Advances in energy systems : the large-scale renewable energy integration challenge 9781119508281, 1119508282

Citation preview

Advances in Energy Systems

Advances in Energy Systems The Large‐scale Renewable Energy Integration Challenge

Edited by

PETER D. LUND Aalto University, Finland

JOHN A. BYRNE University of Delaware, USA

REINHARD HAAS Vienna University of Technology, Austria

DAMIAN FLYNN University College Dublin, Republic of Ireland

This edition first published 2019 © 2019 John Wiley & Sons Ltd All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions. The right of Peter D. Lund, John A. Byrne, Reinhard Haas & Damian Flynn to be identified as the authors of the editorial material in this work has been asserted in accordance with law. Registered Offices John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK Editorial Office The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com. Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats. Limit of Liability/Disclaimer of Warranty While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials, or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. Library of Congress Cataloging‐in‐Publication Data Names: Lund, Peter D., editor. Title: Advances in energy systems : the large-scale renewable energy   integration challenge / edited by Peter D. Lund (Aalto University School   of Science, Finland) [and three others]. Description: Hoboken, NJ : Wiley, [2019] | Includes bibliographical   references and index. | Identifiers: LCCN 2018046918 (print) | LCCN 2018047879 (ebook) | ISBN   9781119508335 (Adobe PDF) | ISBN 9781119508328 (ePub) | ISBN 9781119508281 (hardcover) Subjects: LCSH: Renewable energy sources. | Power resources–Forecasting. Classification: LCC TJ808 (ebook) | LCC TJ808 .A3838 2019 (print) |   DDC 333.79/4–dc23 LC record available at https://lccn.loc.gov/2018046918 Cover Design: Wiley Cover Images: (Left top) © Eviart/Shutterstock, (Right top) © Stocksolutions/Shutterstock, (Left bottom) © metamorworks/Shutterstock, (Right bottom) © guteksk7/Shutterstock Set in 9/11pt Times by SPi Global, Pondicherry, India

10 9 8 7 6 5 4 3 2 1

Contents

List of Contributors

ix

Prefacexi PART I:  ENERGY SYSTEM CHALLENGES

1

  1 Handling Renewable Energy Variability and Uncertainty in Power System Operation Ricardo Bessa, Carlos Moreira, Bernardo Silva and Manuel Matos

3

  2 Short‐Term Frequency Response of Power Systems with High Nonsynchronous Penetration Levels Lisa Ruttledge and Damian Flynn

27

  3 Technical Impacts of High Penetration Levels of Wind Power on Power System Stability Damian Flynn, Zakir Rather, Atle Rygg Årdal, Salvatore D’Arco, Anca D. Hansen, Nicolaos A. Cutululis, Poul Sorensen, Ana Estanqueiro, Emilio Gómez‐Lázaro, Nickie Menemenlis, Charles Smith and Ye Wang

47

  4 Understanding Constraints to the Transformation Rate of Global Energy Infrastructure Joe L. Lane, Simon Smart, Diego Schmeda‐Lopez, Ove Hoegh‐Guldberg, Andrew Garnett, Chris Greig and Eric McFarland

67

  5 Physical and Cybersecurity in a Smart Grid Environment Jing Xie, Alexandru Stefanov and Chen‐Ching Liu

85

  6 Energy Security: Challenges and Needs Benjamin K. Sovacool

111

  7 Nuclear and Renewables: Compatible or Contradicting? Lutz Mez

119

PART II:  PERSPECTIVES ON GRIDS

127

  8 Smart‐Grid Policies: An International Review Marilyn A. Brown and Shan Zhou

129

  9 A View of Microgrids Joao A. P. Lopes, Andre G. Madureira and Carlos Moreira

149

10 New Electricity Distribution Network Planning Approaches for Integrating Renewables Fabrizio Pilo, Gianni Celli, Emilio Ghiani and Gian G. Soma

167

vi Contents

11 Transmission Planning for Wind Energy in the United States and Europe: Status and Prospects Charles Smith, Dale Osborn, Robert Zavadil, Warren Lasher, Emilio Gómez‐Lázaro, Ana Estanqueiro, Thomas Trotscher, John Tande, Magnus Korpas, Frans Van Hulle, Hannele Holttinen, Antje Orths, Daniel Burke, Mark O’Malley, Jan Dobschinski, Barry Rawn, Madeline Gibescu and Lewis Dale

187

12 Opportunities and Barriers of High‐Voltage Direct Current Grids: A State‐of‐the‐Art Analysis Debora Coil‐Mayor and Jürgen Schmid

201

13 Wireless Power Transmission: Inductive Coupling, Radio Wave, and  Resonance Coupling Naoki Shinohara

211

PART III:  FLEXIBILITY MEASURES

221

14 The Role of Large‐Scale Energy Storage Under High Shares of Renewable Energy Shin‐ichi Inage

223

15 The Role of Electric Vehicles in Smart Grids Matthias D. Galus, Marina González Vayá, Thilo Krause and Göran Andersson

245

16 Use of Electric Vehicles or Hydrogen in the Danish Transport Sector in 2050? Klaus Skytte, Amalia Pizarro and Kenneth B. Karlsson

265

17 Comparison of Synthetic Natural Gas Production Pathways for the Storage of  Renewable Energy Sebastian Fendt, Alexander Buttler, Matthias Gaderer and Hartmut Spliethoff

