Application of EVs as a source of frequency regulation

Because of the load/generation imbalance resulting from disturbance occurrence or
congestion in the grid, the frequency will change – increasing if load is less than
generation and decreasing if load exceeds generation.
AGC acts as an interconnected system to ensure that frequency is brought back
to the nominal value and inter-area power flow is regulated. AGC can restore
system frequency in several minutes. So, EVs can be considered as new resources
in secondary frequency control owing to their fast response to disturbances. When
frequency goes downwards, EVs can be easily switched partially or fully from
acting as a load to act as power producer to prevent frequency drop. The feasibility
of EVs as controllable load to provide effective system secondary frequency
response can be approved. Regarding management, the aggregator can alter EVs’
mode with the rate of frequency changes.
0.6
Electricity tariff ($/kWh)
0.5
0.4
0.3
0.2
0 3 6 9 12 15 18 21 0 3 6
Time (h)
Figure 18.11 Financial transactions versus day hours
PEVs not only improve the load curve on peak time, but also can surmount
frequency deviations in a few seconds. Different types of disturbances in IEEE
nine-bus system are simulated such as oscillatory transients (OTs) caused by
switching processes and faults like single line-to-ground (L-G), double line-to-
ground (2L-G) and three-phase short circuit (3-Ph short circuit). The fault is located
on the line between Bus 1 and Bus 4 very close to Bus 1. It occurs at 1:00 a.m. and
at its clearance the system restores its original topology.
For OTs, EVs help damp the frequency fluctuation fast as well as the range of
change is getting less (Figure 18.14(a)). For instance, as in Table 18.6, the time to
reach the allowable frequency change (TR-AFC), which is 0.1% of the rated
value, is reduced from 6 to 3.7 s as well as the maximum range of frequency change
is also reduced from 0.017 to 0.0094 p.u.
When the fault occurs, the frequency drops and EVs act as an energy source to
mitigate the frequency variation. As explained in Chapter 14, Sections 14.4.1 and
14.4.4, the governor response as a primary control is active for the duration of
10–30 s from the instant of fault occurrence and is taken in this simulation as 10 s.
So, V2G is important to play a role in the first 10 s to reduce TR-AFC and damp the
frequency fluctuation fast by reducing the rate of change of frequency (RoCoF). In
the next 10 s, the governor control is in action rather than the EVs. It is to be noted
from Figure 18.14(b)–(d) that the change of frequency is instantly increased and
then starts to decay as well as the instant of fault occurrence is considered as the
zero time for plotting the curves. The numerical results of simulation are given in
Table 18.6.
It is concluded that the EVs can improve power grid transient stability when
the power grid is subject to large disturbances, including bus faults, generator and
branch tripping, and sudden large load changes.
EV behaviour as an aggregate load and source of frequency regulation is better
viewed from the bulk system level. This is mainly because the frequency response
Table 18.3 The average SoC and plug-in state of EV profiles versus day hours
Time Prof. 1 Prof. 2 Prof. 3 Prof. 4 Prof. 5

(h)
SoC Plug-in MW SoC Plug-in MW SoC Plug-in MW SoC Plug-in MW SoC Plug-in MW
(%) state (%) state (%) state (%) state (%) state
indicator indicator indicator indicator indicator
1 40 1 4.8 40 1 4.8 50 1 6 40 1 12 40 0 0
2 50 1 6 50 1 6 60 1 7.2 50 1 15 30 0 0
3 60 1 7.2 60 1 7.2 70 1 8.4 60 1 18 30 0 0
4 70 1 8.4 70 1 8.4 80 1 9.6 70 1 21 30 0 0
5 80 1 9.6 80 1 9.6 90 1 10.8 80 1 24 50 1 6
6 90 1 10.8 90 1 10.8 90 0 0 90 1 27 70 1 12
7 90 0 0 90 0 0 90 0 0 90 1 0 80 1 15
8 80 0 0 90 0 0 80 0 0 90 1 0 90 1 18
9 80 1 0 80 0 0 80 1 1.2 90 1 0 90 1 0
10 80 1 0 80 0 0 80 1 1.2 90 1 3 90 1 3
11 80 1 0 80 0 0 70 1 2.4 80 1 6 80 1 6
12 80 1 1.2 70 0 0 70 1 2.4 80 1 6 80 1 6
13 70 1 2.4 70 0 0 60 1 3.6 70 1 9 70 1 9
14 60 1 3.6 70 0 0 60 1 3.6 70 1 9 70 1 9
15 50 1 4.8 60 0 0 50 1 4.8 60 1 12 60 1 12
16 40 1 6 60 0 0 40 1 6 60 1 12 60 1 12
17 40 0 0 60 0 0 40 0 0 50 1 15 50 1 15
18 30 0 0 50 0 0 30 0 0 40 1 18 40 1 18
19 30 1 1.2 50 1 1.2 30 0 0 30 1 21 40 1 18
20 20 1 2.4 40 1 2.4 30 1 1.2 20 1 24 40 1 18
21 20 1 0 30 1 3.6 20 1 2.4 20 1 0 40 0 0
22 20 1 0 20 1 4.8 20 1 2.4 20 1 0 40 0 0
23 20 1 2.4 20 1 2.4 30 1 3.6 20 1 6 40 0 0
24 30 1 3.6 30 1 3.6 40 1 4.8 30 1 9 40 0 0
Prof. 1 Prof. 2 Prof. 3 Prof. 4 Prof. 5

100
90
80
70
SoC (%)
60
50
40
30
20
10
0
0 5 10 15 20 25
Time (h)
Figure 18.12 The average SoC of EV profiles versus day hours
Table 18.4 IEEE nine-bus simulation results (generators, losses and load) without
EVs versus day hours
Time (h) Gen. 1 (MW) Gen. 2 (MW) Gen. 3 (MW) Load (MW) Losses (MW)
1 80.872799 62.755885 41.911348 184.011671 1.528361
2 77.948526 60.547738 40.416703 177.449282 1.463686
3 75.078261 58.277691 38.932414 170.886708 1.401659
4 71.643945 55.58818 37.1151 163.015989 1.331236
5 77.988806 60.515676 40.411403 177.452391 1.463494
6 80.876709 62.741207 41.923988 184.01357 1.528334
7 92.479023 71.65935 47.946722 210.267017 1.818078
8 104.010048 80.730547 53.934071 236.51584 2.158825
9 115.732023 89.642538 59.941626 262.765404 2.550784
10 121.42111 94.246493 62.990441 275.889382 2.768661
11 124.321996 96.511318 64.500537 282.451403 2.882448
12 130.22057 100.963012 67.51449 295.578123 3.11995
13 133.069069 103.281795 69.033115 302.13891 3.245068
14 134.523535 104.407009 69.797954 305.419777 3.308721
15 135.982403 105.544905 70.54695 308.700927 3.373332
16 138.913513 107.786959 72.068195 315.263491 3.505177
17 135.983248 105.543958 70.546867 308.700726 3.373346
18 133.095961 103.245999 69.040751 302.137932 3.244779
19 130.20566 100.974822 67.514404 295.574855 3.120031
20 127.244122 98.762194 66.004946 289.011676 2.999586
21 121.460248 94.187855 63.000074 275.880054 2.768123
22 103.986806 80.74285 53.938237 236.50899 2.158902
23 92.42523 71.734781 47.920565 210.262147 1.81843
24 80.856502 62.769389 41.914657 184.012092 1.528457
Table 18.5 IEEE nine-bus simulation results (generators, losses, load and net
MW) incorporating EVs versus day hours
Time Gen. 1 Gen. 2 Gen. 3 EV Losses Load Net

(h) (MW) (MW) (MW) (MW) (MW) (MW) (MW)
1 80.8728 62.7559 41.9113 27.6701 1.5284 184.0117 211.6818
2 77.9485 60.5477 40.4167 34.2011 1.4637 177.44923 211.6504
3 75.0783 58.2777 38.9324 40.8005 1.4016 170.8867 211.6872
4 71.6439 55.5882 37.1151 47.4514 1.3312 163.0159 210.4674
5 77.9888 60.5157 40.4114 60.0110 1.4635 177.4524 237.4634
6 80.8767 62.7412 41.924 60.6349 1.5283 184.0136 244.6485
7 92.4790 71.6594 47.9467 15.0120 1.8180 210.2670 225.2790
8 104.0100 80.7305 53.9341 18.0130 2.1588 236.5158 254.5288
9 115.7320 89.6425 59.9416 1.2670 2.5508 262.7654 261.4984
10 121.4211 94.2465 62.9904 7.2147 2.7687 275.8894 268.6747
11 124.3220 96.5113 64.5005 14.4030 2.8824 282.4514 268.0484
12 130.2206 100.9630 67.5145 15.6120 3.1199 295.5781 279.9661
13 133.0691 103.2818 69.0331 24.0608 3.2451 302.1389 278.0781
14 134.5235 104.4070 69.798 25.2106 3.3087 305.4198 280.2092
15 135.9824 105.5449 70.547 33.6060 3.3733 308.7009 275.0949
16 138.9135 107.7869 72.0682 36.1321 3.5052 315.2635 279.1314
17 135.9832 105.5439 70.5469 30.1050 3.3733 308.7007 278.5957
18 133.0960 103.2460 69.0408 36.1380 3.2448 302.1379 265.9999
19 130.2057 100.9748 67.5144 41.4720 3.1200 295.5749 254.1029
20 127.2441 98.7622 66.0049 48.0824 2.9996 289.0117 240.9293
21 121.4602 94.1879 63.0001 6.0579 2.7681 275.8801 269.8222
22 103.9868 80.7429 53.9382 2.4140 2.1589 236.5089 234.095
23 92.4252 71.7348 47.9206 14.4792 1.8184 210.2622 224.7413
24 80.8565 62.7694 41.9147 21.1050 1.5285 184.0121 205.1171
350
Load (MW) Net (MW) EV (MW)
300
250
200
Power (MW)
150
100
50
0
0 5 10 15 20 25
–50
–100 Time (h)
Figure 18.13 Load, EV and net (MW) versus day hours

1.0025
1.004
Under V2G concept Under governor control
1.0020
1.002
With governor
Rated value 1.0015 Instant of response
1 fault
Fcmax with
1.0010 occurrence
governor
Allowable
Fcmax with EVs

Frequency (Hz)
Frequency (Hz)
0.998 variation
Allowable 1.0005 range
range ±0.1%
Fcmax without EVs

0.996
1
Fcmax without EVs

Fcmax with EVs
0.994
0.9995
0.992 With EVs
0.9990
Without both EVs
0.990 0.9985 and governor control
Without EVs
0.988 0.9980
0.986 0.9975
0 2 4 6 8 10 0 2 4 6 8 10 12 14 16 18 20
Time (s) Time (s)
(a) (b)
1.004 1.006
Under V2G concept Under governor control Under V2G concept Under governor concept
1.003
Instant of With governor 1.004
Allowable Instant of Allowable
fault response With governor
1.002 variation fault variation
response
Fcmax with
governor
occurrence range
range occurrence
1.002
Fcmax with
governor
Frequency (Hz)
Frequency (Hz)
1.001
1 1
Fcmax without EVs
Fcmax with EVs
Fcmax without EVs

Fcmax with
0.999
EVs
0.998
0.998 Without both EVs
Without both EVs and governor control
and governor control 0.996
0.997
0.996 0.994
0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20
Time (s) Time (s)
(c) (d)
Figure 18.14 Frequency variation for nine-bus system with/without EVs: (a) OT,
(b) L-G fault, (c) 2L-G fault and (d) three-phase short circuit
Table 18.6 Frequency variation at different faults in the system with/without EV
Type of Maximum range of TR-AFC Maximum range of TR-AFR

fault frequency change (Hz) (s) frequency change (Hz) (s)
without EVs/GR (p.u.) with EVs/GR (p.u.)
From To equal From To equal
OT 0.987 1.004 ¼ 0.017 6 0.9933 1.0027 ¼ 0.0094 3.7
L-G V2GGR 0.998 1.001 ¼ 0.003 6.8 0.9988 1.0005 ¼ 0.0017 5
1.002 0.999 ¼ 0.003 4.7 1.0012 0.9995 ¼ 0.0017 3
2L-G 0.997 1.0015 ¼ 0.0045 3.6 0.9982 1.0009 ¼ 0.0027 1.5
V2GGR 1.0032 0.9982¼ 0.005 3.5 1.0019 0.999 ¼ 0.0029 2
3-Ph sc 0.9955 1.002 ¼ 0.0065 3.5 0.9978 1.0011 ¼ 0.0033 1.6
V2GGR 1.005 0.997 ¼ 0.008 3.5 1.0024 0.9988 ¼ 0.0036 1.7
of central generation needs to be incorporated into a system simulation model to

study how effective EV could be as a source of frequency regulation. As a first step
in modelling EV as distributed energy resources, a transmission level model of the
power system is used and an aggregate EV load will be tied to each load bus in the
system. The power system model is based on a modified version of the power
system simulator.
Power calculations. At the ACC, the aggregated delivered power from each
parking lot can be calculated as shown below.
Some parameters are defined as follows:
● Arrival time (Tarr): The time at which the EV arrives the parking lot.
● Time of request (Treq): The time at which the stored power in EV battery is
required to be delivered to the grid.
● SoCmax: The maximum value of SoC of EV battery.
● SoCmin: The minimum value of SoC of EV battery.
● EV plug-in state coefficient (l): It equals 0 if the EV is on drive, out of service,
SoC SoCmin or in charging, and equals 1 when the EV is in discharge or
standby mode.
● Actual battery capacity (Eactual ): It is the present energy stored in the battery.
● Battery capacity (Echarge/disch): The energy that can be provided or absorbed by
EV battery. It can be calculated by the following equation:
Echarge=disch ¼ Eactual ðSoC max SoC min Þ (18.1)
Each EV in the parking lot is identified by ID number, SoCmax, SoCmin, Tarr

and the load profile number that the EV belongs to. Hence the total energy that can
be supplied to/delivered from EVs’ batteries during charging/discharging in the
three parking lots can be calculated by the ACC using the following equation:
" !#
XN XL X M
Etotal ¼ Ei;j;k
i¼1 j¼1 k¼1 (18.2)
i ¼ 1; . . . :; N; j ¼ 1; . . . : ; L; k ¼ 1; . . . :; M
where
N ¼ number of parking lots (i.e. no. of fleets);
L ¼ number of load profiles;
M ¼ number of EV/load profile.
The time of stay for the kth EV belonging to both jth load profile and ith fleet,
DTi;j;k , can be computed by

DTi;j;k ¼ Treq Tarr i;j;k (18.3)
The power is the energy per unit time, so (18.2) can be adapted for calculating
the total EVs power during discharging state and rewritten as
" !#
XN XL XM
li;j;k Ei;j;k
Ptotal ¼
i¼1 j¼1 k¼1
DTi;j;k (18.4)
i ¼ 1; . . . :; N ; j ¼ 1; . . . : ; L; k ¼ 1; . . . :; M
18.4 V2G challenges

Economic and technical challenges might delay V2G implementation in the real
world as described in the following.
Economic challenges:
● V2G should be in competition with the other deployed ESS technologies (e.g.
pumped hydroelectric, flywheel, concentrating solar power).
● EV battery technology is not preferable to be operated in a frequent switch
between charging/discharging modes. In addition, it is expensive and that
negatively impacts on V2G from becoming a reality.
● Minimization of the overall charging cost is based on the time-of-use/
real-time price.
● It is not clear yet whether the economic incentives justify V2G from the uti-
lity’s perspective. Several issues (e.g. complexity of distribution system
required, battery technology, lack of support for smart grid technologies) may
slow the progress in the deployment of V2G.
● For large-scale EV adoption, batteries have a lack of either low energy density
or power capacity. In addition, batteries need to be large to allow the EV for
distant driving and that makes them heavy, unreliable and expensive.
● The risk of choosing inadequate design, number and location of charging sta-
tions is high.
Technical challenges:
● A reliable two-way communication-enabled infrastructure is greatly needed.
Its effort is anticipated to be high and delays could heavily affect the perfor-
mance of the provided services.
● Two-way power flow control is the key enabling technology that will make
V2G become true.
● Bi-directional inductive power interface should be proposed to facilitate
V2G technology and identify the impact of communication architecture on
the V2G ancillary services in terms of availability, reliability and value of
vehicle-provided ancillary services.
● Controlling the vehicle charging rate remotely based on grid conditions and
user preference as well as reducing additional load at peak times.
● Ensuring high power quality from the sources to allow easier integration of
renewable sources.
References
[1] Su W., Eichi H. R. and Zeng W. ‘A survey on the electrification of trans-
portation in a smart grid environment’. IEEE Transactions on Industrial
Informatics. 2012;8(1):1–10.
[2] Galus M. D., Vayá M. G., Krause T. and Andersson G. ‘The role
of electric vehicles in smart grids’. WIREs Energy Environment. 2013;2(4):
384–400.
[3] Pang C., Dutta P. and Kezunovic M. ‘BEVs/PHEVs as dispersed energy
storage for V2B uses in the smart grid’. IEEE Transactions on Smart Grid.
2012;3(1):473–82.
[4] Quirés-Tortés J., Ochoa L. and Butler T. ‘How electric vehicles and the grid
work together’. IEEE Power and Energy Magazine. 2018;16(6):64–76.
[5] Lu C. F., Liu C. C. and Wu C. J. ‘Dynamic modelling of battery energy
storage system and application to power system stability’. IEE Proceedings
of Generation, Transmission and Distribution. 1995;142(4):429–35.
[6] Su W. and Chow M. ‘Sensitivity analysis on battery modelling to large-scale
PHEV/PEV charging algorithms’. IECON 2011 - 37th Annual Conference of
the IEEE Industrial Electronics Society; Melbourne, VIC, Australia, Nov
2011, pp. 1–6.
[7] Yao L., Damiran Z. and Lim W. H. ‘Optimal charging and discharging
scheduling for electric vehicles in a parking station with photovoltaic system
and energy storage system’. Journal of Energies. 2017;10(4):1–20.
[8] Moses P. S., Deilami S., Masoum A. S. and Masoum M. A. S. ‘Power quality
of smart grids with plug-in electric vehicles considering battery charging
profile’. IEEE OES Innovative Smart Grid Technologies Conference Europe
(ISGT Europe); Gothenburg, Sweden, Oct 2010, pp. 1–7.
[9] Shabaan M. F., Eajal A. A. and El-Saadany E. F. ‘Coordinated charging of
plug-in hybrid electric vehicles in smart hybrid AC/DC distribution
systems’. Renewable Energy, Elsevier. 2015;82:92–9.
[10] Capasso C., Iannuzzi D. and Veneri O. ‘DC charging station for electric and
plug-in vehicles’. The 6th International Conference on Applied Energy
(ICAE2014); Energy Procedia, Elsevier. 2014;61:1126–9.
[11] Vasirani M., Kota R., Cavalcante R. L. G., Ossowski S. and Jennings N. R.
‘An agent-based approach to virtual power plants of wind power
generators and electric vehicles’. IEEE Transactions on Smart Grid.
2013;4(3):1314–22.
[12] Pudjianto D., Ramsay C. and Strbac G. ‘Virtual power plant and system
integration of distributed energy resources’. Renewable Power Generation.
2007;1(1):10–16.
[13] Mwasilu F., Justo J. J., Kim E. K., Do T. D. and Jung J. W. ‘Electric vehicles
and smart grid interaction: A review on vehicle to grid and renewable energy
sources integration’. Renewable and Sustainable Energy Reviews, Elsevier.
2014;34:501–16.
[14] Kempton W. and Letendre S. E. ‘Electric vehicles as a new power source for
electric utilities’. Transportation Research Part D: Transport and Environment,
Elsevier. 1997;2(3):157–75.
[15] Tomic J. and Kempton W. ‘Using fleets of electric-drive vehicles for grid
support’. Journal of Power Sources. 2007;168(2):459–68.
[16] Kempton W. and Tomic J. ‘Vehicle-to-grid power fundamentals: Calculating