279

18 Storage and Demand‐Side Options for Integrating Wind Power Aidan Tuohy, Ben Kaun and Robert Entriken

303

19 On the Long‐Term Prospects of Power‐to‐Gas Technologies Amela Ajanovic and Reinhard Haas

321

20 Wind Integration: Experience, Issues, and Challenges Hannele Holttinen

341

21 Quantifying the Variability of Wind Energy Simon Watson

355

22 Capacity Value Assessments of Wind Power Michael Milligan, Bethany Frew, Eduardo Ibanez, Juha Kiviluoma, Hannele Holttinen and Lennart Söder

369

23 Hydropower Flexibility for Power Systems with Variable Renewable Energy Sources: An IEA Task 25 Collaboration Daniel Huertas‐Hernando, Hossein Farahmand, Hannele Holttinen, Juha Kiviluoma, Erkka Rinne, Lennart Söder, Michael Milligan, Eduardo Ibanez, Sergio M. Martinez, Emilio Gómez‐Lázaro, Ana Estanqueiro, Luis Rodrigues, Luis Carr, Serafin van Roon, Antje Orths, Peter B. Eriksen, Alain Forcione and Nickie Menemenlis

385

24 Contribution of Bulk Energy Storage to Integrating Variable Renewable Energies in Future European Electricity Systems Karl A. Zach and Hans Auer

407

25 Characterization of Demand Response in the Commercial, Industrial, and  Residential Sectors in the United States Sila Kiliccote, Daniel Olsen, Michael D. Sohn and Mary A. Piette

425

Contents vii

26 Simplified Analysis of Balancing Challenges in Sustainable and Smart  Energy Systems with 100% Renewable Power Supply Lennart Söder

445

PART IV:  CHANGING ELECTRICITY MARKETS

459

27 Who Gains from Hourly Time‐of‐Use Retail Prices on Electricity? An Analysis of  Consumption Profiles for Categories of Danish Electricity Customers F. M. Andersen, H. V. Larsen, Lena Kitzing and P. E. Morthorst

461

28 Designing Electricity Markets for a High Penetration of Variable Renewables Jenny Riesz and Michael Milligan 29 Multivariate Analysis of Solar City Economics: Impact of Energy Prices, Policy, Finance, and Cost on Urban Photovoltaic Power Plant Implementation John A. Byrne, Job Taminiau, Kyung N. Kim, Joohee Lee and Jeongseok Seo 30 The Influence of Interconnection Capacity on the Market Value of Wind Power Carlo Obersteiner 31 Research with Disaggregated Electricity End‐Use Data in Households: Review and Recommendations Ian H. Rowlands, Tobi Reid and Paul Parker 32 Household Electricity Consumers’ Incentive to Choose Dynamic Pricing Under Different Taxation Schemes Jonas Katz, Lena Kitzing, Sascha T. Schröder, F. M. Andersen, P. E. Morthorst and Morten Stryg

479

491 507

517

531

Index545

List of Contributors

Amela Ajanovic F. M. Andersen Göran Andersson Atle Rygg Årdal Hans Auer Ricardo Bessa Marilyn A. Brown Daniel Burke Alexander Buttler John A. Byrne Luis Carr Gianni Celli Debora Coil‐Mayor Nicolaos A. Cutululis Lewis Dale Salvatore D’Arco Jan Dobschinski Robert Entriken Peter B. Eriksen Ana Estanqueiro Hossein Farahmand Sebastian Fendt Damian Flynn Alain Forcione Bethany Frew Matthias Gaderer Matthias D. Galus Andrew Garnett Emilio Ghiani Madeline Gibescu Emilio Gómez‐Lázaro Chris Greig Reinhard Haas Anca D. Hansen Ove Hoegh‐Guldberg Hannele Holttinen Daniel Huertas‐Hernando Eduardo Ibanez

Shin‐ichi Inage Kenneth B. Karlsson Jonas Katz Ben Kaun Sila Kiliccote Kyung N. Kim Lena Kitzing Juha Kiviluoma Magnus Korpas Thilo Krause Joe L. Lane H. V. Larsen Warren Lasher Joohee Lee Chen‐Ching Liu Joao A. P. Lopes Peter D. Lund Andre G. Madureira Sergio M. Martinez Manuel Matos Nickie Menemenlis Eric McFarland Lutz Mez Michael Milligan Carlos Moreira P. E. Morthorst Carlo Obersteiner Daniel Olsen Mark O’Malley Antje Orths Dale Osborn Paul Parker Mary A. Piette Fabrizio Pilo Amalia Pizarro Zakir Rather Barry Rawn Tobi Reid

x  List of Contributors

Jenny Riesz Erkka Rinne Luis Rodrigues Serafin van Roon Ian H. Rowlands Lisa Ruttledge Diego Schmeda‐Lopez Jürgen Schmid Sascha T. Schröder Jeongseok Seo Naoki Shinohara Bernardo Silva Klaus Skytte Simon Smart Charles Smith Lennart Söder Michael D. Sohn Gian G. Soma

Poul Sorensen Benjamin K. Sovacool Hartmut Spliethoff Alexandru Stefanov Morten Stryg Job Taminiau John Tande Thomas Trotscher Aidan Tuohy Frans Van Hulle Marina González Vayá Ye Wang Simon Watson Jing Xie Karl A. Zach Robert Zavadil Shan Zhou

Preface

The global energy system confronts huge challenges in the coming decades. The present‐day fossil‐fuel‐ based energy production, which dominates the energy scenery, needs to be replaced by clean energy options to meet the climate change mitigation targets set in Paris in December 2015. At the same time, the demand for energy continues to grow, mainly due to growing prosperity in the less developed world. One of the main challenges will indeed be to secure a clean energy path to the future in the emerging economies, unlike the industrialized countries in the past. Although global carbon dioxide emissions have increased almost by half since the days when the first climate agreement, under the UN auspice, was established in the 1990s, positive news is starting to emerge. In recent years global CO2 emissions have been stabilized, but these now need to be sent on a declining trajectory. Much of the positive development can be contributed to the rapid market share of renewable energy sources, notably solar and wind power. The cost of these technologies is becoming competitive with their fossil counterparts. More importantly, future prospects for renewable energy technologies are bright: there still remains potential for major technology developments, efficiency improvements, and cost reductions, which together could make renewable energy the mainstream energy solution. Indeed, respected energy scenarios, for example those developed by the International Energy Agency (IEA), indicate that in the power (electricity) sector, which is of upmost importance with respect to emissions, a significant share of future generation capacity investments will be concentrated in solar and wind power by the middle of this century. We are already witnessing that these variable renewable electricity forms deliver a major share of the national electricity supply in some countries, such as Denmark, Germany, Ireland, Italy, and the United Kingdom. In the long‐

term, more countries are envisioned to satisfy their clean energy demands through renewable energy. Although renewable energy may play an important role across the entire energy system, it is particularly in the electricity sector where the new technologies will play a dominant role. Moreover, electricity demand is growing much faster than primary energy demand, due to electrification within our societies and everyday life, which stresses the role of electricity in the future energy system. Inherently, most new renewable power production technologies, such as solar and wind, but also marine power, do not rely on a supply of fuel, meaning that their instantaneous power production depends on the prevailing and time‐varying weather conditions. Thus, when transitioning to large‐ scale deployment of renewable energy from variable sources, a key challenge will be matching supply of power against demand, on a range of time scales from seconds to hours, days, and weeks. Large‐scale renewable electricity schemes in conjunction with existing energy systems can cause a range of different systemic issues, which need to be solved to make the best use of clean energy. Bridging the “new” and “old” energy is necessary, and both will need to coexist for some time, although the share from renewable sources will increase. An energy transition cannot be an on–off change, where we switch from old to new overnight! Integration of renewables into the energy system will thus be a critical issue, and of growing importance, in the coming years. We claim here that integration, in broad terms, will actually be the new hot topic in energy, if it is not already, which is not only linked to innovative technology solutions but which will also reshape energy markets, challenge existing business models for companies, and even integrate the consumer in a pivotal role. Energy system integration of renewable energy is a wide field which covers themes ranging from

xii Preface

modifying existing energy systems to better match renewables characteristics, introducing new flexibility measures and the evolution of energy‐limited technologies, exploiting communications and IT advancements within a smarter grid, and reforming markets, incentives, regulation and policy frameworks to obtain an operational, robust and economically viable energy resource portfolio. The energy transition ahead is a huge challenge to all market actors. No one will be left untouched: policy makers, energy planners, businesses, developers, academia, and even end users need to be re‐educated to understand the new rules of the game. This book aims to provide timely guidance on how to prepare for the turbulence, which rushes toward us, presenting implementable solutions and proposing successful pathways moving forward. This book addresses the key areas of large‐scale renewable energy integration and provides an authoritative overview on both the challenges and potential solutions. The book discusses the system challenges associated with renewables, grids, flexibility options, and markets, which encompass the central “integration” themes. The book is a collection of 32 authoritative contributions from specialists in the

r­ elevant disciplines. Through this collective push, our aim is to offer the reader a fresh and skillful insight to the multidisciplinary topic. The original inspiration for the book came through the publisher, when John Wiley & Sons established the Wiley Interdisciplinary Reviews: Energy and Environment journal, which mainly publishes review‐ type articles in energy and environment. Mr.  Tony Carwardine, now retired from Wiley, suggested ­ ­selecting collections of articles from the journal to create reference works in topical areas. The first book published in this manner was Advances in Bioenergy  –  The Sustainability Challenge (ISBN 9781118957875) in 2016. Advances in Energy Systems  –  The Large‐scale Renewable Energy Integration Challenge is the second book in this series. The editors wish to thank Sandra Grayson, Louis Manoharan, Adalfin Jayasingh, Shalisha Sukanya and Peter Mitchell from Wiley for their great help and assistance during the process of finalizing this book. Peter D. Lund John A. Byrne Reinhard Haas Damian Flynn

PART I

ENERGY SYSTEM CHALLENGES

1

Handling Renewable Energy Variability and Uncertainty in Power System Operation Ricardo Bessa, Carlos Moreira, Bernardo Silva and Manuel Matos INESC TEC, INESC Technology and Science (Formerly INESC Porto) and FEUP, Faculty of Engineering, University of Porto, Porto, Portugal