capacity and net revenue’. Journal of Power Sources. 2005;144(1):268–79.
[17] Brooks A. N. Vehicle-to-Grid demonstration project: grid regulation,
ancillary service with battery electric vehicle. California Air Resources
Board and California Environmental Protection Agency report, 2002.
[18] Kempton W. and Tomic J. ‘Vehicle-to-grid power implementation: From
stabilizing the grid to supporting large-scale renewable energy’. Journal of
Power Sources. 2005;144(1):280–94.
[19] Shakoor S. A., Black M., Pudjianto D. and Boppc T. ‘Impact of wind gen-
eration on the operation and development of the UK electricity systems’.
Electric Power Systems Research. 2007;77(9):1214–27.
[20] Guille C. and Gross G. ‘A conceptual framework for the vehicle to grid
(V2G) implementation’. Energy Policy. 2009;37(11):4379–90.
[21] Rı́os A. D. L., Goentzel J., Nordstrom K. E. and Siegert C.W. ‘Economic
analysis of vehicle-to-grid (V2G)-enabled fleets participating in the regula-
tion service market’. Proceedings of 2012 IEEE PES Innovative Smart Grid
Technologies (ISGT); Washington, DC, USA, Jan 2012, pp. 16–24.
[22] Geidl M., Koeppel G., Favre-Perrod P., Kloeckl B., Andersson G. and
Froehlich K. ‘Energy hubs for the future’. IEEE Power and Energy
Magazine. 2007;5(1):24–30.
[23] Sekyung H., Soohee H. and Sezaki K. ‘Development of an optimal vehicle-
to-grid aggregator for frequency regulation’. IEEE Transactions on Smart
Grid. 2010;1(1):65–72.
[24] Shimizu K., Masuta T., Ota Y. and Yokoyama A. ‘Load frequency control in
power system using vehicle-to-grid system considering the customer con-
venience of electric vehicles’. Proceedings of 2010 International
Conference on Power System Technology; Hangzhou, China, Oct 2010,
pp. 1–8.
[25] Tian W. Q., He J. H., Niu L. Y., Zhang W. G., Wang X. J. and Bo Z. Q.
‘Simulation of vehicle-to-grid (V2G) on power system frequency control’.
Proceedings of IEEE PEST Innovative Smart Grid Technologies
Conference; Tianjin, China, May 2012, pp. 1–4.
[26] Baboli P. T., Moghaddam M. P. and Fallahi F. ‘Utilizing electric vehicles on
primary frequency control in smart power grids’. Proceedings of IPCBEE
Petroleum and Sustainable Development Conference; Singapore, May 2011,
pp. 6–10.
[27] Ota Y. ‘Autonomous distributed vehicle-to-grid for ubiquitous power grid
and its effect as a spinning reserve’. Journal of International Council on
Electric Engineering. 2011;1(2):214–21.
[28] Almeida P. M. R., Lopes J. A. P., Soares F. J. and Vasconcelos M. H.
‘Automatic generation control operation with electric vehicles’. Proceedings
of IREP Symposium; Rio de Janeiro, Brazil, Apr 2010, pp. 1–7.
[29] Ota Y., Taniguchi H., Nakajima T., Liyanage K. M., Baba J. and Yokoyama
A. ‘Autonomous distributed V2G (vehicle-to-grid) satisfying scheduled
charging’. IEEE Transactions on Smart Grid. 2012;3(1):559–64.
[30] Kempton W., Tomic J., Letendre S., Brooks A. and Lipman T. Vehicle to
grid power: Battery, hybrid, and fuel cell vehicles as resources for dis-
tributed electric power in California. California Air Resources Board and
California Environmental Protection Agency report, 2001.
[31] Mullen S. K. Plug-in hybrid electric vehicles as a source of distributed
frequency regulation. PhD dissertation, US, University of Minnesota, 2009.
[32] Pillai J. R. and Jensen B. B. ‘Integration of vehicle-to-grid in the Western
Danish power system’. IEEE Transactions on Smart Grid. 2011;2(1):12–19.
[33] Donadee J. and Llic M. ‘Stochastic co-optimization of charging and fre-
quency regulation by electric vehicles’. Proceedings of North American
Power Symposium (NAPS); Champaign, IL, USA, Sep 2012. IEEE; 2012,
pp. 1–6.
[34] Xie F. X., Huang M., Zhang W. G. and Li J. ‘Research on electric vehicle
charging station load forecasting’. Proceedings of International Conference
on Advanced Power System Automation and Protection; Beijing, China, Oct
2011. IEEE; 2012, pp. 2055–60.
[35] Vayá M. G. and Anderson G. ‘Combined smart-charging and frequency
regulation for fleets of plug-in electric vehicles’. IEEE PE, General Meeting;
Vancouver, BC, Canada, July 2013. IEEE; 2013, pp. 1–5.
[36] Mu Y., Wu J., Ekanayake J., Jenkins N. and Jia H. ‘Primary frequency
response from electric vehicles in the Great Britain power system’. IEEE
Transactions on Smart Grid. 2013;4(2):1142–50.
[37] Wu C., Mohsenian H., Huang J. and Jatskevich J. ‘PEV-based combined
frequency and voltage regulation for smart grid’. IEEE-PES Innovative
Smart Grid Technologies (ISGT); Washington, DC, USA, Jan 2012. IEEE;
2012, pp. 1–6.
Chapter 19
Microgrids: modelling and control
19.1 Introduction
The global warming and other environmental hazards of conventional sources of
electrical energy push the energy sector continuously towards distributed energy
resources (DERs), especially renewable sources [1]. The fluctuating power pro-
duced by renewable sources such as wind and solar photovoltaic (PV) systems may
bring challenges in generation/load power balance. With increasing penetration of
renewable energy, the function of such distributed generation (DG) is changing
from an auxiliary role to a primary role in the energy sector. DERs, renewable or
non-renewable, and energy storage systems (ESSs) are integrated into what is
known as a microgrid (MG). An MG deals mainly with the issues associated with
the integration of such intermittent renewable energy sources (RESs) and ESSs. It
gives many advantages not only for power balance but also for environmental
challenges, economic benefits and grid reliability requirements [2]. Moreover, MG
handles many technical issues in decentralized form [3,4]. There are intensive
interests to develop and study the future horizon of MG concept throughout the
world [4]. Generally, MG is a small-size, discrete electricity framework consisting
of a collection of DG units and loads. It may operate in grid-connected or islanded
mode and provide seamless transition between the two modes. MG may also
include conventional generators such as diesel-based synchronous generators to
mitigate the effects of the intermittency nature of RESs. Furthermore, many RESs
act as inverter-based sources as inverters are used to interface them with the host
grid or AC loads.
Many issues pertain to the operation of an MG and the integration of DERs in
an MG. Critical among these are the following: (i) the stable operation of MG at
different modes of operation and loading conditions during normal and after fault
provoked islanding incidents; (ii) the impact of different inverter control strategies
and schemes on the dynamic performance of MG; and (iii) the power quality (PQ)
that should be in compliance with the standards as the interfacing power con-
ditioning systems (PCSs) and non-linear loads may cause negative impact on PQ.
Integration of DERs, constituting an MG, entails a control strategy that can be
categorized into three levels: (i) primary control, which is a local control for each
source and each load in the MG – it is based on the local measurements of the
parameters necessary for managing and controlling the MG; (ii) secondary control
that is located at the point of common coupling (PCC) and is responsible for MG
operation according to the current mode of operation, grid connected or isolated;
and (iii) tertiary control, which is the highest level of control. It coordinates mul-
tiple groups of integrated DERs (i.e. multiple MGs) interacting with one another in
the system and communicates requirements from the utility grid.
Primary and secondary controls are associated with the MG operation itself,
whereas the tertiary control is considered as part of the utility or host grid.
Regardless of the kind of operation mode, MG components should be inter-
operable and with interfaces complying with functional standards defined by the
MGs as well as the secondary control should enable the MGs to manage them-
selves and operate autonomously. One of these standards that must be met is the
MG PQ especially when implying integrated RESs. Harmonics distortion and
voltage stabilization are listed as major PQ issues in the literature because of
load non-linearities and renewable sources’ uncertainties. Nowadays, industrial
and commercial loads have numerous electronic drives, information and com-
munication devices, and computers as well as PCSs are used for RESs interface.
All these devices have a catalyst nature to produce harmonics and cause PQ
degradation. Passive, active and hybrid power filters can be used for mitigating
harmonic distortion. Meanwhile, multifunctional voltage source inverters or
flexible alternating current transmission system (FACTS) devices can be utilized
to process these problems that may yield harmonic distortion and/or voltage
instability in the MG. In the next sections, the methodology of using PQ
enhancement device (PQED) to mitigate the harmonic distortion and keep the
voltage at its nominal value at the PCC is described. It can be summarized by the
following steps:
● Specifying the MG architecture as an integration of DERs.

● Identifying the mathematical model for each of MG components.
● Constructing the simulation model of the MG.
● Selecting the type of PQED to be embedded at the PCC.
● Examining the PQ aspects of the MG with/without PQED.
19.2 MG architecture
Three DERs, solar PV, wind energy system (WES) and proton-exchange membrane
fuel cell (PEMFC), in addition to a battery energy storage system (BESS) with data
given in Tables 19.1–19.4 are integrated into an MG as shown in Figure 19.1. Any
additional data are provided where needed. As a primary control level, a local
control (LC) is connected to each source and each load, which in turn is connected
to the MG central controller (MGCC) as a secondary control level. The LC consists
of a breaker for switching operation and maximum power point tracking (MPPT)
circuit. It is used for achieving reliable and economical operation as well as inter-
acting with the distribution system operator (DSO) through the MGCC to access
the data of PQ requirements (e.g. electricity, fuel costs and emission mitigation).
Microgrids: modelling and control 487
Table 19.1 Parameters of PV module
Open-circuit voltage, Voc (V) 32.9

Short-circuit current, Isc (A) 8.21
Module voltage at the MPP, Vm (V) 26.3
Module current at MPP, Im (A) 7.61
Maximum power, Pm, of a PV module (W) 200
Reference temperature ( C) 25
Reference solar radiation (W/m2) 1,000
Number of series PV modules in a string 8
Number of parallel connected PV strings 63
Table 19.2 Wind generation data
Rated power (kW) Vci (m/s) Vco (m/s) Vr (m/s) Hub height (m) Swept area (m2)
60 2.5 16 12 12 10.235
Table 19.3 PEMFC parameters
Type Stake nominal power (kW) Nernst voltage (En) (V) FC resistance (W)
PEM 50 1.1342 0.66404
Table 19.4 BESS parameters
Capacity Efficiency Minimum Maximum Maximum Maximum

(Ah) (%) charge (%) charge (%) charging rate discharging
(kW/h) rate (kW/h)
2,160 85 20 80 40 20
The MGCC manages the operation of DERs and leads integration with the goal of
the following:
● Providing the individual power and voltage set point for each power flow
according to DER control strategy.
● Satisfying the electrical loads’ requirements.
● Alleviating emissions and network losses.
● Exploiting the operational efficiency of the DERs.
DERs are connected to commercial and residential loads at 380 V, 50 Hz. A
PQED such as filter, compensator or multifunctional device (e.g. filter-compensator)
can be embedded into the MG at the PCC.
19.3 Modelling and simulation of MG components

The components of the MG shown in Figure 19.1 should be mathematically mod-
elled and then simulated as in the following to study how to improve the PQ of an
MG with integrated RESs.
19.3.1 PV system
The equivalent circuit and mathematical model of the PV module and array are
explained in Chapter 3, Sections 3.2.5 and 3.2.6. The I–V and P–V characteristics
are also described in Chapter 3, Section 3.2.7. It is to be noted that PVs have non-
linear characteristics, where the performance and output power are directly affected
by the change of operating conditions. The output power of PV is directly pro-
portional to the amount of solar irradiance falling on it, and inversely proportional
to its temperature. As the temperature and solar irradiance change, the point at
which the power is maximum will accordingly change. This means that a DC–DC
converter (boost converter) must be used to change the array terminal voltage with
a goal of MPPT.
Real values of PV module parameters are given for Kyocera KC 200 GT,
manufactured by Kyocera [5]. I–V and P–V characteristics of the PV module
at different irradiance levels and constant temperature (25 C) are shown in
Figure 19.2.
The output power of PV arrays is mainly influenced by the irradiance (amount
of solar radiation) and temperature. Moreover, for a certain irradiance and
Utility grid
Communication link
DSO
Tr.
MGCC
PCC PQED
Source
bus
S1 S2 S3 S4 S5 S6 S7 Switch S8 S9 S10 S11 Load bus
~
LC LC LC LC
LC LC
FC PV solar system BESS Commercial load Residential load

WES
Figure 19.1 Layout of MG structure

PV array Kyocera KC 200 GT, 8 series, 63 parallel strings

800
1,000W/m2
800 W/m2
Current (A)
400
600 W/m2
200 400 W/m2
200 W/m2
0
0 50 100 150 200 250 300
I–V characteristic Voltage (V)
×104
10 1,000 W/m2
800 W/m2
Power (W)
600 W/m2
5 400 W/m2
200 W/m2
0
0 50 100 150 200 250 300
P–V characteristic Voltage (V)
Figure 19.2 PV characteristics
temperature, the output power of the PV array is a function of its terminal voltage
and there is only one value for the PV’s terminal voltage at which the PV panel is
utilized efficiently. The procedure of searching for this voltage is called ‘MPPT’.
Recently, several algorithms have been developed to achieve the maximum power
point (MPP) such as perturb and observe (P&O), incremental conductance, fuzzy or
neural-based techniques [6,7]. However, the insulation levels and the cell tem-
perature determine only the limits of the best obtainable matching. The array vol-
tage determines the real matching, which can be improved by using the MPPT
controller to locate the local MPP in the P–V characteristic of the solar panel [8].
MPPT: The PV panel has a non-linear characteristic and its output power
depends mainly on the irradiance (amount of solar radiation) and the temperature.
Moreover, at the same temperature and irradiance the output power of a PV panel is
a function of its terminal voltage. The procedure of seeking the voltage where the
power is maximum can be obtained either in a single stage or in a double stage. In a
single stage, a DC–AC converter is employed, whereas in a double stage DC–DC
and DC–AC converters are utilized. PV characteristics, Figure 19.2, show that the
MPP for the panel lies at values approximately between 75% and 80% of array’s
open-circuit voltage. Consequently, the MPPT algorithm scans the P–V curve at
predefined voltage of 75% of the array’s open-circuit voltage. The MPPT technique
is associated with the DC–DC converter, which interfaces the PV array to the MG
through an inverter. This can be achieved by controlling the input voltage of the
DC–DC converter (Figure 19.3).
MPPT system tracks the new MPP in the corresponding curve whenever
temperature and/or insolation variation occurs. MPPT is used for extracting the
DC/DC converter
Iout
To inverter
Vout
IPV
VPV MPPT
Figure 19.3 Block diagram of MPPT system
maximum power from the solar PV array to be transferred to the MG. A DC–DC
(step-up/step-down) converter acts as an interface between the inverter and the
array. The MPPT changes the duty cycle to keep the power transfer from the solar
PV array to the MG at the maximum value [8,9].
P&O algorithm: It is widely used because of its ease of implementation and
its low cost. It is based on the following criterion: in Figure 19.4(a), if the
operating voltage is perturbed in a given direction and dp/dV > 0, it indicates that
the perturbation moves the operating point towards the MPP and the P&O
algorithm would then continue to perturb the operating voltage in the same
direction [10–12].
Otherwise, if dp/dV < 0, the change in operating point moves that point away
from the MPP and the P&O algorithm reverses the direction of the perturbation. In
other words, the system works by increasing or decreasing the operating voltage
and observing its impact on the output power.
P&O has been implemented to extract maximum power at each instant. Its
control action is summarized by the steps shown in Figure 19.4(b). The operating
voltage is perturbed with every MPPT cycle. As soon as the MPP is reached, it will
oscillate around the desired operating voltage and the perturbation is narrowed
down until the oscillations are damped.
Boost converter: Based on the description of boost converters in Chapter 13,
Section 13.3.2, the simulation model of the boost converter control and the local
control action of the PV system are illustrated in Figures 19.5 and 19.6, respec-
tively. Accordingly, the MATLAB/Simulink model of the PV system including
the elements, PV array, boost converter, MPPT system and the local control are
dP/dV
dP/dV==00
Power (W)
dP/dV > 0 dP/dV < 0
MPP
(a) Voltage (V)
Measure V(k), I(k)
Calculate P(k) = V(k) × I(k)
ΔVref (k) = Vref (k) − Vref (k − 1)

ΔP(k) = P(k) − P(k − 1)
Yes
P(k) = P(k − 1)
No
No Yes
ΔP(k) < 0
No Yes No ΔVref (k) < 0 Yes

ΔVref (k) < 0
Increase Vref Reduce Vref Reduce Vref Increase Vref
Return
(b)
Figure 19.4 P&O application for MPPT: (a) searching mechanism and
(b) flow chart of P&O algorithm
19.3.2 Wind energy system

Wind turbine characteristics are discussed in Chapter 2. In particular, the generated
wind power is explained in Sections 2.2 and 2.3. In addition, the wind turbine
generator, in this study, is a permanent magnet synchronous generator (PMSG),
which is also explained in Section 2.4.2.
However, the following relations are rewritten for convenience.
The output power of the wind turbine is given by
8
> 0 for Vw < Vci
>
> 3
< aVw bPr for Vci < Vw < Vr
Pw ¼ (19.1)
>
> Pr for Vr < Vw < Vco
>
:
0 for Vw > Vco
V_PV I_PV
PV model
273 15 Add
Constant + + Duty To inverter
+ T +ive out
Temp I
+V + – +ive
– IPV
G –
–ive out
–ive
Irradiance PV model
DC/DC boost converter
Vmpp
+V –
Vmpp Duty dc
VBC
P&O
Calculation
Figure 19.5 Boost converter control model
Temperature,
irradiance
0° ≤ T ≤ 85°
& No
50 ≤ IR ≤ 1,200
Breaker 2: open
Yes
MPPT
Breaker 2: close
Ppv
Figure 19.6 Local controller of PV system
Pr V3
where a ¼ , b ¼ V 3 ci V 3 , Pr is the rated power of the wind generator, and
ðVr3 Vci3 Þ ð r ci Þ
Vci , Vco and Vr are the cut-in, cut-out and rated speeds of the wind turbine,
respectively. The extracted real power from wind turbine is expressed as

1 Rwopt 3
Pw ¼ rACp ðl; bÞ (19.2)
2 lopt
MPPT controller
MPPT using P&O
parameters technique
Param Param V_PV
Enabled 1
D z D
Enable V_PV V P&O
MPPT [I_PV] I
+v
−
Boost converter
(average model) c 2
1 + i 1
Irradiance −
m_PV D Boost + Conn1 ×
(W/m2) 1 Irr_PV
(average)
Ramp-up/down A −
Ir L1 PV
Temp -K– P_PV

W2kW1
Irradiance 2 Conn2
(W/m2) 2
100-kW PV array
330 * SunPower SPR-305E-WHT-D (Nser = 5 Npar = 66) P_PV
Pmpp @ 1,000 W/m2, 25° = 100.7 kW @ 273.5 V
Temperature Pmpp @ 250 W/m2, 25° = 24.4 kW @ 265.1 V
(°C) Pmpp @ 1,000 W/m2, 50° = 92.9 kW @ 250.2 V
Pmpp @ 1,000 W/m2, 0° = 107.5 kW @ 296.6 V Agent-1
Irr_PV
INITIAL
Irr_PV Agent 1 T_PV
Param
L9
T_PV
Figure 19.7 Model of PV system in MATLAB/Simulink

where r is density of the air (r ¼ 1.2250 kg/m3), A is the blade swept area, R is the
length of the blade, wopt is the optimum speed of wind turbine and lopt is the
optimum tip-speed ratio. The power coefficient Cp is a function of the pitch angle b
and the tip speed ratio l, which is given by
wR
l¼ (19.3)
Vw
where w is the rotational speed of the wind turbine and Vw is the wind speed. P&O
algorithm is used for MPPT, and real WES module data (E3120) manufactured by
Endurance are given in Table 19.2 [13].
The action of the local controller and the corresponding simulation model
of the WES including PMSG and P&O technique for MPPT are depicted in
Figures 19.8 and 19.9, respectively.
19.3.3 Proton-exchange membrane fuel cell

The equivalent circuit and mathematical modelling of PEMFC are described in
Chapter 9, Sections 9.6 and 9.7, respectively. P&O algorithm is applied based on
measuring the current and voltage variables, IFC and VFC, at each instant to calcu-
late the corresponding power, PFC(t), that being compared to the power at the
preceding instant, PFC(t 1) ¼ IFC(t 1) VFC(t 1). The algorithm continues to
perturb the system in the same direction of reference voltage increase when the
Wind speed Vw
No
4 ≤ Vw ≤ 24
Yes
MPPT Breaker 1: open
Breaker 1: close
Pw
Figure 19.8 Local controller of WES

Power 4
V
Dis1 × -K-
I
Dis2 W2kW1
V-I PV power
CP
Vacw Vac-Iac Vac-Iac1 Wp
Torque Lamda
Gspeed
Generator speed (p.u.) Tm (p.u.) Iacw Boost converter CP-Lamda
control (MPPT) 1
Pitch Pitch angle (°) cp CP
RMS Pdcw
Wind speed (m/s) lamda Lamda V V_wind
(discrete) Vab_rms VRmsw
Dis3
Wspeed Vab3
Wind turbine 1 I I_wind Duty
Vab
Dis4 C2
+ v IRmsw Idcw
RMS Enable i
− Vab I D +
(discrete) line_rms C1 + −
Tm Boost
IL B2 Enable Rectifier FW diode 1
Vab4 + i
MPPT − A (average) −
A + i ILine A 1
− L2 Vdc1
+ Vdc2 Boost converter 1 +ve
m_Wind m N +
+ (average model)
B B V v Vdcw
− v −
S
C C −
2
–ve
A 1.2 Gspeed
B Generator speed (p.u.)
C 0 Pitch
Pitch angle 1
Wspeed
Speed 1
12
Wind speed and pitch angle
Figure 19.9 MATLAB/Simulink model of PMSG with MPPT using P&O technique
power and voltage variations are positive. Otherwise (when the variations are
negative) the direction of perturbation is reversed. The flow chart of P&O algo-
rithm is shown in Figure 19.10 and the corresponding simulation model is illu-
strated in Figure 19.11.
19.3.4 Battery energy storage system

The battery bank stores the surplus electrical energy in charging state and delivers
energy in discharging state when the consumption is higher than the energy pro-
duced by the DERs. The most frequent type used is the lead acid battery, which can
be modelled as described in Chapter 10, Section 10.3.
The goal of the battery local controller is to control the charging and dischar-
ging of the battery according to the available renewable power, state of charge
(SoC) and load demand. Real parameters of BESS model (S2-2160GEL) manu-
factured by Rolls battery are given in Table 19.4 [14]. The local control action
follows the steps indicated by the flow chart in Figure 19.12. Thus, the simulation
model of the BESS with LC can be drawn as depicted in Figure 19.13 as well as the
charger controller construction is shown in Figure 19.14.
19.3.5 Load profiles

Real load data of Hurghada city, Egypt, are given. It includes the load curves for
active and reactive power of (i) commercial load, Figure 19.15(a) and (b), and
(ii) residential load, Figure 19.15(c) and (d).
Start
Measure IFC, VFC