Concerns about global warming (greenhouse‐gas emissions), scarcity of fossil fuel reserves, and primary energy independence of regions or countries have led to a dramatic increase of renewable energy sources (RES) penetration in electric power systems, mainly wind and solar power. This has created new challenges associated with the variability and uncertainty of these sources. Handling these two characteristics is a key issue that includes technological, regulatory, and computational aspects. Advanced tools for handling RES maximize the resultant benefits and keep the reliability indices at the required level. Recent advances in forecasting and management algorithms provide a means to manage RES. Forecasts of renewable generation for the next hours/days play a crucial role in the management tools and protocols of the system operator. These forecasts are used as input for setting reserve requirements and performing the unit commitment (UC) and economic dispatch (ED) processes. Probabilistic forecasts are being included in management tools, enabling a move from deterministic to stochastic methods, which lead to robust solutions. On the technological side, advances to increase mid‐merit and base‐load generation flexibility should

be a priority. The use of storage devices to mitigate uncertainty and variability is particularly valuable for isolated power systems, whereas in interconnected systems, economic criteria might be a barrier to invest in new storage facilities. The possibility of sending active and reactive control set points to RES power plants offers more flexibility. Furthermore, the emergence of the smart grid concept and the increasing share of controllable loads contribute with flexibility to increase RES penetration levels. ­INTRODUCTION The integration of renewable energy sources (RES) in a generation portfolio conveys several benefits, such as a reduction in greenhouse gases emissions and in the country’s dependency on imported energy, and it decreases spot prices. However, generation from RES (i.e. wind, solar, hydro, wave, geothermal, and biomass) can be variable and uncertain, in contrast to conventional generation (e.g. coal thermal plants, combined and open cycle gas turbines). Nevertheless, many power systems have had hydropower for a long time in their portfolio, and system operators (SOs)

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

4  Advances in Energy Systems

already have appropriate procedures for its utilization regarding the need to manage its variability and uncertainty. Note that hydropower is more flexible than other RES (such as wind and solar), in particular power plants with a reservoir. The installation of pumped storage units also facilitates water management. Conversely, geothermal generation is invariable, which might create problems because it is incapable of following load variations. The variability of hydropower, biomass, and geothermal is more apparent on yearly and seasonal timescales (run‐of‐river hydropower can also present daily variability), whereas the variability of wind and solar covers all timescales (including daily, hourly, and minutes variability). At present, the penetration of wind and solar generation in many power systems has attained a high level, and this has created new challenges when operating the power system. In order to meet these challenges, the state‐of‐the‐art encompasses new technological and computational advances for dealing with the variability and uncertainty of RES, particularly regarding wind and solar generation, since hydro variability has for a long‐time been tackled in power systems. New forecasting and decision‐aid algorithms, including stochastic information, can improve the ability of a power system to cope with variable and uncertain generation coming from RES, without excessive extra operational cost while maintaining reliability standards. On the technological side, new technological advances to enhance the flexibility of conventional power plants (namely, base‐load and mid‐merit units) are essential. Primary frequency control provided by new RES power plants or the use of storage devices are also relevant research areas. This article describes developments in several interdisciplinary topics related with managing high penetrations of solar and wind, and points toward research trends for the next years. First, the challenges introduced by RES (in the remainder of the chapter only wind and solar are considered) in power system operation are discussed. Then, an overview of the advances in renewable energy forecasting is presented. Renewable energy forecasts are an important input to methods for setting reserve requirements, defining the commitment schedule and performing congestion detection, which are reviewed. Consideration is also given to the electricity market role and the value of storage devices for interconnected and isolated systems. On the technological side, the importance of flexibility (from conventional generators and storage units) in power system operation is described, and some challenges and technological solutions unrelated to resource variability are reviewed, and the capability of active and reactive power control is analyzed.

­ HE CHALLENGES OF RES IN POWER T SYSTEM OPERATION Main Challenges The intrinsic variability and uncertainty of RES ­create several challenges in power system operation and planning[1]. At every instant, generation must follow load variations in order to maintain the generation‐load balance. The variable nature of RES (e.g. rapid generation ramps) represents a challenge, in particular, for systems without hydropower, as it introduces variations in the generation side that can only be smoothed within the physical constraints of the conventional power plants (e.g. ramping up and down, minimum generation limits). In general, the available ramping rates of flexible generation units and fast‐starting units (e.g. hydropower) are used for accommodating this variability. Technological solutions such as control schemes for wind power active and reactive power set points smoothen the impact of variability. For example, a dispatch center for RES with the ability to control the active and reactive power output was created in Spain[2]. RES uncertainty also creates imbalances between generation and load as it is not possible to know (with certainty) the RES generation levels for the next hours/ days. These imbalances originating from forecast errors are handled with additional generation capacity (which is an ancillary service). Computational algorithms such as forecasting algorithms and large‐scale stochastic optimization (instead of deterministic tools/ rules) have been developed for including information about uncertainty in the decision‐making processes. The importance of new and advanced forecasting algorithms for RES, not only for the SO but also for wind power producers (in particular when trading wind power in the market), is shown by the proliferation of companies that sell this service[3]. Storage units can also play an important role in handling RES variability and uncertainty on different timescales. If these new solutions are not adopted, variability and uncertainty of RES could lead to situations with high operational cost. For instance, curtailment of renewable generation during low load periods and the startup of expensive fast‐starting units lead to a cost increase. Moreover, even with a perfect forecast for  the next hours/days, it is necessary to schedule flexible generation units for accommodating the generation ramps. Ela and O’Malley[4] presented a simulation framework for assessing the impact of wind power variability and uncertainty on several timescales. The results showed that the imbalance impacts increase with longer dispatch resolutions (ranging from five