PFC (t) = IFC (t) × VFC (t)
PFC (t − 1) = IFC (t − 1) × VFC (t − 1)
Yes
PFC (t) − PFC (t − 1) = 0
No
No Yes
PFC(t) − PFC(t − 1) > 0
No Yes No Yes
VFC(t) − VFC (t − 1) > 0 VFC (t) − VFC (t − 1) > 0
VFC ↑ VFC ↓ VFC ↓ VFC ↑
Figure 19.10 P&O application to MPPT for PEMFC

MPPT
parameters
V_FC
Param
D
MPPT control 1
Param DFC
Va Duty
Enable Enabled 1
MPPT D z
m_FC <Voltage> V P (kW)
P&O
Ia V_FC
<Current> I
I_FC Vdc
I_FC
× -K-
Idc
+ I-V scope 2 FW diode 2 W2kW1 I-V scope 3
i I-V scope 1
m + – D + + i FC
Boost − 1
ipvf ipvf1
(average) +ve
A −
H2 Air +
Boost converter 1 +
(average model) 1 V
– −
– I-V scope 4
vpvf2
Fuel cell stack
2
–ve
Figure 19.11 PEMFC system simulated in MATLAB/Simulink

Measure
current and
voltage
Calculate SoC
No
40% ≤ SoC ≤ 95%
Yes
Change over Change over

switch = 0 switch = 1
PBattery
Figure 19.12 Local controller of BESS
Gate
S1 Vdc Vdc
S2 SOC
Gate Ibat I (A)
IBatt
Battery/electrolyser V (V)
VBatt
controller Batt
Va (V)
[Va_B1]
Ia (A)
Ibat [Ia_B1]
Grid
S2
SOC
Bp Bp 1 IBatt
Goto +ve
+ + –i Mean
S1
IL3
– 2 Mean value 1 Out 4
–ve
Battery and DC–DC converter 1
Figure 19.13 MATLAB/Simulink model of BESS including local controller

500
+ PI + Boolean NOT
− − × 2 Scope 8
Data type S2
Discrete conversion Product 2
1 PI controller
Vdc 3 1
Ibat S1
SOC
2 -K-
≤
Scope 4
.9
Scope 5
-K- +− PI
Discrete ≥
PI controller 1
1 ramp_out 3
AND Gate
Ramp_generator
-K-
> Plot 1
.8
Figure 19.14 Charger controller construction of the BESS in MATLAB/Simulink

×104
×104 2.3
Commercial load reactive power, Q (VAR)

4.2
Commercial load active power, P (W)
4 2.2
3.8
2.1
3.6
3.4 2
3.2
3 1.9
2.8
1.8
2.6
2.4 1.7
2.2
2 1.6
0 5 10 15 20 0 5 10 15 20 25
(a) Time (h) (b) Time (h)
×104 ×104
5.5 2.2
Residential load reactive power, Q (VAR)
Residential load active power, P (W)
5
2.1
4.5
2
4
1.9
3.5
3 1.8
2.5 1.7
2
1.6
1.5
1.5
1
0 5 10 15 20
Time (h) 1.4
(c) 0 5 10 15 20
(d) Time (h)
Figure 19.15 Load profile curves of active and reactive power for (a and b)
commercial load and (c and d) residential load
Combining the models of individual components of the MG, the integrated

model of the MG in MATLAB/Simulink is shown in Figure 19.16. However, for
better and clearer illustration, the model has been divided into two parts as shown
in Figure 19.17(a) and (b).
Applying the simulation model of the MG with the load profiles shown in
Figure 19.15, the frequency spectrum and the voltage variation at the PCC are shown
in Figure 19.18(a) and (b). It is to be noted that the total harmonic distortion of the
voltage, THDV, without a PQED is 19.07%, whereas the voltage is approximately
constant at 1 p.u.
19.4 Application of D-STATCOM

D-STATCOM is applied as a PQED. It is one of the most popular fast-response
D-FACTS topologies that is used to enhance poor PQ issues throughout various
Vadc
Discrete Vabc_BL1
Uref Iadc
Ts_Power s.
Iabc_BL1
Vdc
Ts_Power = 50 μs Vdc
+ve Ts_Control = 100 μs Out 3 Out 4
P (kW) Loads
Measurements
−ve A a
Gate B b
PQED
Battery and DC–DC converter VSC
DC bus C c
+ve A B C
L3 Iabc_PCC
(average model)
−ve Vabc_PCC
I batt (A) a b c BL2
PCC
60-kW WES V_batt (V) Battery AC bus
(W/m2) Vabc_1 Out5
Irr PV
+ve
T PV
Temp (°C)
Vabc_H BL4 labc_1
P PV
(kW)
V_PV (V) Pind
V PV labc_H
−ve P (kW) Batt Bp
Duty cycle Pres Pind
D pres Group 1 Out5
PV
100-kW PV Wp Pind
P (kW) wind Out 2
V dc Out13 m
PQ
+ve Qres
P (kW) wind 1 m Group 1
Qind
3-phase Out11
−ve Pres Out 9 residential load Out6
Pind Out14
50-kW FC
Figure 19.16 Integrated model of MG in MATLAB/Simulink

Vabc_prim
Vabc_BL1
Discrete Uref Iabc_prim
+ve Ts_Power s. Iabc_BL1
× -K- Vdc_mes Vdc
−ve Ts_Power = 50 μs W2kW1

Ts_Control = 100 μs Out 3 Out 4
P (kW) LOADS
Battery and DC–DC converter
Uref A a
+ve + I
− +
Gate A B b
g
C
−ve +V Vdc VSC B
E
− C c
60-kW WES −
PQED
DCbus L3 C Iabc_PCC
A
VSC Vabc_PCC
a
+ve
I (A) (average model)
IBatt
V (V)
VBatt PCC BL4
−ve Batt ACBus
2 Vabc_H
100-kW PV Ir (W/m )
Irr_PV
Temp (°C)
T_PV
+ve Iabc_H
Pmean_PV (kW) Group 1
P_PV
V_PV (V) Signal 1
V_PV
Duty cycle
−ve D Pres
PV
50-kW FC Group 1 PQ
Signal 1 A
(a)
Figure 19.17 MG simulation model split into two parts: (a) Part I and (b) Part II
Measurements
A a
B b
C c
PQED
Iabc_PCC
a
c
b
B
C
A
Vabc_PCC
B
C
a
c
b
BL2
PCC BL4
ACBus
Vabc_H Out 5
Vabc_I
Iabc_H Iabc_I
Group 1 P_res P_ind.mat
Signal 1
.5
Pres P_ind
Group 1 Out 15
Out 13 Signal 1
Group 1 PQ PQ
Signal 1 A A
m Pind m
B B
Qres
C C
Three-phase Out 11 Group 1
residential load Signal 1
Out 9
Pres.mat Qind
.3
Out 6
P_ind.mat Out 14
.7
(b)
Figure 19.17 (Continued )

Fundamental (50 Hz) = 372.5, THD = 19.07% 1.2

Without D-FACTS
Without D-FACTS
10
1
9
Mag (% of fundamental)
PCC bus voltage (p.u.)

8
0.8
7
6
0.6
5
4
0.4
3
2 0.2
1
0 0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 0 0.1 0.2 0.3 0.4 0.5
Harmonic order
Time (s)
(a) (b)
Figure 19.18 Dynamic response: (a) frequency spectrum and (b) voltage
variation at PCC without PQED
To bus VL
ISS
A
Bridge
rectifier
+
PID controller S
circuit
IGBT switch
Id C
d
Vd
(a)
Vmref
+−
1 1 eVd
VL V RMS 1 + sT3
γVd
mbase
ePd ++
IL 1 1 + eT S
RMS γPd PWM
Imbase PID Limiter
1 + sT3 − +
D3
1 1 + eId
ISS I RMS γId
ssbase 1 + sT3 −
D3
(b)
Figure 19.19 Controller scheme of D-STATCOM: (a) equivalent circuit of

D-STATCOM and (b) three-level error-driven PID of D-STATCOM
system conditions [15]. The equivalent circuit of D-STATCOM and the block
diagram with its three-level error-driven proportional integral derivative (PID)
controller are shown in Figure 19.19(a) and (b). The parameters of the controller
are given in Table 19.5. PID controller gains are determined by trial and error.
Table 19.5 Parameters of D-STATCOM controller
System part Specifications

Vmbase 380 V
Imbase 150 A
Issbase 52.5 A
gVd ¼ gId 1
gPd 0.5
T3 ¼ D3 15 ms
T4 ¼ D4 40 ms
PWM frequency 3,000 Hz
Fundamental (50 Hz) = 373.5, THD = 11.52% 1.2

18 With D-STATCOM (trial and error)
With D-STATCOM (trial and error)
16
1
14
12 0.8
10
0.6
8
6 0.4
4
2 0.2
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 0
Harmonic order 0 0.1 0.2 0.3 0.4 0.5
Time (s)
(a) (b)
Figure 19.20 System response: (a) frequency spectrum and (b) voltage variation
at PCC with D-STATCOM
Their values are Kp ¼ 72.244, Ki ¼ 9.325 and Kd ¼ 2.452. The total error of
D-STATCOM is given by
et ¼ gVd ðeVd Þ þ gId ðeId Þ þ gPd ðePd Þ (19.4)
and the results are represented by both of frequency spectrum and voltage variation
as shown in Figure 19.20(a) and (b).
19.5 Application of SFC
The switched filter compensator (SFC) as reported in [16] has been introduced at
the load bus to enhance the performance of the MG-connected wind system. The
SFC device is a switched/modulated filter formed by a fixed shunt capacitor bank
connected to the AC side of the arm of an uncontrolled rectifier.
The configuration of the SFC is shown in Figure 19.21(a). One mode of
operation is defined for the FACTS device with a controlled switch S installed on
the rectifier side. The switch operation indicates the ON–OFF state of the shunt
capacitor. The ON state of S means that the SFC can compensate the reactive power
and mitigate harmonics because of the series-connected inductance Lf with the
Bus #1 Bus #2
A
B
C
Cf Non-linear
abc load
Rf
Filter
Lf
Rectifier
Error-driven PID
S
g
controller 2
1
IGBT switch
(a)
Voltage loop
VB1 1 1 + eVB1
RMS γV
Vbase 1 + sT1 −
Vmref +
Vmref
VB2 +
1 1
RMS + − γV
Vbase 1 + sT1 eVB2
IB1 eIB1 eA ea
1 1
RMS YI + KA
Ibase 1 + sT1 Limiter
− + eB eb + eT S
D KB PID PWM
+ +
IB2 1 1 + e IB2
eC ec +
RMS YI KC
Ibase 1 + sT1 −
Current loop D
Vmref
VB2 1 1 + eVB2
RMS γV
Vbase 1 + sT1 −
ePL+
X
+ γp
− ++
D
IB2 1 1 + eIB2
RMS γI
Ibase 1 + sT1 −
Power loop D
(b)
Figure 19.21 SFC/PID controller circuit: (a) SFC configuration and

(b) PID controller
capacitor Cf. Filter parameters are selected based on a trade-off between design
goals, and trial and error.
The pulse-width modulation (PWM) switching of the SFC scheme is based on
a dynamic multi-loop error-driven controller that is designed to enhance voltage
stability and PQ, improve power factor (PF) and diminish the effects of voltage
transients and inrush currents.
The primary goals are improving the energy use, stabilizing voltage in a dynamic
manner, enhancing PQ and improving the PF at the common interface AC buses of
the MG-connected DERs. The SFC-switched filter with the modulated capacitive
compensator uses PWM switching strategy triggered by a multi-regulation loop error-

driven controller (Figure 19.21(b)). The controller is comprised of three loops:
(i) voltage stabilization loop to compare the voltage of both sides of the SFC with the
reference value giving the signal error, eA , (ii) current limiting loop to compare the
currents of both sides of the SFC with a reference value giving the signal error, eB ;
and (iii) dynamic power loop to compare the measured voltage, current and calcu-
lated power at Bus #2 with their reference values giving the error eC .
Based on Figure 19.21(b), the total error, eT , which is the input to the PID
controller can be calculated as in the following.
The error, eA ; produced by voltage stabilization loop is given by
eA ¼ gVL ðeVL1 Þ þ gVL ðeVL2 Þ (19.5)
where

1
VL1 1þsT1
eVL1 ¼ Vmref (19.6)
V mbase

1
VL2 1þsT1
eVL2 ¼ Vmref (19.7)
V mbase
Similarly, the error, eB ; produced by current limiting loop can be calculated by
eB ¼ gIL1 ðeIL1 Þ þ gIL2 ðeIL2 Þ (19.8)

IL1 1 1
eIL1 ¼ 1 (19.9)
Imbase 1 þ sT1 1 þ sT2

IL2 1 1
eIL2 ¼ 1 (19.10)
Imbase 1 þ sT 1 1 þ sT2
Finally, the error produced by dynamic power loop can be obtained as
eC ¼ gVL2 ðeVL2 Þ þ gIL2 ðeIL2 Þ þ gPL2 ðePL2 Þ (19.11)
where eVL2 and eIL2 are determined by (19.7) and (19.10), respectively, and ePL2 is
given by

VL2 IL2 1
ePL2 ¼ 1 (19.12)
Vmbase Imbase 1 þ sT2
Thus, the total error eT can be calculated as summation of (19.5), (19.8) and
(19.11), that is,
eT ¼ ea þ eb þ ec (19.13)
where
e a ¼ e A KA (19.14)
1.2
Fundamental (50 Hz) = 379.52, THD = 1.10% With SFC
With SFC (GOA)

1
10

8 0.8
6 0.6
4
0.4
2
0.2
0 0 1 2 3 4 5 6 7 8 9 10 11 12 13
Harmonic order 0
0 0.1 0.2 0.3 0.4 0.5
Time (s)
(a) (b)
Figure 19.22 System response: (a) frequency spectrum and (b) voltage variation
at PCC with SFC
e b ¼ e B KB (19.15)
e c ¼ e C KC (19.16)
where KA ; KB and KC are, respectively, selected weightings for the three regulation
loops to allow adjustment of the control burdens. In addition, SFC-PID gains are
determined by trial and error as Kp ¼ 38.6, Ki ¼ 7.5 and Kd ¼ 9.3.
A PID controller processes the difference between the signals to obtain the
phase angles (delta) required to drive the error to 0. Finally, the control signal of
the PWM of the SFC scheme has the following form u(t) in the time domain for the
PID controller:
ðt
d ð e T ð t ÞÞ
uðtÞ ¼ Kpt eT ðtÞ þ Kit eT ðtÞdt þ Kdt (19.17)
0 dt
where uðtÞis the control variable; eT ðtÞ is the selected system error; and Kp, Ki and
Kd are the PID gains, respectively.
Accordingly, the SFC scheme is applied as a PQED to mitigate the harmonic
distortion and stabilize the voltage at the PCC of the MG using the MATLAB/Simulink
environment. The results illustrate a significant reduction of THDV where it becomes
1.1% rather than 19.07% for the MG without PQED or 11.52% for the MG with D-
STATCOM and the voltage is kept constant at the nominal value (Figure 19.22(a) and
(b)). It may, however, be concluded that the D-FACTSs play an important role in
improving the PQ and ensuring efficient energy utilization with RESs.
References
[1] Justo J. J., Mwasilu F., Lee J. and Jung J. W. ‘AC-microgrids versus DC-
microgrids with distributed energy resources: A review’. Renewable and
Sustainable Energy Reviews, Elsevier. 2013;24:387–405.
[2] Huang W., Lu M. and Zhang L. ‘Survey on microgrid control strategies’.

Energy Procedia. 2011;12:206–12.
[3] Shuai Z., Sun Y., Shen Z. J., et al. ‘Microgrid stability: Classification and a
review’. Renewable and Sustainable Energy Reviews. 2016;58:167–79.
[4] Olivares D. E., Mehrizi-Sani A., Etemadi A. H., et al. ‘Trends in microgrid
control’. IEEE Transactions on Smart Grid. 2014;5:1905–19.
[5] KYOCERA high-efficiency photovoltaic module datasheet [online].
Available from https://www.kyocerasolar.com [Accessed 29 Jan 2020].
[6] Wolf S. and Enslin J. H. R. ‘Economical, PV maximum power point tracking
regulator with simplistic controller’. Proceedings of IEEE Power Electronics
Specialist Conference - PES’93; 1993, pp. 581–7.
[7] Cardona S. and Lopez M. M. ‘Performance analysis of a grid-connected
photovoltaic system’. Energy. 1999;24:93–102.
[8] Esram T. and Chapman P. ‘Comparison of photovoltaic array maximum
power point tracking techniques’. IEEE Transactions on Energy Conversion.
2007;22(2):439–49.
[9] Cheikh S. M. A. and Larbes C. ‘Maximum power point tracking using a
fuzzy logic control Scheme’. Revue des Energies Renouvelables. 2007;
10(3):387–95.
[10] Villalva M. G. and Ruppert F. E. ‘Analysis and simulation of the P&O MPPT
algorithm using a linearized array model’. Proceedings of 35th Annual
Conference of IEEE Industrial Electronics; Porto, Portugal, Nov 2009. IEEE;
2009, pp. 231–6.
[11] Safri A. and Mekhilef S. ‘Incremental conductance MPPT method for PV sys-
tems’. The 24th Canadian Conference on Electrical and Computer Engineering
(CCECE); Niagara Falls, ON, Canada, May 2011. IEEE; 2011, pp. 1–3.
[12] Bennett T., Zilouchian A. and Messenger R. ‘Comparative analysis of the
perturb-and-observe and incremental conductance MPPT methods’. The 8th
International Symposium on Advanced Topics in Electrical Engineering
(ATEE); Bucharest, Romania, May 2013, pp. 1–4.
[13] Endurance Energy MFTG Ltd. Wind turbine models [online]. Available from
https://en.wind-turbine-models.com/manufacturers/247endurance [Accessed
30 Jan 2020].
[14] Rolls Battery Ltd. Battery specifications sheet [online]. 2018. Available
from https://www.rollsbattery.com/battery/6-ehg-31p/?pdf¼7925 [Accessed
30 Jan 2020].
[15] Li S., Li Y., Cao Y., Tan Y. and Keune B. ‘Capacity optimization method of
distribution static synchronous compensator considering the risk of voltage
sag in high-voltage distribution networks’. IET Generation, Transmission
and Distribution 2015;9(16):2602–10.
[16] Gandoman F. H., Sharaf A. M., Abdel Aleem S. H. E. and Jurado F.
‘Distributed FACTS stabilization scheme for efficient utilization of dis-
tributed wind energy systems’. International Transactions on Electrical
Energy Systems 2017;27(11):1–20.
Part IV
Associated systems
Chapter 20
Communications in power systems
20.1 Perspective of future power grid

As explained in Chapter 1, great attention is continuously paid by power engineers
and researchers in making the power grid a smarter grid (SG). The power grid is
getting more complicated in configuration as it includes not only the conventional
elements, such as generation, transmission and distribution systems with uni-
directional power flow, but now also the distributed generation (DG) units and
energy storage systems. DG units can be directly connected to the transmission and/
or distribution system or integrated together constituting microgrid(s), which can
be connected to the distribution system at different locations.
This developing configuration makes the power flow bi-directional rather than
uni-directional. To achieve the desired technical, environmental and economic
benefits, e.g. reliability and flexibility enhancement, emission reduction and power
quality improvement, the power grid must be augmented with assets and systems
that will enable it to operate in an optimal manner keeping customer satisfaction.
Smart metering, communication and cybersecurity systems are of utmost impor-
tance to be incorporated into the power grid with a goal of modernizing the legacy
electricity network and achieving smart characteristics. Cybersecurity is discussed
in Chapter 21, whereas a brief introduction to smart metering and communication
systems is given in this chapter.
Advanced metering infrastructure (AMI). It is a key factor in SG architecture
where advanced meters called smart meters (SMs) are used rather than the con-
ventional meters which just read the power consumption. SMs can record the power
consumption online and can communicate with the utility or other monitoring and
control centres through the power grid. This facilitates (i) the utility to collect
information for load monitoring, billing process, online control and managing
power quality and (ii) the customers to recognize how much power they are using
and accordingly could manage their consumption and, thus, carbon dioxide emis-
sion as well as they can participate in managing the peak load.
Communication systems. SG as an anticipated next generation power grid
should provide monitoring, protecting and automatically optimizing the operation of
the interconnected components in the power grid. It is characterized by bi-directional
connection of electric power and information flows to establish an automated wide
distributed delivery network. In addition, it covers different types of generation
Sending a message
Electrical
Message signal Transmitted signal
Sensing
Transmission medium
element Transmitter
Transmitter Receiver
Message Received signal
Electrical
signal
Receiving a message
Figure 20.1 Main components of electronic communication system
(central conventional, renewable and non-renewable DGs) through transmission and

distribution networks to customers (industrial, commercial and residential/home
users). Each of these elements has its own controls; for instance, renewable sources
are inverter-based sources equipped with power conditioning system, and home users
with their thermostats, electric vehicles (EVs) and intelligent appliances. Based on all
these aspects, two-way communication systems are critically required to be a com-
plementary part of the power grid for information flow in real time from and to each
node to another. Electronic communication system comprises three primary com-
ponents: a source, a destination and a transmission medium, by which four basic
processes are implemented as in the following [1,2]:
● conversion of data into signals in a form that is adequate for transmission;
● modulation of data signal for transmission over selected medium;
● determination of adequate transmission medium; and
● retrieving information for the modulated signals.
A simplified block diagram of an electronic communication system is shown in
Figure 20.1.
SG can be envisioned as an interconnected multi-area system in which each
area can incorporate a communication technology that meets specific requirements
such as reliability, latency, bandwidth and security. Some of these areas are local
area network (LAN), home area network (HAN), building area network (BAN),
neighbourhood area network (NAN) and wide area network (WAN). This classi-
fication, from communication engineering point of view, helps in developing a
Communications in power systems 515
reliable communication infrastructure for establishing robust real-time data trans-

portation through WANs [3].
Communications is a discipline by itself. The intent of this chapter is to simply
provide awareness and a general understanding of the communications systems
from the perspective of a practising power systems engineer rather than an in-depth
treatment of the communications technology.
20.1.1 Classification of communication networks