Handling Renewable Energy Variability and Uncertainty in Power System Operation  5

Other Challenges and Technological Solutions Specific technological characteristics of RES conversion systems, which do not depend on resource variability, bring also new operational challenges as integration levels increase. There are many important ancillary services traditionally provided by conventional thermal or hydro‐based generation units, such as voltage and frequency control. Additionally, conventional generation units intrinsically provide inertia to the system, which is a fundamental characteristic in order to ensure its stability. The large‐scale integration of RES naturally displaces conventional generation units, thus strongly affecting ancillary services provision and global system security as a result of a general degradation of adequate frequency response[5]. This has led SOs to define very restrictive rules and conditions for allowing increasing RES integration, which are referred to as grid codes[6]. In general, existing grid codes require wind farms to withstand several disturbances and to support network stability through the provision of ancillary services similar to those provided by conventional synchronous units. Focussing on international grid code requirements for wind power integration, they can be generally organized in five main categories: (i) fault ride‐through (FRT) requirements, (ii) active and reactive power responses following disturbances, (iii) extended variation range for voltage‐frequency, (iv) active power control or frequency regulation support, and (v) reactive power control or voltage regulation capability. From a technological point of view, energy converter manufacturers have been actively responding to SO operational requests through the development of innovative conversion systems, where the controllability and flexibility of power electronic interfaces have been assuming an increasing role[7, 8].

The large‐scale penetration of non‐FRT‐compliant units may lead to significant amounts of generation tripping, thus compromising system security. In  order to overcome these difficulties, system operators defined specific requirements that all generators should fulfill in order to be connected to the grid[9]. These requirements are generally specified in a voltage‐time curve that defines the region in which generators are not allowed to trip. As an illustrative example, Figure  1.1 depicts a general shape of the voltage‐time curve where the points A−H define key time‐voltage values delineating the region in which wind generators should remain connected to the grid during low‐voltage periods. The values shown in ­Table 1.1 illustrate different parameterizations of the voltage‐time curve for different SOs, which reflect specific grid characteristics in terms of generation units and protection philosophies. More recently, interest in large photovoltaic (PV) plants has gaining additional attention. Therefore, specific control requirements for photovoltaic inverters regarding FRT capabilities and voltage ­control are beginning to become a research topic, which will be developed in the near future[10].

Voltage (%) at point of connection

minutes to one hour) and with installed wind power. Assessment of the uncertainty impacts lead to the following conclusions: the uncertainty impact increases with the forecast error, but it is not significant until the forecast error reaches a threshold; large forecast errors have a significant impact on the generation costs and branch congestion of day‐ahead scheduling, but not in the real‐time dispatch. The next subsection discusses challenges and solutions for aspects unrelated to variability and uncertainty of the resource (e.g. wind and solar). Nevertheless, situations with wind turbine tripping following voltage dips are a source of uncertainty and variability to the system, but they are essentially technological and not related with the natural resource.

A

F

H

E

G

Tripping allowed B C

D Time (s)

Figure 1.1  Generic FRT voltage versus time characteristic curve.

Table 1.1  FRT voltage and time values for European grid codes. Grid Code

BC

BD

AF

FE

AH

HG

Denmark Germany Ireland Spain Spain (Canary Islands) United Kingdom Portugal

25% 0% 15% 20% 0%

0.1 s 0.15 s 0.625 s 0.5 s 0.5 s

0.75 s 0.15 s 3 s 1 s 1 s

25% 30% 10% 20% 20%

10 s 0.7 s N.A. 15 s 15 s

N.A. 10% N.A. 5% 5%

1.2 s 1.5 s

20% 2.5 s 15% 20% 10 s N.A.

0% 0.14 s 20% 0.5 s

6  Advances in Energy Systems

The consequent displacement of synchronous generators affects power system inertia and primary ­frequency control capabilities, limiting further integration of RES. In order to overcome such limitations, the possibility of endowing wind energy converters with additional control loops in order to provide either inertia emulation capabilities and primary frequency regulation has been discussed[8, 11]. The general approach is to operate RES with a certain reserve margin through the appropriate use of deloading mechanisms. Thus, the reserve margin can be autonomously deployed in case of frequency deviations through the use of a power reference‐frequency droop control. Additionally, a supplementary control‐loop based on the rate of change of the frequency deviation enables RES to emulate an inertial response. Similar operational characteristics are being envisioned for large offshore wind farms connected to a mainland grid through multiterminal high‐voltage direct current grids[12]. Energy storage systems can also provide an important contribution to primary frequency control, namely, in island systems[13], but also in interconnected systems[14]. ­ADVANCES IN RENEWABLE ENERGY FORECASTING During the last 15 years, research work has been conducted on developing forecasting algorithms which achieve a forecast error reduction and an expansion of forecasting products[3]. Furthermore, the number of companies selling forecasting services for RES has proliferated, and the reliability and availability of their service has improved. Presently, SOs use forecasts in their daily operation, embedded in decision‐making processes[15]. Research on wave energy forecasting is also being conducted[16], although this technology is not at the maturity levels of wind and solar technologies. At a regional/national level, wind power forecasting (WPF) literature reports a normalized mean absolute error (NMAE) of around 6–10% and a normalized root mean square error (NRMSE) of around 8–12% of the installed capacity for 24 hours ahead, rising to 11–14% and 14–17% for 48 hours ahead[3]. For solar power forecasting (SPF), the NRMSE is around 4.3–4.9% for day‐ahead forecasts[17]. The load forecast error is generally measured as a mean absolute percentage error and its value is around 1–2% for 24‐hours ahead forecasts[18]. Note that the real impact of the RES forecast errors can only be measured by an increase in power system operating costs (e.g. increased use of reserves for balancing forecast errors).