First, it is important to understand what a network is. It is simply two or more devices
linked with a goal of communicating or sharing resources. This connection can either
be wired through use of cables and telephone lines or be wireless through use of
satellites and radio waves [4,5]. Second, communication networks can be classified
based on the size of the network or the number of devices included in its area, in
addition to the communication technology required for communicating the network
entities. Regarding SGs, a brief description of such networks is given in the following:
LAN: It is a network connecting a cluster of devices in a local area through use
of Ethernet cables with length limited by the communication speed. Wi-Fi can also
be used to make LAN wireless through using a wireless router. LAN is the most
common and basic type of communication network.
HAN: It is a network connecting devices within a user’s home and is used to
provide monitoring and control over energy usage. It connects to the utility’s
electric meter and monitors energy usage to enable the home owner to recognize
how much energy is being used and where. It is also used to give the utility a way to
control, for example, the heating and air conditioning to save energy. HAN also
provides a way to conveniently turn appliances off and on and controlling all smart,
digital and thermostatic devices at home such as washing machine, dish washer,
refrigerator, computers and peripherals, EV and TVs with a goal of saving energy
and optimizing the consumption. HAN may use wireless or communications over
the AC power line, as well as a router to which some home assistants are connected.
A schematic diagram of a general vision for a HAN is depicted in Figure 20.2. It
can be considered as a type of LAN [6].
BAN: A BAN is a LAN that covers an entire building. It may be a collection of
smaller LANs. For example, if each floor is considered a single LAN, then the
combination of each floor LAN is considered a BAN. A BAN is also known as a
business area network [7].
NAN: It is an offshoot of Wi-Fi hotspots and wireless LANs (WLANs) which
enable users to connect to the Internet quickly and at very little expense. A NAN is
generally installed by an individual to serve a family or some of neighbours. NANs
cover only a small number of blocks close to a specific access point. With the help
of an omnidirectional antenna, a single access point could cover a radius of more
than half a mile (0.8 km). Users who wish to connect to a NAN can then make use
of a directional antenna to get an improved signal from the access point.
NAN providers are usually individuals or a group that join to share an Internet
broadband connection. If a user with a broadband connection, either digital
Utility grid
Data
concentrator
Distribution
box gateway
Energy
management
system
Charge/
discharge
SM controller
Power line
Data line
Control signal
Electric
vehicle
Figure 20.2 HAN vision
subscriber lines (DSLs) or cable modem, plans to share it, a NAN makes it possible
to share this with anyone within range. For the receiver (usually neighbours within
reach) to connect wirelessly to the shared Internet, they need to have a Wi-Fi-
enabled laptop. This concept differs from the wireless deployment for hotspots.
Hotspots are usually commercial Internet access points with a reach of only 100 m.
They are used to attract tech-savvy customers to a coffee shop, airport or restaurant.
NANs, on the other hand, offer a much wider radius of Internet connection.
Therefore, commercialization of NANs is an effective way to speed up the exten-
sion of neighbourhood Wi-Fi networks. NANs allow the users to reduce their
Internet expenditure by sharing a connection with neighbours. The downside is that
this strategy reduces the bandwidth speed and sometimes leads to the violation of
the Internet service provider’s agreement. Some service providers do not allow
individual broadband users to share their connection, making NANs a violation of
that agreement. A typical structure of a NAN network is shown in Figure 20.3.
WAN: In contrast to the smaller LAN and HANs, a WAN is a wide area
network, covering any distance necessary. The Internet could be considered a
WAN that covers the entire Earth.
The SG includes many networks (domains) with various boundaries that must be
connected to supply end-to-end services. The challenge is to design network archi-
tectures that can meet the interoperability requirements for inter-domain and intra-
domain communications. Such networks, according to the US Energy Independence
and Security Act of 2007, should supply secure and reliable end-to-end two-way
communications. In this respect, one of the most important challenges is to supply a
reliable last unit distance communication that covers the connectivity from the meters
Data
aggregator
Smart
appliances
Laptop
TV
Figure 20.3 A typical NAN structure
to the AMI headend. The SG communication infrastructure may include different area
communication networks. For instance, an SG that distributes electricity between
generators (conventional and DG sources) and consumers (commercial, industrial and
residential) may form HAN, BAN, NAN and data centres using bi-directional infor-
mation flow for monitoring and controlling the devices as shown in Figure 20.4. The
design of such networks depends not only on the application layer requirement, but
also on the nature of its medium access control. These are some of diverse commu-
nication technologies and standards deployed for NAN [8].
Various technologies of communications in power systems can be used to
transfer the information aggregated by SMs from device to others. Wired and
wireless communication technologies in respect of SG needs are briefly over-
viewed in the next sections [9].
20.2 Wired communications

20.2.1 Power line communications
The technology of power line communication (PLC) system implies the use of
existing power network, which transmits the power from generation sources to a
variety of customers over a wide region, for communication purpose without the
elevated cost and implementation required by the other solutions based on wireless
and optical fibres. Since many years, amplitude-modulation carrier-based com-
munication has been used by the power network operators to transfer status, and
alert messages between power stations and substations as well as operator internal
telephony. Nowadays, with the continuous development of communication tech-
nologies, the PLC can be used for data transmission [10].
The power network includes power lines (overhead and underground cables),
transformers (step-up and step-down) and protection and control devices. Power
DG units (renewable and non-renewable)
DSO
PV
Diesel gen.
Wind set
Generation
Data centre
Conventional power system

NAN
Transmission
Communication line
TSO
Po
BAN
we
r fl
ow
DSO Distribution
HAN
From DGs
Figure 20.4 A typical SG communication infrastructure
lines work at different voltage levels: extra-high voltage (EHV), high voltage (HV),
medium voltage (MV) and low voltage (LV) [11].
● EHV and HV lines: They transmit electricity from generating stations to dis-
tribution substations at voltage levels, typically, of the order of hundreds of
kilovolts over long distances of the order of tens of kilometres. In the begin-
ning, the PLC technology using long-wave frequencies (e.g. in the range of
24–500 kHz) with bandwidth of a few kilohertzs was enough for transmitting
messages or internal telephony. Currently, the availability of optical fibre
communication links in underground cables or overhead lines increases the
bandwidth significantly and allows high data rate.
HV lines are excellent carriers for radio frequency (RF) energy because of
the presence of open wire equipment with very few crossovers. Small trans-
mitted power (~10 W) is enough to overcome distances more than 1,000 km. In
addition, because of the rapid progress of digital communications automation
and through modern digital modulation and coding schemes, the bandwidth
efficiency for the carrier-frequency system has been highly improved.
Typical lines of voltages in the range from 110 to 380 kV are used for nation-
wide or even international power transfer and consist of long overhead lines with
little or no branches. This makes them acceptable wave guides with less attenua-
tion per line length as for their MV and LV counterparts. However, their potential
for broadband SG communication services has up to the present day been limited.
Time-varying HV arcing and corona noise with noise power fluctuations in the
order of several tens of decibels and the practicalities and costs of coupling com-
munication signals in and out of these lines have been an issue. Further, there is a
fierce competition of fibre optical links. In some cases, these links might even be
spliced together with the ground conductor of the HV system [12].
● MV lines: These connect the distribution substations to transformers’ kiosks or
pole-mounted transformers. The voltage levels are of the order of a few kilo-
volts and they run over distances of the order of a few kilometres. Numerous
cross connections and several types of cables and overhead lines characterize
the MV lines leading to hindering the RF signal propagation as the attenuation
is high in addition to the impedance matching problems. It is to be noted that
fibres in MV networks are rarely included in the power cabling and the MV
substations are connected to the communication network mainly through DSL,
private pilot cables or cellular radio techniques. Meanwhile, communications
to substations may be realized by PLC.
Typical MV lines, overhead lines or underground cables, working at vol-
tages in the range from 10 to 30 kV, are connected to zone transformer sub-
stations. They are used to distribute electricity between cities and large
industrial consumers. For monitoring and control purpose, MV lines are
directly connected to intelligent electronic devices (IEDs) that require low data
rates, and thus, NB-PLC can be relevant and economic communication solu-
tion for this task. The IEDs such as capacitor banks, reclosers, sectionalizers,
ring main units (RMUs) and phasor measurement units (PMUs) are connected
to the MV lines forming an MV distribution network with fewer branches than
those of LV distribution network.
● LV lines: They have the same nature as the MV lines. Recently, more attention
is being paid for data communication over LV lines due to the growth in
telecommunication industry and the advances in digital devices that use either
low or high bandwidth.
LV lines, with voltages in the range of 110–400 V, are connected to the MV
lines via MV/LV transformer substations. A communication signal on an MV
line can pass through these transformers to the LV line with heavy attenuation
in the order of 55–75 dB. Consequently, a special coupling device (inductive,
capacitive) or a PLC repeater is frequently required to establish a high data rate
communication path.
The frequency bands used by the PLC systems have a direct impact on signal reach
and effective bandwidth, which define the data rate and then the applications of PLC to
the grid. Therefore, based on the frequency band, PLC technology for SG can be
classified into three classes: ultra-narrowband PLC (UNB-PLC), narrowband PLC
(NB-PLC) and broadband PLC (BB-PLC) [13,14].
● UNB-PLC systems use very narrow bandwidth for data transmission in fre-
quencies less than 3 kHz. These systems are usually used for data transmission
when the electrical signals cross zero to be not affected by the high amplitude
of the main signal and harmonics. Because of the very low frequency, these
systems are less affected by transmission losses and can reach long distances
without repeaters. However, the main drawback of this technology is the low
data rate (a few hundred bits per second, bps).
● NB-PLC systems work with medium data rate in frequencies from 3 to 500 kHz.
Thus, this class of technology can be categorized as both low-data rate and high-
data rate technologies. For the first, the technology is based on single-carrier
modulations conveying data rates of a few kbps, whereas the second is based on
multi-carrier modulations and send data rates of hundreds of kbps.
● BB-PLC includes a large variety of systems that work with high-data rates in
frequencies from 1 to 250 MHz.
Role of PLC in SGs: In the beginning, the use of PLC was for tele-
communication, tele-protection and tele-monitoring between electrical substations
through EHV and HV power lines at voltages such as 110, 220 and 400 kV. The
carrier frequency range was used for audio signals, protection and a pilot frequency.
Nowadays, applications of PLC in SGs also pertain to power distribution
systems and communication between the consumers and the utility. Current
examples of these applications are AMI systems, EV charging systems, tele-control
applications, smart cities and DG systems. For these applications of PLC in SGs,
both NB-PLC and BB-PLC technologies are used in addition to some other appli-
cations of grid topology connectivity, cable health monitoring and fault location.
NB-PLC and BB-PLC technologies provide real-time data to the devices connected
to the grid, enabling an easy understanding of the network, as well as an efficient
management of events and failures.
A typical example of using PLC system in SG can be envisioned as shown
in Figure 20.5. The SG encompasses conventional generation, transmission through
HV power lines, and MV and LV distribution through overhead lines and under-
ground cables as well as incorporating renewable energy sources at the distribution
level [15].
Regarding the HV power lines, the phase conductors are used for signal
transmission by applying a coupling scheme to define how the communication
equipment is connected to the power line. The PLC channel is characterized based
on power line characteristics, coupling devices and line trap unit (LTU) char-
acteristics, and coaxial linking cable characteristics. It consists of signal path
between the transmitter and coupling device on the transmission side (power line
conductor) and coupling device and receiver on the receiving side. LTU is used to
block the signal propagation in the unwanted direction. It is to be noted that the
PLC communication system requires additional equipment (not needed for the
Communications
Coupling equipment
device Large
commercial
Coupling Communications building
capacitor equipment
HV o Coupling Coaxial cable
verhe
ad po
Conventional LTU wer li device
nes
generation Coupling SCADA
capacitor MV/MV
substation
BAN
LTU
RESs
MV switchboard
cablerground
PLC modem
ine ad
r l he
s
Repeater
we er
po V ov
s
nde
M
MV/LV
MV u
Sectionalizer MV/LV transformer
HAN transformer kiosk
RMU
Data aggregator and

energy management Factory (SCADA)
system
Recloser
Figure 20.5 Vision of PLC in power system
other technologies) to protect communication devices and personnel from the HV

and to match impedances between the coaxial cable and the power line.
NB-PLC and BB-PLC solutions can be implemented for the management of
MV substations, EVs and EV charging points, RMUs and data aggregators, smart
homes and distributed energy resources. In this case, PLC can be used in both MV
and LV sections. The major driver for the deployment of PLC market is the
increase of SG installations. Nowadays, hundreds of millions of PLC devices are
deployed all over the world. The objective of achieving secure and reliable com-
munications with utility control centres is a challenge. The main problems for that
are caused by noise generated by electrical devices, which may affect the proper
performance of the communications. In AMI systems, especially, the electrical
noise can block the communication between the meters and the headend systems in
the worst case and produce datagram losses. Other factors that may or have slowed
down the deployment of the PLC technology are security issues related to the
regulations in different countries, interoperability of different PLC solutions and
network topology of each country; for example, a meshed electricity network,
neutral secondary substations connectivity and others could challenge the deploy-
ment of PLC in SG if not properly engineered [12].
The LV lines lead directly or over street cabinets to the end customers’ pre-
mises. It is to be noted that a considerable regional topology difference exits. For
example, in the USA a smaller secondary transformer on a utility pole might
service a single house or a small number of houses. In Europe, however, it is more
common that up to 100 households get served from a single secondary transformer
substation. Further, significant differences exist between building types. They may
be categorized as multi-flat buildings with riser, multi-flat buildings with common
meter room, single family houses and high-rise buildings. Their different electrical
wiring topologies influence signal attenuation as well as interference between
neighbouring PLC networks.
20.2.2 Optical communications

Technologies of optical communications are widely used for building the com-
munication backbone that interconnects the transformer substations with control
centres. They have the following advantages:
● Ability of data transmission over long distance (several kilometres) with a
bandwidth of tens of Gbps.
● Adequacy of use in HV environment because of the robustness against elec-
tromagnetic and radio interference.
● Optical power ground wire as a special type of optical cables combines the
functions of grounding and optical communications as well as allowing high
data rate transmitted over long distance.
● Expandability for an application of providing SG services directly.
● Use in access networks, known as fibre-to-the-home (FttH) through using the
passive optical network (PON) technologies, which do not require electrical
power switching equipment. They use optical splitters to separate and collect
optical signals.
● Ability of serving multiple premises in a point-to-multi-point fashion by using
PON technology.
● Enabling the use of the standard Ethernet communication protocol over an
optical network by applying the Ethernet PON technology which is one of the
PON technologies. Consequently, the interoperability with existing Internet
protocol (IP)-based network is facilitated.
20.2.3 Digital subscriber lines

DSL enables digital data transmission over telephone lines. As a technology, it
allows the electric utilities to interconnect residential consumers to control centres
without any additional cost of deploying their own communication infrastructure,
just paying a communication fee to the telecommunications operator. DSL has
different categories based on the range of data rate. For instance, asymmetric DSL
(ADSL) supports up to 8 Mbps in the downstream and up to 640 kbps in the
upstream, ADSL2þ with a maximum download speed of 24 Mbps and upload
speed of 1 Mbps, and very-high-bit-rate DSL (VDSL or VHDSL) supports faster
data transmission over copper wires up to 52 Mbps downstream and 16 Mbps
upstream, and on coaxial cable up to 85 Mbps downstream and upstream.
20.3 Wireless communications

20.3.1 Microwave radio systems
Generally, radio frequencies, specifically the microwaves, are the transmission
medium, generally described as waves with frequencies that range from approxi-
mately 500 MHz to 300 GHz or more. Microwave radios propagate signals through
Earth’s atmosphere between transmitters and receivers often located on top of
towers spaced about 90–180 km apart. Microwave radio systems have the advan-
tage of having the capacity to carry thousands of individual information channels
between two points without the need for physical facilities such as coaxial or
optical fibre cables. The main advantages of microwave radio systems are the
following: (i) because of the high operating frequencies, they can carry large
quantities of information; (ii) high frequencies mean short wavelengths, which
require relatively small antennas; (iii) minimum delay times are introduced; and
(iv) minimal crosstalk exists between voice channels.
Microwave radio system capacities range from less than 12 voice-band chan-
nels to more than 22,000 channels. Early microwave systems carried frequency-
division-multiplexed voice-band circuits and used conventional, non-coherent
frequency modulation techniques. Most recent developed microwave systems
carry pulse-code time-division-multiplexed voice-circuits and use more modern
digital modulation techniques, such as phase-shift keying or quadrature amplitude
modulation [16–19]. A block diagram of a microwave radio system, transmitter and
receiver is shown in Figure 20.6. A local oscillator is used with a mixer to convert
the intermediate frequency (IF) or the baseband into RF as an up-converter during
signal transmission and as down-converter by changing the RF signal to IF signal
during receiving process.
20.3.2 Cellular communication networks

Cellular systems (e.g. the fourth generation ‘4G’) are more public than the other
wireless communication technologies because of their larger coverage area.
These systems use technologies such as global system for mobile communica-
tions ‘GSM’, general packet radio services ‘GPRS’ and enhanced data rates
for GSM evolution ‘EDGE’. Utilities have used such technologies for data
communications in supervisory control and data acquisition (SCADA) and
advanced meter reading systems. New generations continuously arise aiming at
supporting higher data rates, more data communication systems, reducing the
IF signal RF signal
S1 D
Mixer
Baseband Power
MUX IF MUX
in S2 amplifier
Up converter
C ENB
Bandpass
filter
LO
Down converter
Baseband Frequency Power
IF Demux
out divider amplifier
IF signal RF signal
Figure 20.6 Simplified block diagram of a microwave radio system: transmitter

and receiver
expenses and improving the latency performance that depends on the number of users
served by the same base station.
20.3.3 Satellite communication systems

Satellite systems have the capability of supporting communications with variable
bandwidth and latency performance by stationing satellites on orbits at different
altitudes. Electric utilities can consider such satellite communication systems for
SCADA systems as well as services supplied to rural or remote areas that may be
out of the coverage area of the present communication network or the high
expenses required to reach these areas with fibre optic cables.
20.4 Internet access

Internet architecture presents several benefits such as connecting large number of
devices, transmitting data over multiple media, connecting multiple types of sys-
tems and maintaining reliability. That is because IP has the capability of
● offering straightforward addressing and routing for large networks;
● running over any link layer network (encompassing ethernet, wireless
networks, etc.);
● identifying any type of system, when the data is sent from a point of access and
delivered to multiple destinations; and
● having more tools and applications to help manage the network.
Therefore, similarities between Internet facilities and requirements of SG
communication structure can be recognized as the SG communications network
needs interconnection between numerous devices, integration of highly hetero-
geneous areas (or individual electric subsystems) and supporting different degrees
of autonomous control at various times and scales. Thus, Internet technologies may
be an adequate solution for the problem of SG communication interoperability.
However, that is accompanied with the huge challenges of inherent threats and
vulnerabilities of the Internet and must be addressed [20].
20.5 Concluding remarks
Based on the anticipated SG infrastructure, a few concluding remarks are given as

follows:
● SG has the data communication network integrated into the electrical grid.
● In an SG, data about transmission, distribution and consumption are collected
and analysed in near real time.
● In addition to the integration of communication network into the power grid,
security of data is of utmost importance.
● Communications network in an SG includes different communication subnet-
works such as HAN, NAN, BAN and WAN, or SCADA. Each is dedicated to a
specific area, which has its own characteristics and requirements.
● Communications technology must be properly selected to be relevant, techni-
cal and economical, for each area or subnetwork.
● Based on the structure of the electric power system, the following steps should
be considered when designing the infrastructure of the communications
network:
– Specifying the main devices of monitoring and control for generation
sources (conventional, DGs), transmission and distribution systems.
– Identifying the functions and location of control centres at each system.
– Specifying the connections through the power system, such as connections
at transformer substations, switchgear, distribution transformers, RMUs,
PMUs and so on.
– Determination of system parameters that may affect the signal propaga-
tion, especially when using wired communication systems.
– Specifying the functions and locations of SMs required through the power
system and at consumers’ premises.
– Decomposition of the power system, from communications perspective,
into subsystems based on which subsystem is relevant, e.g. HAN or NAN,
etc. Each subsystem covers a specific area [21].
– Selection of communications technology that is well fitted to each sub-
system based on requirements such as the bandwidth, speed of data
transmission and data rate.
– Integration of different communication technologies through, for example,
specific protocols as each area in the system may have a technology that
differs from others.
– As the function of the communication system is data transmission,
security is of utmost importance.
References
[1] Yousuf M. and Vajuhudeen M. ‘Current communication media in power
systems’. International Journal of Research in Electrical Engineering
(IJREE). 2017;4(2):145–7.
[2] Ancillotti E., Bruno R. and Conti M. ‘The role of communication systems in
smart grids: Architectures, technical solutions and research challenges’.
Journal of Computer Communications, Elsevier. 2013;36:1665–97.
[3] Yan Y., Qian Y., Sharif H. and Tipper D. ‘A survey on smart grid commu-
nication infrastructure: Motivations, requirements and challenges’. IEEE
Communication Surveys & Tutorials. 2013;15(1):5–10.
[4] Frenzel Jr L. E. Electronics explained: Fundamental for engineers, techni-
cians, and makers. 2nd edn. Oxford, UK: Elsevier Inc, Newnes; 2018,
pp. 217–42.
[5] Ho Q. D. and Le-Ngoc T. ‘Smart grid communications networks: Wireless
technologies, protocols, issues, and standards’ in Obaidat M. S., Anpalagan A.
and Woungang I. (eds.). Handbook of Green Information and Communication
Systems. Oxford, UK: Elsevier Inc, Academic Press; 2013, pp. 115–46.
[6] Beal V. HAN-home area network [online]. Available from https://www.
webopedia.com/TERM/H/HAN.html [Accessed 4 Mar 2020].
[7] Sechilariu M., Wang B. and Locment F. ‘Building-integrated microgrid:
Advanced local energy management for forthcoming smart power grid com-
munication’. Journal of Energy and Buildings, Elsevier. 2013;59:236–43.
[8] Wang W., Xu Y. and Khanna M. ‘A survey on the communication archi-
tectures in smart grid’. Journal of Computer Networks, Elsevier. 2011;55:
3604–29.
[9] Ruiz-Romero S., Colmenar-Santos A., Mur-Perez F. and Lopez-Rey A.
‘Integration of distributed generation in the power distribution network: The
need for smart grid control systems, communication and equipment for a
smart city – Use cases’. Journal of Renewable and Sustainable Energy
Reviews, Elsevier. 2014;38:223–34.
[10] Berganza I., Bumiller G., Dabak A., Lehnert R. and Sendin A. ‘PLC for
smart grid’ in Lampe L., Tonello A. M. and Swart T. G. (eds.). Power Line
Communications: Principles, Standards and Applications from Multimedia
to Smart Grid. 2nd edn. New York: John Wiley & Sons; 2016, pp. 509–61.
[11] Oborkhale L. I. and Shoewu O. ‘Power line communication technology’. The
Pacific Journal of Science and Technology. 2007;8(2):398–405.
[12] Zajc M., Suljanović N., Mujčić A. and Tasič J. F. ‘High voltage power line
constraints for high-speed communications’. Proceedings of the 12th IEEE
Mediterranean Electrotechnical Conference; Dubrovnik, Croatia, May
2004, pp. 285–8.
[13] Galli S., Scaglione A. and Wang Z. ‘For the grid and through the grid: The
role of power line communications in the smart grid’. Proceedings of IEEE.
2011;99(6):998–1027.
[14] Berger L. T., Schwager A. and Escudero-Garzás J. J. ‘Power line commu-

nications for smart grid applications’. Journal of Electrical and Computer
Engineering. Special Issue; Power-line communications: Smart grid,
Transmission, and Propagation. 2013;(2013):1–16.
[15] Kabalci E., Kabalci Y. and Develi I. ‘Modelling and analysis of a power line
communication system with QPSK modem for renewable smart grids’.
Journal of Electrical Power and Energy Systems, Elsevier. 2012;34:19–28.
[16] Abdalla J. H., Carvalho P., Luis F., Evangelista, C. and Bermundez L. ‘Tools
for microwave radio communications system design’. Journal of Microwaves
and Optoelectronics. 2003;3(1):26–38.
[17] Office of the National Coordinator for Smart Grid Interoperability. NIST
framework and roadmap for smart grid interoperability standards. Release
1.0. U.S. Department of Commerce, Report, 2010, pp. 1–145.
[18] U.S. Department of Energy. Smart grid system report. Report, 2009, pp. 1–84.
[19] NETL. West Virginia smart grid implementation plan (WV SGIP) project,
Advanced Power and Electricity Research Center (APERC). Assessment of
as-is grid by non-utility stakeholders. Report, 2009, pp. 1–28.
[20] Gao J., Xiao Y., Liu J., Liang W. and Philip Chen C. L. ‘A survey of
communication/networking in smart Grids’. Future Generation Computer
Systems, Elsevier. 2012;28:391–404.
[21] Kuzlu M., Pipattanasomporn M. and Rahman S. ‘Communication network
requirements for major smart grid applications in HAN, NAN and WAN’.
Journal of Computer Networks. 2014;67:74–88.
Chapter 21
Cybersecurity in power systems
21.1 Introduction
The future smarter grid as envisaged in Chapters 1 and 20 illustrates that the grid of
the future will be subject to energy diversification, digitalization and advanced
metering infrastructure (AMI) systems that will include numerous new devices to
monitor, control and protect the grid. In addition, all grid components will be able
to communicate and collaborate through a high-speed two-way communication
system integrated into the power grid. This is because of the presence of conven-
tional, and renewable and non-renewable distributed energy sources as well as their
interdependency on information and communication technology (ICT) to enhance
grid efficiency and performance.
Power system is a massive and complex system. Being highly dependent on
ICT, with communications and networking systems, exposes the grid to potential
vulnerabilities to physical attack or cyberattack [1]. Thus, with the need to mitigate
the risk of compromising reliability and security of power system operation, and
surmounting the probable severe consequences from customer information leakage
to cascade failures that may lead to complete blackout and eradication of the
infrastructure, cybersecurity is becoming an issue of utmost importance in the
design of the information network. Therefore, cybersecurity issues are highlighted
in this chapter with the aim of getting the power engineers and power system
operators acquainted with the issues within the knowledge domain of cybersecurity.
Cybersecurity refers to the technologies, processes and practices designed to
protect networks, devices, programs, and data from attack, damage or unauthorized
access. The source of cyberattacks in electric power systems is mainly malicious
threats in the integrated communication networks [2]. Consequently, attack coun-
termeasures should be integrated into network protocols to achieve reliable infor-
mation exchange and real-time message delivery. Relevant transmission control
protocol/Internet protocol (TCP/IP) is described in the next section.
21.2 Communication network protocols in power grids

The structure of network protocol suites is mostly a series of layers, sometimes also
referred to as a protocol stack. Each layer is designed for a specific purpose and
exists in both the sending and the receiving systems. A specific layer on one system
sends or receives the same object that another system’s peer process sends or
receives. Layer activities occur independently from activities in layers above or
below the layer under consideration or other layers in the same system. Meanwhile,
each layer acts in parallel with the same layer on the other systems [3].
The basic unit of information is formed as a packet that is transferred across a
network. The basic packet consists of a header with the addresses of the sending
and receiving systems, and a body, or payload, with the data to be transferred. As
the packet travels through the TCP/IP protocol stack, the protocols at each layer
either add or remove fields from the basic header. When a protocol on the sending
system adds data to the packet header, the process is called ‘data encapsulation’.
Moreover, each layer has a different term for the altered packet.
The life cycle of a packet, which starts when a send message command is
issued, is summarized in this section. The life cycle finishes when processing
through all layers is completed and the appropriate application on the receiving
system receives the packet. Protocol layers of communication network in power
grids are shown in Figure 21.1 with functions described as follows:
● Application layer: It defines standard Internet services and network applica-
tions that anyone can use. These services work with the transport layer to send
and receive data. Many application layer protocols exist such as TCP/IP,
domain name system (DNS), simple network management protocol (SNMP)
and so on.
● Transport layer: The protocols at this layer start the process of data encapsu-
lation when the data arrive. The transport layer encapsulates the application
data into transport protocol data units. It ensures that packets arrive in
sequence and without error, by swapping acknowledgements of data reception
and retransmitting lost packets. This type of communication is known as end to
end. TCP is one of the protocols used at this level.
Network media
Physical layer Physical layer

Data link–MAC layer Data link–MAC layer
Frame
Network layer Network layer
Frame Transport layer Transport layer
IP datagram Application layer Application layer
TCP segment
Packet
Sending host Receiving host
Figure 21.1 Typical TCP/IP model packet travel through protocol layers of
communication network in a grid
Cybersecurity in power systems 531
● Network layer (IP layer): It accepts and delivers packets for the network. This
layer includes the powerful IP, the address resolution protocol (ARP) and the
Internet control message protocol (ICMP).
IP prepares them for delivery by formatting them into units called IP data-
grams. IP then determines the IP addresses for the datagrams, so that they can
be delivered effectively to the receiving host.
● Data-link layer: It is also called ‘medium access control (MAC) layer’. It identifies
the network protocol type of the packet (e.g. TCP/IP) and provides error control
and formatting the IP datagram into a frame in accordance with protocols such as
IEEE 802.2 framing and the point-to-point protocol (PPP). These protocols attach a
third header and a footer to ‘frame’ the datagram. The frame header includes a
cyclic redundancy check field that checks for errors as the frame travels over the
network media. Then, the data-link layer passes the frame to the physical layer.
● Physical network layer: It specifies the physical characteristics of the hardware
used in the communication network media. For instance, the physical network
layer on the sending host receives the frames and converts the IP addresses into
the hardware addresses appropriate to the network media. The physical net-
work layer then sends the frame out over the network media.
21.3 Cybersecurity objectives and requirements
As reported in [4], the three major objectives of cybersecurity for the grid are
availability, integrity and confidentiality of data. Availability is the primary
objective to ensure timely and reliable access to the information. Time latency of
the availability is application dependent, for example, data availability for protec-
tive relaying is less than 4 ms, for substation supervisory control and data acqui-
sition (SCADA) is in seconds, for market pricing in minutes and for meter reading
and longer-term pricing information in hours.
Data integrity implies source and quality of data. Modification or destruction
of original data leads to loss of data integration, which may also occur due to the
intrusion in the cyber domain by the attackers, unauthorized employees or human
error. Loss of data integration degrades the system reliability rapidly, especially for
a complex system such as the modern power grid.
Data confidentiality mainly pertains to protecting personal privacy and pro-
prietary information. It becomes especially important in systems involving inter-
action with customers (e.g. demand response and AMI networks).
To satisfy the objectives of cybersecurity, various issues, such as (i) attack
detection through monitoring the communication network traffic status to identify and
detect abnormal events due to attacks, if any, (ii) identification authentication and
access control to verify the identity of a device or user to be allowed to access the grid
resources as well as ensuring that the resources are accessed only by the appropriate
person, (iii) secure and efficient communication protocols to satisfy time-criticality
and security of message delivery in the grid, etc., should be addressed.
Referring to the National Institute of Science and Technology (NIST), the grid
infrastructure model is based on different segments [5]: centralized generation,
power transmission, energy distribution, distributed renewable and non-renewable
energy sources, HV/MV and MV/LV substations, system operations and control,
market, service providers, and the industrial, commercial and residential classes of
customers. Each class has its own characteristics, needs and corresponding com-
munication network such as HAN, BAN and IAN (Figure 21.2).
These area networks are interconnected through communication infrastructures
for forwarding any information such as alarms, commands, readings of voltage
values and energy consumption to system operations and control by means of IP-
based protocols (e.g. IEC 61968, IEC 61850, IEEE 1815 and ZigBee) as well as
controlling the message routing through use of gateways and routers. Gateways are
used for telecommunication networks to allow data to flow from one discrete net-
work to another irrespective of their protocol. They have the capability of con-
verting one protocol to another protocol. Consequently, they are enabling the traffic
flow and communication between dissimilar networks. The router is a device that
provides the function of routing IP data packets between similar networks. A packet
is typically forwarded from one router to another router through similar networks
until it reaches its destination node [6]. Therefore, a gateway can be viewed as a
router, but a router does not have to be a gateway.
21.4 Network attacks in power grids

Security attacks in communication networks come from either misbehaving users
or malicious operators. Misbehaving users are those who violate communication
System operations BAN

Service provider Market
IAN
Information flow
Router HAN
Secure communication line
Router
Gateway
Transmission substation HV/MV Renewable sources Substation MV/LV

Centralized
generation
Figure 21.2 Cybersecurity of smarter grid infrastructure

protocols with the aim of obtaining more network resources than those allocated for
the legitimate users. In contrast, malicious operators are dedicated to illegally
acquiring, modifying or disrupting information in the network. However, irre-
spective of the users’ type, security threats must be met. Countermeasures against
attacks by misbehaving users can be incorporated in the numerous electronic
computing devices used for monitoring and control purposes. Malicious operators’
attacks may result in severe damage to power supplies and probably widespread
power outage. Based on specific objectives, malicious attacks can be categorized
and described as in the following subsections.
21.4.1 DoS attacks

Denial-of-service (DoS) attacks encompass a series of actions that prevent the
functioning of any part of an information system concerning a specific device [7].
DoS can simultaneously attack an array of devices, and in this case, it is named as
‘distributed denial-of-service (DDoS)’. Thus, these attacks attempt to delay, block
or corrupt the communication resource in the power grids such as bandwidth; that
is, they target the availability objective of the network through effects that can
range from delay in service response time to complete inaccessibility.
DDoS is a distributed attack over a group of dispersed compromised devices in
the grid (called ‘bots’ or ‘slaves’) controlled by an attacker (called ‘bot-master’)
through command and control (C&C) channels and simultaneously used to attack a
target victim. The impact of the individual attack of compromised bots can be
aggregated to magnify the effect against the victim through communication net-
work infrastructure. The network of compromised bots is called a ‘botnet’. A server/
client model exposed to a DDoS attack is shown in Figure 21.3. The botnet includes
an array of clients that represents bots. The bot-master compromises the bots by
infecting them, e.g. with a Trojan or a back-door program in addition to controlling
them through C&C channels. The compromised bots are used to attack the server,
which is the target victim [8].
In power networks, DDoS attacks against communication system can happen
in a variety of communication layers as described in the following:
● Physical layer: One of the most efficient ways to launch physical layer attacks
is channel jamming as it is a very easy method for attackers, especially for
wireless communications. It becomes the primary physical layer attack in such
networks and can lead to a wide range of damage to the network performance
of, say, power substation systems, because of the delayed delivery of time-
critical messages as a result of jamming.
● MAC layer: Spoofing is a harmful threat at the MAC layer. The spoofing
attacker can open the address fields in the MAC frame and masquerade itself as
another device to send fake information to other devices. As an application of
this attack to a power substation network, a malicious node can broadcast
forged ARP packets to shutdown connections of all intelligent electronic
devices (IEDs) to the substation gateway node.
Co two
ne
m rk
m in
Botnet
un S
ic G
at
io
n
els
ann Bot#1
ch Server
C
&
C
Bot#2
Bot-master
ffic
k tra
ac
Bot#N Att Victim
Clients
Figure 21.3 A typical server/client model with DDoS attack
● Network and transport layers: These two layers need to provide reliable con-
trol for information delivery over communication networks according to TCP/
IP protocol model. Thus, DDoS attacks, such as traffic flooding and buffer
flooding, can highly degrade the end-to-end communication performance of
the power systems.
● Application layer: Attacks at this layer are mainly on transmission bandwidth
in communication channels, computers or routers. Thus, application layer
attacks can easily overflow the computers with limited computing resources by
computation flooding. Accordingly, the computing and communication devi-
ces in the grid can be potential victims of this attack as they are equipped with
limited abilities.
21.4.2 Attacks targeting integrity and confidentiality

These attacks occur at the application layer as they attempt to manipulate data
information in the power system. Attacks targeting integrity attempt to modify data
to corrupt information exchange for either consumers, for example, pricing infor-
mation and account balance, or status values of power systems such as voltage
readings and device running status. Therefore, integrity-check methods should be
applied by utilities to protect data integrity against any false data injection from
attackers.
Attacks targeting confidentiality have least impact on the grid reliability, but
because of deployment of AMI and decentralized control within the grid, an
increased attention is paid to protect the grid against such attacks.
21.5 Grid privacy

As reported by cybersecurity working group [9,10], privacy is not the same as
confidentiality and personal information is not the same as confidential informa-
tion. This is because access to confidential information is limited to only those with
a business need to know and consequently may result in compromise to a system,
data, application or other business function if inappropriately shared. Privacy can
simply be the right to be left alone. Privacy in the power system can be described
by four dimensions:
1. Privacy of personal information: It is any information relating to an individual,
who can be identified, directly or indirectly, by that information. In addition, it
involves the right to control when, where, how, to whom and to what extent an
individual can share their own personal information.
2. Privacy of the person: It is the right to control the integrity of one’s own body.
It covers things such as physical requirements, health problems and required
medical devices.
3. Privacy of personal behaviour: It is the right of the individuals to keep any
knowledge of their activities from being shared with others.
4. Privacy of personal communications: It is the right to communicate without
undue monitoring or censorship.
Smart meter (SM) reading at consumers’ premises is conveyed through the
communication network. Information can be exchanged between utilities and
consumers. More interoperability between consumers and service providers is
required for intelligent control and energy consumption management. Unprotected
energy-related data will allow invasions of privacy, for example, radio waves in
AMI may reveal information about where people were and when, and what they
were doing. Hence, such invasions are not accepted by regulators and consumers.
Privacy issues may include the following [11–13]:
● Personal information such as name, address, account number, meter reading,
home electrical appliances, life style, meter IP and service provider. Personal
information should be secured and monitored in a proper way.
● Privacy concerns with secure mechanisms to prevent privacy attacks on local
data. Attacks may present the following:
– fraud of energy consumption, which can be surmounted by not allowing
any personnel abuse or modifying the collected data;
– data in the SM could help deduce certain activities of home appli-
ances, which should not be recognized and tracked by malicious
intruder; and
– near-real-time data of energy consumption can reveal some privacy
of the individual power usage, which is not desired by the public.
Thus, data transfer from home meters to the data centre should be
secured.
21.6 Security of grid components

Power system is a large and complex arrangement, and it encompasses many dif-
ferent interlinked components. Security risks of some major components are
highlighted in this section to briefly illustrate the impact of cyberattacks on such
components [14,15]. The information flows between the different nodes in the grid,
and the system operations and control centre through the communication network is
shown in Figure 21.4. Each node in the network represents the location of IEDs
such as SMs, sensors or remote terminal units to monitor their own system states.
21.6.1 Process control systems

Process control systems (PCSs) or operations and control centres are the compo-
nents responsible for monitoring and control of physical properties of the electrical
power system. In power networks, these PCSs monitor and control large geo-
graphical areas of the power grid as they are traditionally designed to work for a
specific system. Consequently, many entry points exist to get into the grid. SCADA
system is one of the many types of PCSs being commonly used in the electrical
power system.
Security of PCSs is a paramount important factor to keep PCSs function
properly, especially as they control the physical aspects of the power grid. For
instance, compromising a computer loaded with data may cause some hardware
damage. Similarly, when a PCS is compromised, a number of equipment can be
physically damaged in addition to data being lost. The most important security
objective is the availability. So, the PCS must always be available to satisfy power
system availability. Lack of ability to deliver correct decisions based on false input
System operations
and control Node#IAN
State estimator
Node#BAN
Information flow
Router Router Node#HAN
Secure communication line
Gateway
Node#HV/MV Node#renewable Node#MV/LV

Node#generation Node#MV/LV
sources
Figure 21.4 Typical information flow through communication network

in power systems
data can in turn affect the integrity of decisions, the second security objective.
Regarding confidentiality, the third security objective, it may be too time-
consuming to meet latency as the PCS needs to run in real time. Accordingly,
PCSs must address security concerns of large networks [16].
21.6.2 Power system state estimation

Power system state estimation (SE) model is a tool that the grid PCSs use to model
sensor and agent data. It is important because it is used by the operators to maintain
the electric power system. As the security objectives important to PCSs are also
important for SE, it means that the security of SE model is also important.
Availability is the most important followed by the objectives of integrity and
confidentiality that adds overhead to a real-time system.
Maintaining the security of power system SE model is a challenge because of
the possibility of receiving false input data into the model. System instability and
financial gain are a couple of motivations for attackers. PCSs may have the false
data injection security issue, and distinguishing between actual and false data is a
difficult problem. There are mechanisms to distinguish bad data from normal data,
but this is not enough as they must also be effective against false data attacks.
21.6.3 Smart meters

SMs are devices installed at a customer’s site to measure the amount of power used
and send the information back to the power utility at regular intervals. Developed
electronic version of the current power meters are recommended not just for
recording the amount of energy consumption but are also used to give the utility a
feedback mechanism that can be used to model power usage requirements at a
much more detailed level for energy management.
The security of SMs is important because altered readings from the device,
especially if the device installed at a customer’s site is vulnerable to attack, can
lead to incorrect billing and false power usage approximations, as well as providing
attackers with monetary gains. It is important that, to satisfy the security objective
of integrity, the SM readings are correct and not modified. Confidentiality of the
SM reading is also important. Tools that can profile a user’s electrical usage
readings to determine which household appliances are being used are already
available. This information can be used by different companies and individuals, and
it is a privacy concern. The availability of SMs is more flexible than other grid
components. The maximum latency of SMs is in terms of hours, far greater than the
real-time latency requirements that are in terms of milliseconds.
SM security is a challenge because it is easy to gain physical access to the SM
device and there are instant monetary gains from altering these systems. The
integrity of SMs and their data need to be verified before use. Confidentiality of the
SM readings is also a challenge. SMs need to be networked to the power suppliers,
so that they can carry out their function. This means that it may be possible for
anyone else connected to that same network to observe others’ power usage read-
ings. This information can then be used to profile a user’s behaviour.
21.6.4 Communication protocol

The network comprises various intercommunication components to provide proper
functions. Each component has its own communication requirements and a set of
security needs. Thus, the network will need several communication protocols to
meet the varying connection requirements. The security of communication proto-
cols is important as the network communication is the backbone of the modern
grid. Many of the major grid functions cannot take place without communication.
The security objectives that are important depend on which components are com-
municating, and what data they are exchanging.
The network communication protocol security is a challenge because of the
presence of many different components that may communicate with different
communication protocols. Another issue is that the new grid technology needs to
integrate with legacy power systems, and many of these devices have constraints
that must be considered. Legacy devices can typically introduce security vulner-
abilities into the system because of a lack of security support.
21.6.5 False data injection attacks