Wind Power Forecasting In general, SOs only need WPFs for a horizon up to 3 days ahead. This time horizon can be divided into two classes[3]: (i) very short term, for a maximum of 6 hours ahead with different time steps (10, 15, 30, and 60 minutes), (ii) short term, for a maximum of 72 hours ahead in 30 and 60 minute time steps. The algorithms used to produce a WPF for these two horizons can differ in type and input data. Forecasting algorithms for very short‐term horizons use as input past values of the time series. Classic examples are models based on the autoregressive integrated moving average[19], but recently, regime‐ switching models[20], such as Markov‐switching autoregressive, are being used for capturing the ­ influence of some complex meteorological variables on the power fluctuations. Furthermore, the combination of dispersed meteorological on‐site and off‐site observations and mesocale numerical weather predictions (NWP)[21] can lead to significant improvements in very‐short‐term horizons[22]. Forecasting algorithms for a short‐term horizon require NWP data. Several statistical and machine learning algorithms have been applied for converting the forecasted meteorological variables into wind power generation forecasts[3]. The use of cost functions for fitting prediction models under non‐Gaussian errors[23], as well as the combination of different statistical and NWP models[24], can improve accuracy. In addition to research work seeking a forecast error reduction, different forecasting products have been developed: regional forecasts, uncertainty forecasts, and ramp forecasts. These forecasts are inputs to the power system management algorithms reviewed in the following sections. Regional (or aggregated) forecasts, resulting from a process called upscaling, consists of estimating the total wind power, using information from forecasts of representative wind farms, for which NWP and/ or online observations are available[25]. This process is justified when the observations from some wind farms are unavailable or the data quality is poor. From the literature, it is known that the aggregation of wind farms reduces the forecast error because of spatial smoothing effects[26]. A point forecast does not give any information about the errors associated with the forecasted value. This motivated the development of advanced physical and statistical methods for estimating and communicating wind power uncertainty. This uncertainty can have different representations: probabilistic, statistical, and meteorological‐based scenarios.

Handling Renewable Energy Variability and Uncertainty in Power System Operation  7

In the literature, a broad set of statistical methods for producing probabilistic WPF can be found, such as local quantile regression[27] or conditional kernel density estimators[28]. The probabilistic forecasts can be expressed as a set of quantiles or interval forecasts (as depicted in Figure  1.2a), or a probability density function (pdf), or moments from the distribution. However, the forecasts produced by these methods do not include the temporal and spatial dependency of the forecast errors. The dependency of the errors is valuable information for time‐dependent decision‐ making problems, such as storage management[29] or UC[30]. The state‐of‐the‐art in WPF represents this time dependency by scenarios that are time/spatial trajectories or random vectors (as depicted in Figure 1.2b). The scenarios can be statistical or physical based. Statistical‐based scenarios are generated with simulation techniques that sample random vectors using a dependency structure (e.g. covariance matrix) capturing the temporal/spatial dependencies of the forecast errors. Pinson et  al.[31] described a method for sampling random vectors from the forecasted marginal cumulative probability functions. The method uses a multivariate Gaussian distribution where the covariance matrix represents the temporal dependency between the forecast errors. This method can also be extended to include spatial dependency[32]. Physical‐based scenarios (or meteorological ensemble) capture two different sources of error: initial conditions and the model (i.e. representation of the dynamics and physics of the atmosphere). These scenarios are obtained with three approaches: (i) different initial conditions or numerical representations of the atmosphere are used in each run of the NWP system[33]; (ii) outputs of different NWP forecast models or from the same NWP model with different parameterizations; (iii) different forecasts made at different times with the same NWP model[34]. These NWP ensembles can be converted to wind power ­scenarios[35]. Recently proposed, a different type of scenario is associated with spatial fields[36], which are statistically transformed NWP points from a grid covering the wind farm to an equivalent value representing the surface roughness and terrain at a chosen reference point for the wind farm location. Compared with other uncertainty representations, NWP spatial fields capture phase errors in NWP and sampling errors[37]. Finally, although the scenarios already capture ramps and extreme weather events, another forecast product is ramp forecasting[38]. A ramp forecast provides information (point and probabilistic) about the magnitude and timing (i.e. phase error) of future wind  power rapid variations. Bossavy et  al.[38] used