During power system operation, SE is important for power system planning, ana-
lysis and studies such as optimal power flow calculations, contingency analysis and
automatic generation control. SE plays a vital role for smooth operation of different
energy management system applications. Basically, these studies take the SE output
data as an input to make an intelligent decision. For instance, SE receives data from
the SCADA network for processing purposes. Due to a false data injection attack,
SCADA may send the SE wrong information and make the smart grid vulnerable.
21.7 Cloud computing

Integration of distributed generation (DG) sources, especially renewable sources
such as solar and wind, plays a big role in meeting energy demand. To intensify the
visibility of DG within power markets, the idea of virtual power plant (VPP) has
emerged. VPP combines various small DG units to form a single virtual generating
unit that can act as a conventional one and is capable of being manageable on
individual bases [17]. VPP is outlined in Chapter 18, Section 18.3.1.
Based on the features that should be verified in the grid, huge amount of
information must be collected, stored and processed, and information exchange
between millions of devices is needed. The computing resources can be hardware
such as networks, servers and storage, as well as software such as applications and
services. Therefore, a model based on providing the required computation as a
service, whereby storage, software and information are provided to the computers
and other devices as a commodity over the Internet, called cloud computing, has
been developed. A single cloud computing data centre might have storage and
computing capabilities tens or hundreds of times greater than the multiple com-
bined supercomputing facilities, which will be more expensive.
Cloud providers own a large number of networked servers with low expenses.
This infrastructure includes massive pooled systems linked together and works with
virtualization techniques to provide high performance of data storage units and runs
alongside a local network connection that can run from a few to trillions of com-
putations per second depending on the demand.
As reported by NIST in [18], the term ‘cloud computing’ is defined as ‘a model
for enabling ubiquitous, convenient, on-demand network access to a shared pool of
configurable computing resources (e.g. networks, servers, storage, applications and
services) that can be rapidly provisioned and released with minimal management
effort or service provider interaction’.
This cloud model is composed of (i) five essential characteristics – on-demand
self-service, broad network access, resource pooling, rapid elasticity and measured
service; (ii) three service models – software as a service (SaaS), platform as a
service (PaaS) and infrastructure as a service (IaaS); and (iii) four deployment
models – private cloud, community cloud, public cloud and hybrid cloud as shown
in Figure 21.5 [19].
Cloud-based systems are sharing of an enormous amount of information
technology infrastructures, such as computational and database resources in the
form of service, which focus on maximization the efficiency of operation, scal-
ability, maintainability and reliability by decreasing cost.
NIST identifies seven domains within the power grid: generation, transmis-
sion, distribution, operations, customers, markets and service providers as shown in
Figure 21.6 [5]. Accordingly, power grids should support both electrical power
Hybrid cloud
Deployment
models
Community cloud Public cloud Private cloud
Service SaaS PaaS IaaS

models
Broad network On-demand Resource Elasticity Measured

service self-service pooling service service service
Characteristics
Figure 21.5 Cloud computing model

Service
Operations
Markets provider
Bulk
generation
Customer
Transmission Distribution
Secure communication interface Electrical interface

Domain
Figure 21.6 NIST – power grid model [5]
flows (generation, transmission and distribution of energy) and information pro-

cessing flows (collecting, processing and distributing data).
Information flow in a power network is required to monitor and control the
flow of energy (like SCADA systems) as well as monitoring and controlling future
and new energy-based services.
Electric utilities are focused on (i) improving the efficiency and reliability of
the delivery systems, (ii) getting solutions to empower the end consumer and (iii) to
better manage the demand and supply balance along the distribution chain, besides
the application of monitoring and diagnostic technologies to optimize assets use
and reduce the risk of catastrophic failure of electrical infrastructure.
21.7.1 Advantages, disadvantages and threats

of cloud computing
Cloud computing has strengths and weaknesses that must be understood to recog-
nize its applicability and how to develop this approach.
Advantages of cloud computing:

● Customers’ data are stored externally at the location of the cloud computing
provider.
● Clouds manage themselves in case of failures or the performance degradation;
that is, they introduce high reliability services.
● Customers pay for the service rather than buying software and hardware.
● Cloud services can be used instantly and can be scaled up and down easily.
● Sharing of computing resources between different customers.
Disadvantages of cloud computing:
● Dependency on Internet connectivity as constant connection is required.
● Pay per use is varied as the cost of computing differs with the time.
● Security concerns.
Threats to cloud computing:
The top seven prime threats to cloud computing are listed as follows [20]:
● Abuse and immoral use of cloud computing.
● Insecure application programming.
● Interfaces.
● Malevolent insiders.
● Shared technology liabilities.
● Data losses or leakages.
● Account, service and traffic hijacking.
References
[1] Arghandeh R., Meier A., Mehrmanesh L. and Mili L. ‘On the definition of
cyber-physical resilience in power systems’. Journal of Renewable and
Sustainable Energy Reviews, Elsevier. 2016;58:1060–69.
[2] Pearson I. L. G. ‘Smart grid cyber security for Europe’. Journal of Energy
Policy, Elsevier. 2011;39:5211–18.
[3] Amin S. M. and Giacomoni A. M. ‘Smart grid – safe, secure, self-healing’.
IEEE power & energy magazine. 2012. IEEE; pp. 33–40.
[4] Kuzlu M., Pipattanasomporn M. and Rahman S. ‘Communication network
requirements for major smart grid applications in HAN, NAN and WAN’.
Journal of Computer Networks. 2014;67:74–88.
[5] NIST. NIST framework and roadmap for smart grid interoperability stan-
dards, Release 2.0. NIST Special Publication 1108R2 report, 2012.
[6] Wang W. and Lu Z. ‘Cyber security in the smart grid: Survey and chal-
lenges’. Journal of Computer Networks, Elsevier. 2013;57:1344–71.
[7] US Committee on National Security Systems. National Information
Assurance (IA) Glossary. CNSS, Instruction 4009 report, 2006.
[8] Bhatia S., Behal S. and Ahmed I. ‘Distributed denial of service attacks and
defense mechanisms: Current landscape and future directions’ in Conti M.,
Somani G. and Poovendran R. (eds.). Versatile Cybersecurity. Advances in
Information Security. Springer, Cham Press; 2018, pp. 55–97.
[9] The Smart Grid Interoperability Panel, Cyber Security Working
Group. Guidelines for smart grid cyber security. Vol. 2, Privacy and the
smart grid. NIST and US Department of Commerce report, 2010.
[10] The Cyber Security Coordination Task Group. Smart grid cyber security
strategy and requirements. NIST report, 2009.
[11] Liu J., Xiao Y., Li S., Liang W. and Philip Chen C. L. ‘Cybersecurity and
privacy issues in smart grids’. IEEE Communications Surveys & Tutorials.
2012;14(4):981–97.
[12] Siddiqui F., Zeadally S., Alcaraz C. and Galvao S (eds.). ‘Smart grid priv-
acy: Issues and solutions’. 21st International Conference on Computer
Communications and Networks (ICCCN); Munich, Germany, Jul/Aug 2012.
IEEE; pp. 1–5.
[13] Lu R., Liang X., Li X., Lin X. and Chen X. ‘EPPA: An efficient and
privacy-preserving aggregation scheme for secure smart grid communica-
tions’. IEEE Transactions on Parallel and Distributed Systems. 2012;23(9):
1621–32.
[14] Anwar A. and Mahmood A. ‘Cyber security of smart grid infrastructure’ in
Anwar A. and Mahmood A. (eds.). The State-of-the-art Intrusion Prevention
and Detection. USA: CRC Press, Taylor & Francis Group; 2014, pp. 449–72.
[15] Kundur D., Feng X., Kiu S., Zourntos T. and Butter-Purry K. L. (eds.).
‘Towards a framework for cyber-attack impact analysis of the electric smart
grid’. First IEEE International Conference on Smart Grid Communications;
Gaithersburg, MD, USA, Oct 2010. IEEE; pp. 244–49.
[16] Mahapatra K. and Chaudhuri N. R. ‘Online robust PCA for malicious attack-
resilience in wide-area mode metering application’. IEEE Transactions on
Power Systems. 2019;34(4):2598–610.
[17] Markovic D. S., Zivkovic D., Branovic I., Popovic R. and Cvetkovic D.
‘Smart power grid and cloud computing’. Renewable and Sustainable
Energy Reviews, Elsevier. 2013;24:566–77.
[18] Mell P. and Grance T. The NIST definition of cloud computing. NIST Special
Publication 800-145 report, 2011.
[19] Bera S., Misra S. and Rodrigues J. P. C. ‘Cloud computation applications for
smart grid: A survey’. IEEE Transactions on Parallel and Distributed
Systems. 2016;26(5):1477–94.
[20] Sivapragash C., Thilaga S. R. and Kumar S. S. (eds.). ‘Advanced cloud
computing in smart power grid’. Third International Conference on
Sustainable Energy and Intelligent Systems (SEISCON 2012), VCTW;
Tiruchengode, Tamil Nadu, India, Dec 2012, pp. 349–54.
Appendix A
Probability density and Weibull distribution
functions
It is essential to obtain the probability density function (pdf) of the actual power
output of wind or solar photovoltaic energy systems, especially for planning and
analysing a power system incorporating such renewable sources of energy.
Probability density function is a statistical expression that defines a probability
distribution for a continuous random variable as opposed to a discrete random
variable.
A.1 Random variables
A random variable is a real-valued function that assigns a numerical value to each

possible outcome of the random experiment. For instance, the random variable X
assigns the value 0 to the outcome GenOutage, the value 1 to the outcome
GenInService and so on. Hence, the random variable X is a mapping function from
the sample space S ¼ fGenOutage; GenInService; ; GenRestoreg to the real
numbers. That is,
X :S!R
Random variables usually are written in capital letters such as X, Y and Z. As a
random variable is a function, it has a range. The range of a random variable X is
denoted by Range(X) or RX and is the set of possible values for X.
There are two important classes of random variables: discrete random vari-
ables and continuous random variables.
A.1.1 Discrete random variables

A random variable is discrete if its range is a countable set. The set A is countable if
either
● A is a finite set such as {1,2,3,4}, or
● it can be put in one-to-one correspondence with natural numbers (in this case,
the set is said to be countably infinite).
A.1.2 Continuous random variables and their distributions

One big difference that is noticed as opposed to discrete random variables is that
the distribution function of a random variable X is the function FX ðxÞ of x, which is
called ‘cumulative distribution function (cdf)’ and given by FX ðxÞ ¼ pðX xÞ for
all values x in the range of X. It is a continuous function; that is, it does not have
any jumps. That jumps in the cdf correspond to points x for which PðX ¼ xÞ > 0.
Thus, the fact that the cdf does not have jumps is consistent with the fact that
PðX ¼ xÞ ¼ 0 for all x, where probability cannot be assigned to individual values
of x; instead, probability is assigned to intervals. The continuous random variable is
defined as follows: ‘A random variable X with cdf, FX ðxÞ, is said to be continuous if
FX ðxÞ is a continuous and differentiable function for all x [ R’.
A.2 Probability density function
To determine the distribution of a discrete random variable, probability mass

function (pmf) or cdf should be provided. For continuous random variables, the cdf
is well defined, so it can be provided. However, the pmf does not work for con-
tinuous random variables. Instead, the pdf can usually be defined where it is the
density of probability rather than the probability mass. The pdf can be thought of as
the infinite limit of a discrete distribution, i.e. a discrete distribution with an infinite
number of possible outcomes. Specifically, suppose a discrete distribution with N
possible outcomes is created, each corresponding to a range on the real number
line. Then, suppose N is increased towards infinity, so that each outcome shrinks to
a single real number; a pdf can be considered as the limiting case of this
distribution.
The concept is very similar to mass density in physics: its unit is probability
per unit length. To get a feeling for pdf, consider a continuous random variable X
and define the function fX ðxÞ as follows (wherever the limit exists):
Pðx < X < x þ hÞ
fX ðxÞ ¼ limþ (A.1)
h!0 h
where
Pðx < X x þ hÞ ¼ FX ðx þ hÞ FX ðxÞ (A.2)
The function fX ðxÞ gives the probability density at point x. It is the limit of the
probability of the interval ðx; x þ h divided by the length of the interval as the
length of the interval goes to 0. Substituting (A.2) into (A.1) obtains
FX ðx þ hÞ FX ðxÞ dF ðxÞ
fX ðxÞ ¼ lim ¼
D!0 h dx
0
¼ FX ðxÞ if FX ðxÞis differentiable at x (A.3)
Probability density and Weibull distribution functions 545
Thus, the definition of the pdf of continuous random variables follows as

given next.
Considering a continuous random variable X with a continuous cdf, FX ðxÞ, the
pdf, fX ðxÞ; is given by
dF ðxÞ 0
f X ðx Þ ¼ ¼ FX ðxÞ if FX ðxÞis differentiable at x (A.4)
dx
As the pdf is the derivative of the cdf, the cdf can be obtained by integrating
(A.4) as follows:
ðx
F X ðx Þ ¼ fX ðuÞdu (A.5)
1
From (A.2), the following form can be written:

ðb
Pða < X bÞ ¼ FX ðbÞ FX ðaÞ ¼ fX ðuÞdu (A.6)
a
and integration over the entire real line must get 1, that is,
ð1
fX ðuÞdu ¼ 1 (A.7)
1
That is, the area under the pdf curve must be equal to one. However, con-
sidering a continuous random variable X with pdf, fX ðxÞ; the properties of the pdf
can be summarized as
g
fXððxÞ 0 for all xR
1
fX ðuÞdu ¼ 1
1
ðb
(A.8)
Pða X bÞ ¼ FX ðbÞ FX ðaÞ ¼ fX ðuÞdu
ðx a
F X ðx Þ ¼ fX ðuÞdu
1
According to (A.6), fX ðuÞ can be plotted for RX, and then, the probability that x
falls within the range can be obtained by computing the area under the curve, fX ðuÞ
versus u. It is important to note that a probability density is not a probability per se.
For one thing, there is no requirement that FX ðxÞ 1. Moreover, the probability
that x attains any one specific out of the infinite set of possible values is always
zero. For example,
ð5
FX ðx ¼ 5Þ ¼ fX ðxÞdx ¼ 0 for any pdf ; fX ðxÞ
5
Joint distributions are defined in a natural way. For two variables X and Y, the
joint pdf, fX ;Y ðx; yÞ, defines the probability that (x, y) lies in a given domain D:
ð
fX ;Y ðx; yDÞ ¼ fX ;Y ðx; yÞdxdy (A.9)
x;yD
For instance, the probability that a 2D-coordinate (x, y) lies in the domain
Ð2 Ð2
ð1 x 2; 1 y 2Þ is x¼1 y¼1 fX ;Y ðx; yÞdxdy. The pdf over a vector may also
be written as a joint pdf of its variables.
Conditional distributions are defined as well; fX ðxjAÞ is the pdf over x, if the
statement A is true. This statement may be an expression on a continuous value, say
y ¼ 5. As a shorthand, fX ðxjyÞÐ can be written to provide a pdf for x for every value
of y. It must be the case that fX ðxjyÞdx ¼ 1, as fX ðxjyÞ is a pdf over values of x.
A.3 The Weibull distribution function

The Weibull distribution is widely used for several applications such as mathe-
matical modelling for system analysis, failure analysis, warranty analysis, relia-
bility growth, life-cycle cost and production process control. Weibull models are
used to represent pdf and cdf [1, 2]. The non-negative random variable X with
distribution function
8 x k
<
F X ðx Þ ¼ 1 e c ; x > 0 (A.10)
:
0; x 0
is said to have a Weibull distribution with shape parameter k > 0 and scale para-
meter c > 0. The pdf of the Weibull distribution is
8 x k
<k
x e c ; x>0
k1
0
f X ðx Þ ¼ F X ðx Þ ¼ (A.11)
: ck
0; x0
It is to be noted that the exponential distribution is a special case of Weibull
distribution with k ¼ 1.
The expectation and variance of a Weibull distribution could be computed by
ð1
k k ðx=cÞk 1
E ðX Þ ¼ t e dx ¼ cG 1 þ (A.12)
0 c k
k
ð1
k kþ1 ðx=cÞk 2
E X2 ¼ t e dx ¼ c 2
G 1 þ (A.13)
0 c k
k
!
2 2 2 2 1 2
VarðX Þ ¼ E X ðEðX ÞÞ ¼ c G 1 þ G 1þ (A.14)
k k
Probability density and Weibull distribution functions 547
A.3.1 Calculation of mean and variance

● For discrete random variables, the formula of the mean, expected, value is
based on the probability function. It is calculated by
X
E ðX Þ ¼ xpX ðxÞ (A.15)
X
and the variance is defined by

n o
VarðX Þ ¼ E ½X EðX Þ2 (A.16)
● For continuous random variables with pdf, fX ðxÞ, the expectation of a function
g(x) is defined as
ð
1
Efg ðX Þg ¼ gðxÞfX ðxÞdx (A.17)

1
Hence, the mean is given by

ð
1
E ðX Þ ¼ xfX ðxÞdx (A.18)

1
The best way, in practice, to calculate the variance is by using the relations
given in (A.19) and (A.20):

VarðX Þ ¼ E X 2 fEðX Þg2 (A.19)
and
ð
1
2

E X ¼ x2 fX ðxÞdx (A.20)
1
Example A.1 The lifetime, measured in years, of a brand of mobile device has a
Weibull distribution with parameters k ¼ 2 and c ¼ 13. Compute the probability of
a mobile device failing before the expiration of a 2-year warranty.
Solution

FX ðx 2Þ ¼ 1 exp 2=132 0:0234:
In general, the computation program to provide the Weibull probability
distribution function versus X at k ¼ 2 and c ¼ 4, 7, 10 and 13 is as in the
following:
0.25
Weibull(2,4)
Weibull(2,7)
Weibull(2,10)
0.20 Weibull(2,13)
0.15
Fx(X)
0.10
0.05
0
0 2 4 6 8 10 12 14 16 18
X
Figure A.1 Weibull probability distribution function with c ¼ 2 and k ¼ 4, 7,

10 and 13
X ¼ 0:1 : 0:2 : 18
k ¼ 2; c ¼ 4; Y1 ¼ ðk=c ^ k Þ X: ^ ðk 1Þ: expððX=cÞ: ^ k Þ
plot(X,Y1,‘m-’,X,Y2,‘g-’,X,Y3,‘b-’,X,Y4,‘r-’) legend(‘Weibull(2,4)’,‘Weibull
(2,7)’,‘Weibull(2,10)’, ‘Weibull(2,13)’)

title ’Weibullðk; k Þ : f ðxÞ ¼ ðk=c ^ k Þx ^ fk 1gexp½ðx=cÞ ^ k 0
The results are plotted as shown in Figure A.1.
References
[1] Lai C. D., Murthy D. N. and Xei M. ‘Weibull distributions and their
applications’. Springer Handbook of Engineering Statistics, Part A. 2006,
pp. 63–78. Available at: http://www.researchgate.net/publication/37628953.
[2] Khan M. S., Pasha G. R. and Pasha A. H. ‘Theoretical analysis of inverse
Weibull distribution’. WSEAS Transactions on Mathematics. 2008;7(2):
30–38.
Appendix B
Test system
The modified IEEE 24-bus system without hydro generation is taken as a test
system. Its transmission network consists of 24-bus bars connected by 38 lines and
transformers and the generation system includes 26 thermal units at locations as
shown in the single-line-diagram (Figure B.1).
The characteristics of generating units are given in Tables B.1 and B.2. The
hourly peak load and bus loads as a percentage of the system load are listed in
Table B.3.
Bus 17 Bus 18
Bus 21 Bus 22
Bus 23
Bus 16
Bus 19 Bus 20
Bus 15
Bus 14 Bus 13
Bus 24 Bus 11 Bus 12
Bus 3 Bus 9 Bus 10

Bus 6
Bus 4
Bus 5 Bus 8
Bus 7
Bus 1 Bus 2
Figure B.1 Single-line diagram of the test system

Table B.1 Generating unit data
Unit Pi (MW) P i (MW) ai (k$/MW2) bi (k$/MW) ci (k$) Bus #

1 2.4 12.0 0.02533 25.5472 24.3891 15
2 2.4 12.0 0.02649 25.6753 24.4110 15
3 2.4 12.0 0.02801 25.8027 24.6382 15
4 2.4 12.0 0.02842 25.9318 24.7605 15
5 2.4 12.0 0.02855 26.0611 24.8882 15
6 4.0 20.0 0.01199 37.5510 117.755 1
7 4.0 20.0 0.01261 37.6637 118.108 1
8 4.0 20.0 0.01359 37.7770 118.457 2
9 4.0 20.0 0.01433 37.8896 118.820 2
10 15.2 76.0 0.00876 13.3272 81.1364 1
11 15.2 76.0 0.00895 13.3538 81.2980 1
12 15.2 76.0 0.00910 13.3805 81.4641 2
13 15.2 76.0 0.00932 13.4073 81.6259 2
14 25 100.0 0.00623 18.0000 217.895 7
15 25 100.0 0.00612 18.1000 218.335 7
16 25 100.0 0.00598 18.2000 218.775 7
17 54.25 155.0 0.00463 10.6940 142.734 15
18 54.25 155.0 0.00473 10.7154 143.028 16
19 54.25 155.0 0.00481 10.7367 143.317 23
20 54.25 155.0 0.00487 10.7583 143.597 23
21 68.95 197.0 0.00259 23.0000 259.131 13
22 68.95 197.0 0.00260 23.1000 259.649 13
23 68.95 197.0 0.00263 23.2000 260.176 13
24 140.0 350.0 0.00153 10.8616 177.057 23
25 100.0 400.0 0.00194 7.4921 310.002 18
26 100.0 400.0 0.00195 7.5031 311.910 21
Table B.2 Generating units operating parameters
Unit Min. Min. UH (h) DH (h) UR DR aj (h) bi (h) ti (h)

Up (h) Down (h) (MW/h) (MW/h)
1–5 0 0 0 0 48.0 60.0 0 0 1
6–9 0 0 1 0 30.5 70.0 20 20 2
10–13 3 2 2 1 38.5 81.0 50 50 3
14–16 4 2 2 2 51.0 74.0 70 70 4
17–20 5 3 3 2 55.0 78.0 150 150 6
21–23 5 4 4 2 55.0 99.0 200 200 8
24 8 5 5 3 70.0 120.0 300 200 8
25–26 8 5 8 4 50.5 100.0 500 500 10
Test system 551
Table B.3 Hourly load demand and bus load as a percentage of total load
Hour Load Hour Load Hour Load Bus Bus load Bus Bus load
(MW) (MW) (MW) # (%) # (%)
1 2,000 10 2,257 19 2,539 1 3.8 10 6.8
2 1,847 11 2,308.5 20 2,488 2 3.4 13 9.3
3 1,744 12 2,334 21 2,411 3 6.3 14 6.8
4 1,693 13 2,308.5 22 2,360 4 2.6 15 11.1
5 1,642 14 2,257 23 2,231.5 5 2.5 16 3.5
6 1,643 15 2,231.5 24 2,077.5 6 4.8 18 11.7
7 1,693 16 2,231.5 – – 7 4.4 19 6.4
8 1,795.5 17 2,334 – – 8 6.0 20 4.5
9 2,052 18 2,565 – – 9 6.1 – –
1.0
0.9
0.8
Load demand (p.u.)
0.7
S + 0.05
0.6
S
0.5
S – 0.05
0.4 Covered range of
training process
0.3
0.2
0.1
0 6 12 18 24
Time (h)
Figure B.2 Load demand over 24-hourly time periods
The patterns of load demand over 24-hourly time periods with a range of 5%
load forecast error are taken as samples for training the ANN (Figure B.2).
The data of the transmission network are reported by the reliability test system
task force of the application of probability methods subcommittee and published in
IEEE Transactions Power Apparatus and Systems, vol. PAS-98, no. 6, Nov./Dec.
1979.

FCi Pij ¼ aij P2ij þ bi Pij þ ci :
Pi, Pi ¼ rated lower and upper generation limit, respectively.

h h ii
SCi;j ¼ uij ai þ bi 1 exp Ti;joff =ti :
UH, DH ¼ ramp-up and ramp-down time, respectively.

UR, DR ¼ ramp-up rate limit and ramp-down rate limit, respectively.
B.1 Data of renewable energy sources
1. Each wind turbine of the wind farm of 30 MW rating has the following
characteristics:
● Cut-in speed ¼ 5 m/s
● Nominal speed ¼ 15 m/s
● Cut-off speed ¼ 45 m/s.
2. The solar PV plant of rating 30 MW consists of 400,000 modules. Each module
has the following characteristics:
● Watt peak – 50 W
● Open-circuit voltage ¼ 38 V
● Short-circuit current ¼ 1.8 A
● Voltage at maximum power ¼ 38 V
● Current at maximum power ¼ 1.32 A
● Voltage temperature coefficient ¼ 194 mV/ C
● Current temperature coefficient ¼ 1.4 mA/ C
● Nominal cell operating temperature ¼ 43 C.
Index
AC–AC conversion 330 asynchronous induction generators 41

ON–OFF control 330–1 doubly fed induction machines
phase control 331 (DFIM) 46
AC–DC conversion 313 dynamic model in d-q frame
diode rectifiers cascaded with boost 55–6
converter 317–19 power and torque relations 51–4
phase-controlled rectifiers 315–17 steady-state model 47–9
pulse-width modulation (PWM)- steady-state phasor diagram
controlled rectifiers 319–20 49–51
uncontrolled rectifiers 314–15 squirrel cage induction generators
activation losses 229–30 43
adiabatic compressed air energy equivalent circuit in transient
storage (A-CAES) 203 state 45–6
advanced metering infrastructure relations and equivalent electric
(AMI) 513, 521 circuit at steady state 43–5
aggregator control centre (ACC) wind turbine (WT) with doubly fed
466–9 induction generator (DFIG)
air density, factors affecting 25 control system 56–60
pressure impact 25–6 attenuators 146
temperature impact 26 automatic generation control (AGC)
airfoils 35–9 463–4, 473
air mass ratio (AM) 73 average system availability index
air turbines 149–51 (ASAI) 358
amorphous silicon 70 average system interruption
anolyte 251 duration frequency index
Aquamarine Power Oyster 147 (ASIDI) 359
area control error (ACE) 463 average system interruption frequency
artificial intelligence (AI) 411 index (ASIFI) 359
artificial neural network (ANN) Avogadro’s number 222
411–12 axle angular speed 300
evolutionary algorithms 412–13 axle torque 300
heuristic and expert systems 411
simulated annealing (SA) 411 back electromotive force (BEMF) 286,
Tabu search (TS) 412 293
artificial neural network (ANN) 402, band-gap energy 75
411–12, 416–17 base-load plants 185
battery charging infrastructure, burning municipal solid waste or

of electric vehicles 301 wood waste 176
chargers 304 burning wood 175–6
Level 1 and Level 2 chargers collecting landfill gas/biogas
304–6 176–7
Level 3 chargers 306 disposing ash from waste-to-
charging stations 303–4 energy plants 176
charging times 306–8 liquid biofuels 177
battery electric vehicles (BEVs) 277, biomass materials 172
282 black moon 102
battery energy storage (BES) blades–rotor–gearbox–generator set
technologies, comparison of 36
255 boost converter 322–4, 490
economical comparison 256–8 Brayton cycle engines 94
technical comparison 256 breakers 133
battery energy storage system (BESS) breaking waves 133
457, 486, 496 Bristol cylinder 152
local controller of 498 broadband PLC (BB-PLC) 520–1
battery management system (BMS) brushed DC motors 282–9
304 brushless DC (BLDC) motors 292
battery modelling 252–5 equivalent circuit and general
battery packs, large 280 equations 294–6
beta-alumina ceramic 250 operational theory 292–3
Betz limit 30 switching process 293–4
bi-directional tidal turbine generators buck-boost converter 324
116 buck converter 321–2
biochemical technology 175 building area network (BAN) 515
biodiesel 177 bulk storage 5
biodiesel fuel 347 burning municipal solid waste or wood
bioenergy 172 waste 176
biofuels 177 burning wood 175–6
biological materials 347
biological technology 174–5 calendar life 264
biomass 169 capacitance 255
properties 172–3 capacity factor (CF) 39–41
as a renewable source of energy 171 capillary wave 131–2
types of 169–71 carbon cycle from crops to a liquid
biomass energy 171–2 fuel 172–3
conversion technologies 173–5 catholyte 251
biochemical technology 175 cellular communication networks
biological technology 174–5 523–4
thermal technology 173–4 central receiver system 95
thermochemical technology 174 chargers 281
pros and cons of 175 of electric vehicles 304
Index 555
Level 1 and Level 2 chargers communication network protocols in

304–6 power grids 529–31
Level 3 chargers 306 communication protocol 538
charging, defined 183 communications in power systems 513
charging cycle (compression cycle) future power grid, perspective of
204–5 513
charging management system (CMS) classification of communication
461 networks 515–17
charging stations, of electric vehicles Internet access 524–5
303–4 wired communications
charging times, of electric vehicles digital subscriber lines (DSLs)
306–8 522–3
chemical energy storage systems 217 optical communications 522
factors affecting VOC 226 power line communication (PLC)
activation losses 229–30 system 517–22
concentration losses 230–1 wireless communications
ohmic losses 230 cellular communication
fuel cell: see fuel cell (FC) networks 523–4
polymer electrolyte membrane fuel microwave radio systems 523–4
cell (PEMFC) characteristics satellite communication systems
233 524
calculation of resistances communication systems 513–15
233–4 compressed air energy storage (CAES)
computer modelling 234–5 system 203
power-to-gas (PtG) energy system advantages and disadvantages of
217, 236 207–8
gas turbines 239–41 modelling 206–7
microturbines 241 principles of CAES operation 204–5
principle of operation 236–9 concentrating solar power (CSP) 65,
choke filter 331 97, 268
choppy waves 132 concentration losses 230–1
circuit breaker 305 concentration PV 69
circuit interrupting device 305 constant losses 289
Claude cycle 159 consumers’ role in the electricity
clearness index 374 market 8–9
closed-cycle OTEC 160–3 continuous random variables and their
cloud computing 538 distributions 544
advantages of 540–1 controller area network (CAN) bus
disadvantages of 541 functionality 304
threats to 541 conventional economic dispatch (CED)
cogeneration 346 with thermal units 379–80
combined heat and power (CHP) 346 conventional power systems 2–4
combustion turbines 345 conventional secondary BES
command and control (C&C) technologies
channels 533 lead-acid (PbA) batteries 243–4
lithium ion (Li-ion) batteries 245–6 denial-of-service (DoS) attacks

metal air (Me-air) batteries 246–7 533–4
nickel-based batteries 244–5 objectives and requirements 531–2
sodium nickel chloride (NaNiCl) cycle life 264
batteries 248–50 cycling capacity 190
sodium sulphur (NaS) batteries
247–9 data confidentiality 531
Copenhagen Accord 382 data encapsulation 530
copper losses 288 data integrity 531
Coulombic efficiency 263 DC–AC conversion 324
critical clearing time (CCT) 452, 454 single-phase inverters, operation of
crystalline silicon 68 327–9
cube of the wind speed, average value sinusoidal PWM 325–6
of 34 three-phase inverters (two-level),
cumulative distribution function (cdf) operation of 329–30
544–6 DC–DC conversion 320
current-source inverter (CSI) 324 boost converter 322–4
customer average interruption duration buck-boost converters 324
index (CAIDI) 358 buck converter 321–2
customer average interruption DC motors
frequency index (CAIFI) 358 brushed 282–9
Customers Experiencing Multiple brushless 292
Interruptions (CEMIn) 358 equivalent circuit and general
customer total average interruption equations 294–6
duration index (CTAIDI) 358 operational theory 292–3
cut-out wind speed 39 switching process 293–4
cybersecurity, in power systems 529 dynamic and regenerative braking
cloud computing 538 in 289–92
advantages of 540–1 decomposition methods 415–16
disadvantages of 541 deepwater wave 134–5
threats to 541 corresponding group velocity in 142
communication network protocols fundamentals of 137–9
529–31 degree of hybrid station (DOH) 279
grid components, security of 536 demand response (DR) 8
communication protocol 538 demand side management (DSM) 350
false data injection attacks 538 denial-of-service (DoS) attacks 533–4
power system state estimation depth of discharge (DoD) 190, 245,
(SE) model 537 253
process control systems (PCSs) Deterministic Dynamic Economic
536–7 Emission Dispatch (DDEED)
smart meters (SMs) 537 concept 392, 395
grid privacy 535 deterministic economic dispatch
network attacks in power grids 532 (DED) incorporating wind
attacks targeting integrity and power 380
confidentiality 534 diesel engine generator sets 346
Index 557
diesel fuel 347 load-based indices 359

digital subscriber lines (DSLs) 522–3 momentary indices 359
diode rectifiers cascaded with boost power system restoration 355–7
converter 317–19 sustained interruption indices
direct drive permanent magnet 357–8
(DDPM) generator 123 virtual inertia 359–62
discharge efficiency: see storage-tank voltage and frequency ride
effectiveness through functionality 353–5
discharging, defined 183 voltage regulation 352
discharging cycle (expansion cycle) distributed energy source (DES) 460
205 distributed generation (DG) systems
discrete random variables 543 4–5, 347–8, 373, 513
dish-engine systems 94–5 economic benefits 349
dish-Stirling system 94–5 of DERs integration in power
distributed denial-of-service (DDoS) systems 349–51
533–4 emissions reduction (ER) benefits
distributed energy resources (DERs) with integration of renewable
181–2, 345, 485–7 DG units 351–2
classification of 346 environmental benefits 348
distributed generation (DG) systems, non-renewable resources 347
powering 347 renewable resources 347–8
non-renewable resources 347 technical benefits 348–9
renewable resources 347–8 distribution system operator (DSO)
distributed generations (DGs), 8–9, 363, 460, 486
benefits of 348 diurnal tides with monthly variation
economic benefits 349–51 105
emissions reduction (ER) benefits double-layer capacitors (DLCs) 261
with integration of renewable doubly fed induction generator
DG units 351–2 (DFIG) 41, 53, 123, 154–5,
environmental benefits 348 449
technical benefits 348–9 wind turbine (WT) with 56–60
microgrids 362 doubly fed induction machines
challenges facing MG (DFIM) 46, 196
implementation 364–5 d-q frame, dynamic model in
concept 363 55–6
control of 367–9 power and torque relations 51
energy storage systems (ESSs), reactive power calculations
role of 365–6 53–4
islanding detection 369 steady-state model 47–9
operation functions 363–4 steady-state phasor diagram 49–51
protection strategy 366–7 downwind machines 20–1
operation requirements for DERs drag force 299
integration into power systems D-STATCOM, application of 500–5
352 dynamic dispatch (DD) problem 385
frequency regulation 353 dynamic programming (DP) 409
earth, gravitational forces of moon and electric double-layer capacitors: see

sun on 101–4 supercapacitors (SCs)
ebb generation 114–15 electric generators 280
economic dispatch (ED) 401, 408 electricity generation 75–6, 153
implementation 417–19 direct drive wave energy 154–6
economic dispatch (ED) of hybrid from wind and solar 1
thermal and wind plants 373 electric motors 280, 282
economic/emissions dispatch brushed DC motors 282–9
problem 382 brushless DC (BLDC) motors 292
multi-objective stochastic equivalent circuit and general
dispatch 383–5 equations 294–6
energy storage system (ESS), role operational theory 292–3
of 385–96 switching process 293–4
optimization problem, formulation dynamic and regenerative braking in
of 377–8 DC motors 289–92
of power system comprising thermal switched reluctance (SR) motors
and wind energy sources 374 296–9
single-objective ED model 379 electric power systems (EPSs) 1, 459
conventional ED (CED) with electric vehicle management system
thermal units 379–80 (EVMS) 461
deterministic ED (DED) electric vehicles 277, 457
incorporating wind power 380 batteries 282
stochastic ED (SED) battery charging infrastructure 301
incorporating wind power chargers 304–6
380–2 charging stations 303–4
economic/emissions dispatch problem charging times 306–8
373–4, 382 integration with renewable energy
multi-objective stochastic dispatch sources (RESs) 459–60
383–5 performance 299–301
eddy currents 288 plug-in electric vehicle (PEV)
electrical efficiency 263 technologies 278
electrical energy storage (EES) 181 design features 280–1
characteristics 190–2 operation modes 279–80
classification 189 electric vehicles (EVs) for power grid
concept 183 frequency regulation 461
critical challenges 184–5 application 473–9
elements of 183 control architecture 466
fields of energy storage aggregator control centre (ACC)
applications 192–3 468–9
functions of 185–9 local controller (LC) 467–8
need for 183–4 large-scale EV participation in
and renewable energy 182–3 frequency regulation 462
electrical generators in wind systems: primary frequency control 463
see wind systems, electrical secondary frequency control
generators in 463–4
Index 559
tertiary control 464–6 specific energy 262

scheduling the charging/discharging lifetime aspect 264
processes for EVs fleet 470–3 supercapacitor energy storage
system simulation 469–70 (SCES), working principle of
electrification of transportation sector 261–2
457–8 energy loss in solar cell 73–5
electrochemical energy storage (EcES) energy storage device 183
systems 243 energy storage systems (ESSs) 5–6,
battery energy storage (BES) 10, 71–2, 347, 386, 432, 449,
technologies, comparison of 459, 485
255 role of 365–6, 385–96
economical comparison 256–8 environmental economic dispatch
technical comparison 256 (EED) problem 382
battery modelling 252–5 Environmental Protection Agency
conventional secondary BES (EPA) 176
technologies equivalent per-phase impedance 45
lead-acid (PbA) batteries 243–4 ethanol 172, 177
lithium ion (Li-ion) batteries exchange current density 229
245–6 extra-high voltage (EHV) 518–19
metal air (Me-air) batteries 246–7
nickel-based batteries 244–5 false data injection attacks 538
sodium nickel chloride (NaNiCl) Faradaic processes 264
batteries 248–50 Faraday constant 222
sodium sulphur (NaS) batteries fetch 129
247–9 field windings 287
flow batteries 250 fill factor (FF) 86
hybrid flow batteries (HFEs) 252 first-law efficiency 274
redox flow batteries (RFBs) Fleming’s left-hand rule 283
251–2 flexible alternating current
electrochemical storage technologies transmission system (FACTS)
258 devices 486
electromagnetic radiation 73 floating cylinder 152
electromotive force (EMF) 217 float life 264
electron–hole pair 75 flood generation, tidal barrage 115
electronic communication system, flow batteries 250
main components of 514 hybrid flow batteries (HFEs) 252
electrostatic energy storage systems redox flow batteries (RFBs) 251–2
261 flow rate of mass 22
advantages and disadvantages 265 fly ash 176
applications 264–5 flywheel battery 210
electrical performance 262 flywheel energy storage (FES) 208
discharge behaviour 263 applications 212–13
efficiency 263–4 architecture and operation 209–10
power 263 characteristics 212
reversibility and cycle life 264 mathematical relations 210–12
force to accelerate the vehicle 300 branch and bound 410

fossil-fuelled steam thermal units 402 dynamic programming (DP) 409
fuel cell (FC) Lagrangian relaxation (LR) 410
efficiency 224–5 linear programming 410–11
equivalent circuit 231–3 priority list 409
inputs and outputs of 221 genetic algorithm (GA) 402
physical structure 219 implementation 413–15
principles of FC operation 218–21 Gibbs free energy 221–2, 224, 226
reversible open-circuit voltage of global positioning system (GPS) 12
221–4 gravitational forces 101–6
types of 235–6 greenhouse gases (GHGs) 382, 457
fuel cell electric vehicles (FCEVs) 278 grid 1
fuel costs 404–5 grid components, security of 536
full moon 102 communication protocol 538
false data injection attacks 538
gasoline–ethanol blends 177 power system state estimation (SE)
gas turbine 239–41, 345, 402 model 537
generation scheduling for power grids process control systems (PCSs)
with renewables 401 536–7
economic dispatch 408 smart meters (SMs) 537
problem representation and solution grid-connected DFIG WT (case
strategy 413 study) 334–41
artificial neural network (ANN) grid-connected PV (case study) 334
implementation 416–17 grid privacy 535
decomposition methods 415–16 grid side converter (GSC) 47
economic dispatch (ED) grid tariffs 8
implementation 417–19
genetic algorithm (GA) heat pump storage 273
implementation 413–15 heat recovery steam generator
priority list (PL) implementation (HRSG) 240
416 heat transfer fluid (HTF) 271–2
security-constrained generation heliostats 95
scheduling (SCGS) 419–27 heuristic algorithms 377
thermal generating units 402 heuristic and expert systems 411
characteristic of 402–3 high-head hydraulic turbines 151
thermal scheduling problem, high-pressure oil hydraulic motors
formulation of 403 152
fuel costs 404–5 direct mechanical interface 152–3
problem objective function 404 high voltage (HV) lines 518–19
system constraints 405–6 hill climbing force 300
transition costs 405 home area network (HAN) 515–16
unit commitment 404 Hopfield neural network 412
unit constraints 407–8 horizontal axis turbines 122
thermal scheduling techniques 408 horizontal axis wind turbines
artificial intelligence (AI) 411–13 (HAWT) 18, 21
Index 561
downwind machines 20–1 linear Fresnel reflector systems 92, 94

upwind machines 20 linear permanent magnet generator
hybrid electric vehicle 277 (LPMG) 155–6
hybrid flow batteries (HFEs) 252 liquid biofuels 177
hybridization 97 lithium ion (Li-ion) batteries 245–6
hydraulic turbines 151–2 basic components of 246
hydrogen FCs 217 load-based indices
hydrogen storage 238 average system interruption duration
hysteresis band (HB) 319 frequency index (ASIDI) 359
hysteresis loss 288 average system interruption
frequency index (ASIFI) 359
IEEE 14-bus system 435 load flow calculations
IEEE 24-bus system 549 flow chart for 434
induction generators (IGs) 41–2 steady-state 432
information and communication energy storage system (ESS), role
technology (ICT) 529 of 443–9
intelligent electronic devices (IEDs) load flow algorithm 433–43
519, 533 for transient stability 449–54
internal combustion (IC) engine 277 load-following plants 185
internal electromotive force (EMF) load-forecast uncertainty 416
254 load profiles 496–500
Internet access 524–5 local area network (LAN) 515
inverter interfaced DG (IIDG) unit local control (LC) 368–9, 486
367 local controller (LC) 466–8
inverters 280 Lorentz force 283
iron losses 288 low voltage (LV) lines 519
lunar cycle 102
Kaplan turbines 150
kinetic energy 22 magnetic energy storage,
Kyoto Protocol 382 superconducting 266
applications 267
Lagrangian method 412 main branch resistance 255
Lagrangian multipliers 410 mass, flow rate of 22
Lagrangian relaxation (LR) 410 MATLAB/Simulink model 500,
lambda iteration 377 508
landfill gas/biogas, collecting 176–7 of battery energy storage system
Laplace variable 254–5 (BESS) 498
latent heat storage (LHS) 269–70 charger controller construction of
LC filters 331–3 BESS in 499
LCL filters 333 integrated model of microgrid (MG)
lead-acid (PbA) batteries 243–4 in 501
discharging and charging of 244 of photovoltaic (PV) system 490,
L filter 331 493
life cycle cost (LCC) 191–2, 257 of PMSG with MPPT using P&O
linear concentrator systems 92–4 technique 495
proton-exchange membrane fuel cell microgrids (MGs) 4–5, 362–3, 485