a filtering/thresholding approach for detecting and forecasting ramps using, as input, forecasts produced with meteorological ensembles. The forecast gives the probability of observing a ramp within a set of time intervals. Ferreira et al.[39] presented a different methodology also based on a high‐pass filter and using statistical scenarios as inputs. Solar Power Forecasting Although research on WPF has reached a maturity level, research on SPF could be classified as “under development.” In contrast to WPF, SPF requires specific forecast models for the different technologies, such as concentrated photovoltaic and concentrated solar thermal. Nevertheless, current research is more concentrated on the weather forecasting side. Differently from WPF, pure statistical models with past observations can produce SPF with an acceptable quality for hours/day‐ahead horizons ­ with hourly time steps[40]. The main reason is that the serial dependency presents a strong daily and weekly seasonal pattern. In fact, the daily cycle of solar irradiation is rather easy to predict. Nevertheless, as ‘shown’ by Ahlstrom and Kankiewicz[41] and ­depicted in Figure 1.3, intrahourly forecasts present a high variability caused by clouds. In fact, clouds are the main cause of short‐term variation, and clouds of different types, speed, and size represent different impacts in solar power generation. Heinemann et  al.[42] identified different input data sources for the various time horizons of interest. For very short‐term horizons (between 30 minutes and 6 hours), the authors suggest the use of satellite‐based cloud motion vector fields. For short‐term horizons (up to 2 days ahead), the forecasts should be based on NWP data. Heinemann et al.[42] describe very short‐term forecasting algorithms based on cloud‐index images that are predicted with motion vector fields[43] derived from two consecutive images. For short‐term horizons, the authors forecast surface solar irradiance with a NWP mesoscale model. The results showed a good accuracy for clear sky situations, but with broken clouds and overcast conditions, the error increased significantly. A second approach described by the authors is to use a mesoscale NWP model to predict variables describing cloudiness, and this information is converted to solar irradiance. Wittmann et al.[44] examined the importance of combining NWP and aerosol‐based forecasts for day‐ahead horizons. According to the authors, aerosol information is of great relevance for clear sky situations, which are the most frequent situations at solar farm locations.

8  Advances in Energy Systems

(a) 1 90% int. 80% int. 70% int. 60% int. 50% int. 40% int. 30% int. 20% int. 10% int. Forecast Measured

Power (p.u.)

0.8

0.6

0.4

0.2

0 00:00

03:00

06:00

09:00

12:00

15:00

18:00

21:00

Time horizon (h)

(b) 1 Scenarios Forecast Measured

Power (p.u.)

0.8

0.6

0.4

0.2

0 00:00

03:00

06:00

09:00

12:00

15:00

18:00

21:00

Time horizon (h) Figure 1.2  (a) Forecast intervals, centered in the median, and limited by their lower and upper bounds, which are forecasted quantiles. (b) Twenty statistical‐based scenarios with temporal dependency of errors that respect the marginal distribution of the probabilistic forecasts (i.e. plot in (a)).

Handling Renewable Energy Variability and Uncertainty in Power System Operation  9

(a)

(b) 8

8 6 MW

MW

6 4 2

4 2

0

0 5

10

15

20

Hourly time steps

0

20

40

60

80

15-min time steps

Figure 1.3  Generation variability of solar power due to clouds: (a) hourly average; (b) 15 minute average.

Lorenz et  al.[45] showed that, as with wind farms, the aggregation of photovoltaic (PV) panels and farms decreases the forecast error, by smoothing the cloud effect. A forecast model for a regional forecast is also described[17]. Finally, note that most statistical‐based uncertainty forecasts techniques from WPF can be applied to this problem. Furthermore, meteorological ensembles also have a potential application in SPF. ­ HE IMPORTANCE OF GENERATION T FLEXIBILITY In the past, flexible power plants were mainly used for handling load variability and uncertainty. RES introduces variability and uncertainty in the supply side, and with a higher magnitude in a scenario with high RES penetration. The wide regional distribution of multiple RES power plants decreases the variability and uncertainty of RES‐based generation[1, 26]. The flexibility of the power plants can be defined as the ability to react (i.e. modify generation levels in response to a command from the SO) quickly enough to rapid generation/load ramps and to d­eviations ­between scheduled and realized values. Generation flexibility is characterized by several parameters, such as ramp rate (% of MW per minute), technical and economic minimum generation level, start‐up and shut‐down time. Lannoye et al.[46] propose a methodology to compute the Insufficient Ramping Resource Expectation index, which is a metric that measures power system flexibility for long‐term planning studies. The following separation between power plants can be made[47]:

• Peak units are able to start, shutdown, or change their generation level very quickly in response to a command from the SO. Units of this type can handle minute‐to‐minute variations and uncertainty, and some examples are open‐cycle gas turbines and hydropower plants. • Mid‐merit units have the ability to ramp up to their maximum power and down to their minimum, but are slower than the peak units. Examples of these units are combined cycle gas turbines (CCGT) and biomass, and they cover interhour variations and uncertainty. • Base‐load units have slower response times and normally operate across the entire day or during predefined periods. Units such as coal‐fired plants, nuclear, and geothermal might require around six hours to start or provide a flexible response. These units can provide some flexibility for daily variability. In addition to these three categories, there are also must‐run units that ensure the dynamic stability of the power system (e.g. minimum online inertia in the system). The minimum generation level of these units may be a critical factor during periods with high penetration of RES. In some cases, such as France and Germany, nuclear power plants can also provide load‐following capabilities[48]. However, this has economic costs, mainly related with the load factor, as it is only profitable if operating at high load factors (i.e. above 90% of rated power). Furthermore, coal‐ and gas‐fired power plants are becoming more flexible with recent technology. For instance, in Germany, older coal power plants have ramp rates of 2%/minute, whereas new plants have rates of 4–6%/minute[49].