(PEMFC) system simulated in architecture 486–7
497 battery energy storage system 496
maximum power point (MPP) 84, 86, challenges facing implementation
489 of 364
maximum power point tracking financial issues 365
(MPPT) system 235, 307, 486, regulatory issues 365
488–90 technical issues 364–5
mechanical energy storage (MES) concept 363
systems 195 control of 367
compressed air energy storage local control (LC) 368–9
(CAES) system 203 microgrid central control
advantages and disadvantages of (MGCC) 367–8
CAES 207–8 D-STATCOM, application of
modelling 206–7 500–5
principles of CAES operation energy storage systems (ESSs),
204–5 role of 365–6
flywheel (FW) energy storage 208 islanding detection 369
applications 212–13 load profiles 496–500
FES architecture and operation operation functions 363–4
209–10 photovoltaic (PV) system 488
FES characteristics 212 boost converter 490
mathematical relations 210–12 maximum power point tracking
pumped hydroelectric storage (MPPT) 489–90
(PHES) system perturb and observe (P&O)
applications 201 algorithm 490
disadvantages 203 protection strategy 366–7
functions of the main components proton-exchange membrane fuel
of 197 cell 494–6
main advantages of 202 simulation model 502–3
modelling of PHES dynamics switched filter compensator (SFC),
197–201 application of 505–8
physical principles 195–7 wind energy system 491–4
types of PHES 201–2 microturbines 241, 345
mechanical interfaces 149 microwave radio systems 523–4
air turbines 149–51 mixed tides 105
high-pressure oil hydraulic motors momentary average interruption
152–3 event frequency index
hydraulic turbines 151–2 (MAIFIE) 359
mechanical losses 289 momentary average interruption
medium voltage (MV) lines 519 frequency index (MAIFI) 359
metal air (Me-air) batteries 246–7 moon phases 102–3
methanation 239 moon’s gravitational attraction 102
microgrid central control (MGCC) motor convention 51
367–8, 486–7 multi-junction cells 70
Index 563
multi-objective PSO (MO-PSO) site selection 164–5

algorithms 383 technical limitations and challenges
multi-objective stochastic dispatch 164
383–5 tidal current energy conversion
devices 122–4
nacelle 17–18 tidal energy
narrowband PLC (NB-PLC) systems environmental impacts of 125
520 estimation 105–12
natural gas 347 extraction 112–22
neap tides 103–5, 108 tidal phenomena, physics of 101–5
neighbourhood area network (NAN) tidal power
515–17 cons of 126–7
Nernst equation 227 pros of 126
Nernst voltage 227 tidal power plants connected to
network attacks in power grids 532 utility grid 124–5
attacks targeting integrity and wave energy
confidentiality 534 estimation of 139–43
denial-of-service (DoS) attacks 533–4 advantages of 157
new moon 102 disadvantages of 157–8
nickel-based batteries 244–5 wave energy concept 129
nickel cadmium (NiCd) batteries wave energy converters (WECs)
advantages 245 143–50
disadvantages 245 wave propagation 133–7
nickel metal hydrate (NiMH) batteries ocean power technology’s (OPT)
245 power buoy 145
nominal operating cell temperature ocean thermal energy conversion
(NOCT) 87 (OTEC) operation 158
non-renewable resources 347 closed-cycle OTEC 160–3
nuclear steam thermal units 402 and environment 163–4
open-cycle OTEC 159–60
ocean energy 101, 129 ocean waves 130
advantages 165–6 ocean waves creation 129
deepwater waves, fundamentals of ripples 130
137–9 capillary wave 131
disadvantages 166 swell waves 131–2
electricity generation 153–6 shallow waves/shore waves 133
mechanical interfaces 149–53 ohmic losses 230
ocean thermal energy conversion ON–OFF control 330–1
(OTEC) operation 158 open-circuit voltage 254
closed-cycle OTEC 160–3 open-cycle OTEC 159–60
and environment 163–4 operating life 264
open-cycle OTEC 159–60 operating reserves units 185
ocean waves creation 129–33 optical communications 522
power take-off (PTO) systems optimal power flow (OPF) 418
156–7 organic PVs 69
oscillating hydrofoil turbines 123 temperature and insolation impact

oscillating water column (OWC) 86–9
148–9 photovoltaic (PV) effect 67
oscillating wave surge converter photovoltaic (PV) system 347, 488
147–8 boost converter 490
outflow generation: see ebb generation maximum power point tracking
output filters 331 (MPPT) 489–90
LC filters 331–3 perturb and observe (P&O)
LCL filters 333 algorithm 490
overtopping device 148 photovoltaic electric vehicles
(PVEVs) 278
parasitic branch current 255 pitch-control system 38
Pareto front 383 plug-in battery electric vehicle
Pareto-optimal front 382 (PBEV) 278
particle swarm optimization (PSO) plug-in electric vehicle (PEV)
378, 382 technologies 278, 457, 474
peaking units 185 design features 280–1
Pelamis wave power 146 operation modes 279–80
pendulum system 152 plug-in hybrid electric vehicle
permanent magnet (PM) synchronous (PHEV) 278–9, 301, 457
machines 60 point absorber 144–6
connected to power grid 63 point of common coupling (PCC) 5,
structure, equivalent circuit and 47, 363, 486
phasor diagram 60–2 polymer electrolyte membrane fuel cell
permanent magnets (PMs) 42 (PEMFC) 233
permanent magnet synchronous computer modelling 234–5
generator (PMSG) 41, 123, resistances, calculation of 233–4
491 power conditioning systems (PCSs)
perturb and observe (P&O) algorithm 266, 313, 485
490 AC–AC conversion 330
phase-change materials (PCMs) ON–OFF control 330–1
269–70 phase control 331
phase-change storage: see latent heat AC–DC conversion 313
storage (LHS) diode rectifiers cascaded with a
phase control 331 boost converter 317–19
phase-controlled rectifiers 315–17 phase-controlled rectifiers
phase-controlled single-phase full- 315–17
bridge rectifier 316 PWM-controlled rectifiers
phasor measurement units (PMUs) 11, 319–20
519 uncontrolled rectifiers 314–15
photosynthesis process 170 case studies 334
photovoltaic (PV) cell equivalent grid-connected DFIG WT
circuit 76–81 334–41
photovoltaic (PV) cell power curve 84 grid-connected PV 334
shading impact 89–91 DC–AC conversion 324
Index 565
single-phase inverters, operation proportional integral derivative (PID)

of 327–9 controller 504
sinusoidal PWM 325–6 proton-exchange membrane fuel cell
three-phase inverters (two-level), (PEMFC) 486, 494–7
operation of 329–30 perturb and observe (P&O)
DC–DC conversion 320 application to MPPT for 496
boost converter 322–4 pulse-width modulation (PWM) 294
buck-boost converters 324 -controlled rectifiers 319–20
buck converter 321–2 switching 506
output filters 331 pumped hydroelectric storage (PHES)
LC filters 331–3 system
LCL filters 333 advantages of 202
power conversion system 183 applications 201
power curve 38 disadvantages of 203
power electronic circuits 71 functions of the main components
power formulae, derivation of 142–3 of 197
power from wind 22–5 modelling of PHES dynamics
power grid 1 197–201
power grid examination 12 physical principles 195–7
power intensity 23 types of 201–2
power line communication (PLC)
system 517–22 radio frequency (RF) energy 518
power quality enhancement device random variables 543
(PQED) 486 continuous random variables and
power system reliability 2 their distributions 544
power system security 2 discrete random variables 543
power system stability 4 Rankine cycle engines 94
power system state estimation (SE) Rankine-cycle thermal efficiency 163
model 537 rate of change of frequency (RoCoF)
power take-off (PTO) systems 156–7 359–61, 474
power-to-gas (PtG) energy system Rayleigh pdf 33, 376
217, 236 rectification 313
applications 239 rectifier 313
gas turbines 239–41 redox flow batteries (RFBs) 251–2,
microturbines 241 256
principle of operation 236–9 regenerative braking 290
power tower systems 95–9 reliability 357
priority list (PL) 402 load-based indices 359
implementation 416 momentary indices 359
probability and statistics 33–5 sustained interruption indices 357–8
probability density function (pdf) virtual inertia 359–62
32–3, 373–4, 543–6 renewable energy, electrical energy
probability mass function (pmf) 544 storage and 182–3
process control systems (PCSs) renewable energy sources (RESs) 4–5,
536–7 12, 157, 307, 431, 459, 485
electric vehicles (EVs) integration stochastic ED (SED) incorporating

with 459–60 wind power 380–2
renewable resources 347–8 single-phase inverters, operation of
resistance 254 327–9
resultant acceleration 138 single-tank thermocline storage 272–3
ring main units (RMUs) 519 sinusoidal PWM 325–6
ripples 130 smarter grid (SG) 6, 513–14
capillary wave 131 communication infrastructure 518
swell waves 131–2 consumers’ role 8–9
rolling resistance force 299 definition 9–12
rotor current 44 distribution system operators’
rotor efficiency 29–30 (DSOs) role 9
rotor equivalent resistance 45 power line communication (PLC)
rotor impedance 44 in 520
rotor power, apparent 45 smart grid (SG), defined 10
rotor reactive power 53 smart meter (SM) 9, 461, 513, 535,
rotor-side converter (RSC) 47, 451 537
sodium-heat engines 94
satellite communication systems 524 sodium nickel chloride (NaNiCl)
second-law efficiency 274–5 batteries 248–50
security 2 sodium sulphur (NaS) batteries 247–9
security attacks in communication solar-electric efficiencies 97
networks 532 solar energy, disadvantages of 73
security-constrained generation solar photovoltaic technology
scheduling (SCGS) 402, 419–27 basics 66–73
security constraints 3 electricity generation 75–6
semidiurnal tides 104–5 energy loss in solar cell 73–5
monthly cycle 105 photovoltaic (PV) cell equivalent
physical explanation of 106 circuit 76–81
sensible heat storage (SHS) 269 photovoltaic (PV) cell power curve
service life 264 84
shallow waves/shore waves 133 shading impact 89–91
shelf life 264 temperature and insolation
Shockley diode equation 76 impact 86–9
shore waves 133 solar arrays 83–4
short-term storage 5 solar modules 81–3
silicon (Si) crystals 67 solar power, concentrating 91
simulated annealing (SA) 411 dish-engine systems 94–5
single-crystal silicon 70 linear concentrator systems 92–4
single-objective economic dispatch power tower systems 95–9
model 379 solar PV 486
conventional ED (CED) with solar radiation 374
thermal units 379–80 solar systems, advantages of 73
deterministic ED (DED) solid biomass 172
incorporating wind power 380 spinning reserves 185
Index 567
spring tides 103–4, 108 Customers Experiencing Multiple

squirrel cage induction generator Interruptions (CEMIn) 358
(SCIG) 41, 43, 123 customer total average interruption
relations and equivalent electric duration index (CTAIDI) 358
circuit at steady state 43–5 system average interruption duration
transient state, equivalent circuit in index (SAIDI) 357
45–6 system average interruption
stability phenomenon 4 frequency index (SAIFI) 357
state of charge (SoC) 253, 255, 457–8 swell waves 131–2
statistics, probability and 33–5 switched filter compensator (SFC),
stator reactive power 53 application of 505–8
stator voltage 44 switched reluctance (SR) motors
steady-state circuit equation 294 296–9
steady-state load flow calculations 432 symmetrical ripples 131
energy storage system (ESS), role synthetic inertia: see virtual inertia
of 443–9 system average interruption duration
load flow algorithm 433–43 index (SAIDI) 357
step-up converter: see boost converter system average interruption frequency
Stochastic Dynamic Economic index (SAIFI) 357
Emission Dispatch (SDEED) system constraints, in power system
concept 392, 395 405
stochastic economic dispatch (SED) active power balance 405–6
incorporating wind power emission constraint 406
380–2 load forecast error and system
storage-tank effectiveness 274 reserve requirements 406
stored energy, amount of 269 transmission network constraints
submerged pressure differential 406
device 147 system reliability 2
supercapacitor energy storage (SCES)
261 Tabu search (TS) 412
working principle of 261–2 Tafel equation 229
supercapacitors (SCs) 261 terminal resistance 255
superconductive magnetic energy terminator devices 146–7
storage (SMES) systems 266–7 test system 549
supervisory control and data renewable energy sources, data of
acquisition (SCADA) 523–4, 552
531, 538 single-line diagram of 549
surface wave 134 thermal energy storage (TES) 267
sustained interruption indices classification and characteristics of
average system availability index 268–9
(ASAI) 358 latent heat storage (LHS) 269–70
customer average interruption performance 274
duration index (CAIDI) 358 first-law efficiency 274
customer average interruption second-law efficiency 274–5
frequency index (CAIFI) 358 storage-tank effectiveness 274
sensible heat storage (SHS) 269 thermochemical heat (TCH) energy

storage systems 271 storage 270–1
heat pump storage 273 thermochemical technology 174
single-tank thermocline storage thermocline storage tank 273
272–3 thin-film PVs 68–9
two-tank direct storage 271–2 three-phase inverters (two-level),
two-tank indirect storage 272 operation of 329–30
thermochemical heat (TCH) energy tidal barrage 113
storage 270–1 ebb generation 114–15
thermal generating units 402 energy calculations 117–19
characteristic of 402–3 flood generation 115
thermal scheduling problem, pumping at high tide 115–16
formulation of 403 two-way generation scheme 116–17
fuel costs 404–5 tidal current generators 123
problem objective function 404 direct drive permanent magnet
system constraints 405 (DDPM) generator 123
active power balance 405–6 doubly fed induction generator
emission constraint 406 (DFIG) 123
load forecast error and system permanent magnet synchronous
reserve requirements 406 generator (PMSG) 123
transmission network constraints squirrel cage induction generator
406 (SCIG) 123
transition costs 405 wound rotor induction generator
unit commitment (UC) 404 (WRIG) 123
unit constraints 407 tidal current turbines 122
real power operating limits 407 horizontal axis turbines 122
unit minimum up and downtime oscillating hydrofoil turbines 123
407 vertical axis turbines 122
unit ramp rate limits 407–8 tidal energy 105
thermal scheduling techniques 408 environmental impacts 125
artificial intelligence (AI) 411 tidal energy converter (TEC) 119
artificial neural network (ANN) tidal frequency 112
411–12 tidal phenomena, physics of 101
evolutionary algorithms 412–13 types of tides 104–5
heuristic and expert systems tidal power 125
411 cons of 126–7
simulated annealing (SA) 411 pros of 126
Tabu search (TS) 412 tidal power plants connected to utility
branch and bound 410 grid 124
dynamic programming (DP) 409 capacity limits 124
Lagrangian relaxation (LR) 410 power quality 124–5
linear programming 410–11 system analysis 125
priority list 409 tidal range 102
thermal technology 173–4 sinusoidal variation of 118
by direct combustion 174 tidal resonance 112
Index 569
tidal stream generation (TSG) system vehicle to grid (V2G) concept 460–1
119 vertical axis turbines 122
resource assessment 121–2 vertical axis wind turbine (VAWT) 18,
tidal energy calculation 120–1 21–2
tidal stream generators 120 virtual inertia 359–62
drawbacks of 124 virtual power plant (VPP) 460–1, 538
tidal streams 105 voltage efficiency 263
tidal wave 110–12 voltage induced on the rotor 44
water motion in 111 voltage-source inverter (VSI) 324
tide, height of 110
tide formation 104 waste-to-energy furnace 176
tide-generating force ratio (TGFR) waste-to-energy plants, disposing ash
107 from 176
tide generating forces 106–10 water electrolyser 238
tides 102 water flow 199
tip-speed ratio (TSR) 30–2, 41 water starting time 200
torque 285 wave amplitude 112
total capital cost (TCC) 191, 257 Wave Dragon 148
transient stability, load flow wave energy 139, 157
calculations for 449–54 advantages of 157
transition costs 405 concept 129
transmission system operator (TSO) disadvantages of 157–8
460 power formulae, derivation of
transportation sector, electrification 142–3
of 457–8 wave energy converters (WECs) 129,
two-tank direct storage 271–2 143, 150
two-tank indirect storage 272 location of 144
nearshore 144
ultra-capacitors: see supercapacitors offshore 144
(SCs) onshore 144
ultra-narrowband PLC (UNB-PLC) modes of operation 147–9
systems 520 oscillating water column (OWC)
uncontrolled rectifiers 314–15 148
unit commitment (UC) 401, 404 oscillating wave surge converter
unit constraints, in power system 407 147–8
real power operating limits 407 overtopping device 148–9
unit minimum up and downtime 407 submerged pressure differential
unit ramp rate limits 407–8 device 147
upwind machines 20 types of 144–7
attenuators (line absorbers) 146
vanadium redox flow battery (VRFB) point absorber 144–6
251–2 terminator devices 146–7
vehicle to grid (V2G) challenges 480 wave-formed ripples 131
economic challenges 480 wavelength 110
technical challenges 480 wave propagation 133–7
wave surface velocity 139 squirrel cage induction

Weibull distribution function generators 43–6
546 WT with DFIG control system
mean and variance, calculation of 56–60
547–8 permanent magnet (PM)
Weibull pdf 33 synchronous machines 60
Weibull probability density 375 connected to power grid 63
weighted aggregation (WA) method structure, equivalent circuit and
383 phasor diagram 60–2
wide area network (WAN) 516 wind turbines (WTs) 17, 377
wind, power from 22–5 with doubly fed induction generator
wind characteristics 35 (DFIG) control system 56–60
airfoils 35–9 horizontal axis wind turbines
capacity factor (CF) 39–41 (HAWT) 18
wind energy 23, 347 downwind machines 20–1
wind energy system (WES) 486, upwind machines 20
491–4 vertical axis wind turbines
local controller of 494 (VAWT) 21–2
wind power, average 32, 34 wired communications
probability and statistics 33–5 digital subscriber lines (DSLs)
probability density function (pdf) 522–3
32–3 optical communications 522
wind power, factors affecting 25 power line communication (PLC)
air density 25 system 517–22
pressure impact 25–6 wireless communications
temperature impact 26 cellular communication networks
rotor efficiency 29–30 523–4
tip-speed ratio (TSR) 30–2 microwave radio systems 523–4
wind speed, factors affecting 26–8 satellite communication systems
wind power modelling using proper 524
pdf 374–7 wound rotor induction generator
wind power per unit area 23 (WRIG) 123
wind systems, electrical generators in wound rotor induction machine
41 (WRIM) 46
asynchronous induction generators
41 ZEBRA batteries: see sodium nickel
doubly fed induction machines chloride (NaNiCl) batteries
(DFIM) 46–56 Zn-air battery 247–8

Power Grids With Renewable Energy

Uploaded by

Copyright:

Available Formats

Power Grids With Renewable Energy

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Power Grids With Renewable Energy

Uploaded by

Copyright:

Available Formats

What are some of the different types of renewable energy sources discussed?

What are some of the different types of renewable energy sources discussed?

What are some challenges associated with integrating renewable energy into power grids?

What are some challenges associated with integrating renewable energy into power grids?

IET ENERGY ENGINEERING SERIES 67

Power Grids with

The Institution of Engineering and Technology

The Institution of Engineering and Technology

British Library Cataloguing in Publication Data

ISBN 978-1-83953-027-2 (hardback)

Typeset in India by MPS Limited

About the authors xix

1 Towards the new trend of power grids 1

Part I: Renewable energy sources 15

3.2 Solar PV technology 66

4 Ocean energy: tidal energy 101

5 Ocean energy: wave and thermal energy 129

5.3 Wave propagation 133

6 Biomass energy 169

Part II: Energy storage systems 179

7.3.1 Elements of EES 183

8 Mechanical energy storage systems 195

9 Chemical energy storage systems: fuel cells and power-to-gas 217

10 Electrochemical energy storage systems 243

11 Electrostatic, magnetic and thermal energy storage 261

12 Energy storage system application: electric vehicles 277

12.4.3 Brushless DC motors 292

Part III: Renewables integration 311

14 Integration of distributed energy resources 345

15 Economic dispatch of hybrid thermal and wind plants 373

16 Generation scheduling for power grids with renewables 401

16.5 Review of thermal scheduling techniques 408

17 Load flow analysis for power systems with renewables 431

18 Integration of electric vehicles with renewables into power grids 457

19 Microgrids: modelling and control 485

Part IV: Associated systems 511

21 Cybersecurity in power systems 529

Appendix A: Probability density and Weibull distribution functions 543

Appendix B: Test system 549

Abdelhay A. Sallam is Professor Emeritus in Port Said University, Egypt. He has

Om P. Malik is Professor Emeritus in the University of Calgary, Canada. His pub-

amount of research and development is taking place in ensuring utilization and

1.1 Conventional power systems

Figure 1.1 A typical conventional power system

MGs are small distribution systems encompassing a group of electricity consumers

1.3 Energy storage systems

Figure 1.2 Schematic diagram of an MG connected to utility distribution network

electricity storage systems at different locations in the MGs (decentralized storage).

1.4 Smarter grid

Utility distribution network

Figure 1.3 Power grid supported by MG and compressed air storage

system components’ integration, interface and better utilization of the existing

1.4.1 Consumers’ role

intelligent appliances and smart chargers. Smart meters as a kind of intelligent

1.4.2 DSOs’ role

1.4.3 Smart grid definition

SCADA for generation

Main coordinator and

Figure 1.4 SG perspective of the future

MG, distribution system, transmission system and central generation is locally