10  Advances in Energy Systems

A critical limitation is the minimum load limit of conventional generation units (both technical and economic), which, in general, is around 40–50% of the rated power. Nicolosi[50] reported a lower flexibility in the German power system for the period between October 2008 and December 2009. In situations with generation surplus, all the generation sides together were unable to reduce the generation below 46%, in particular the base‐load technology, which showed high minimum generation levels (due to technical and economic constraints). Soder et al.[51] reported that the Portuguese power system was able to accommodate a very high wind power penetration level (maximum of 81% at 6:45 a.m.) by stopping several hydropower plants between 0 and 4 a.m., and only one gas‐fired plant was kept in operation at 25% of rated power. The flexibility of hydropower units and its pump storage was used to balance wind power fluctuations, and the interconnection with Spain allowed exporting generation surplus. The cross‐border interconnections also play an important role in increasing the flexibility. It facilitates the use of more flexible power plants where they are needed to balance RES. For example, a high penetration of wind power generation is feasible in Denmark because of the good interconnections with Sweden, Norway, and Germany. The interconnection with Norway provides access to its fast starting hydropower plants, which can be used as reserve capacity contracted in the NordPool electricity market[52]. However, a higher transfer capacity between countries may also create overload problems in situations with high RES‐based generation in one control area[53]. Therefore, cross‐border interconnections should be combined with suitable operational procedures (e.g. forecasting, stochastic power flows) and investment in flexible AC transmission system (FACTS). The SO can use this flexible generation as reserve capacity, and methods reviewed in the following section optimize the use of this available flexibility.

­METHODS FOR HANDLING THE VARIABILITY AND UNCERTAINTY FOR STEADY‐STATE OPERATION In the operational domain, the SO has a limited time for conducting studies and examining alternatives. The uncertainty in this domain is relatively small when compared with the planning domain, which leads the SO to analyze the current conditions by a series of deterministic methods or rules. The main problem with deterministic approaches is that the user does not know the level of risk associated with

future operating conditions. This normally leads to a conservative operation with high operating costs or to unanticipated high risk during operations. As the penetration of renewable generation increases, new management procedures and algorithms emerge for handling generation variability and uncertainty. The use of uncertainty forecasts for renewable generation becomes a relevant input for decision‐aid algorithms. It is important to stress that modeling forecast uncertainty in decision‐making algorithms with the average historical forecast errors (a priori estimation) is not the same as using uncertainty forecasts (a posteriori estimation) directly in the methods, and might lead to conservative and expensive operating strategies. This section presents an overview of new management algorithms for supporting power ­system operation in steady‐state conditions with a significant penetration of renewable generation. The majority of the methods in the literature only consider the presence of wind power in the system. Nevertheless, a generalization to other RES (such as solar) is in general possible and straightforward if forecasts for solar power generation are available. Methods for Setting the Reserve Requirements in the Operational Domain In each time instant, the SO is responsible for maintaining the balance between generation and load in the power system. This leads to the need for a reserve capacity composed of loads and generation units able to respond to possible problems. Upward and downward reserves are normally necessary. The upward reserve consists of generation units (or loads) online or offline able to, in a short period, increase their generation levels (or decrease their consumption levels). The downward reserve consists of online generation units able to decrease their generation levels or loads able to start consuming (or increase consumption) in a short period. The classical categories of reserves[54] are: primary (or frequency response), secondary (or regulation reserve), and tertiary (or replacement reserve). Primary reserve is used for stabilizing the system frequency at a stationary value after a disturbance or incident in the time frame of seconds. Secondary reserve is dispatched via automatic generation control for restoring the frequency and interarea power exchanges to their nominal values in a time frame of 15 minutes. The SO restores or supplements secondary reserve levels by manually activating tertiary reserve, in a time frame between 15 minutes and 1 hour. High penetration levels of renewable energy (mainly wind and solar) in the system require a revision of these

Handling Renewable Energy Variability and Uncertainty in Power System Operation  11

reserve categories. Large shares of renewable generation do not create problems (discounting the sudden disconnection due to voltage dips), in interconnected systems, for the timescale of primary reserve[55]. Holttinen et al.[56] divided the reserve categories in: nonevent (normal operation) for variability and forecast errors inside the scheduling period; fast event (contingency operation) for an unplanned outage of a generator or cross‐border transmission line; slow event for expected net‐load ramps; and forecast errors that can occur on longer timescales (tens of minutes to hours). The authors argue that generation from RES does not change fast enough to be a contingency event, and the impact on secondary reserve is lower than the impact on tertiary reserve (both reserves are included in the nonevent category). High penetrations of RES will introduce slow events characterized by high rapid generation ramps and forecast errors. On the timescale of tertiary reserve, several countries have already created a reserve category for slow events; for example, the manual reserve in Denmark and the balancing reserve in Spain and Hydro‐Quebec. Deterministic methods (or rules) for setting the reserve requirements have served many SOs in the past[57], mainly because they are easier to be understood and applied by operators, and continued to high security levels with minimum study effort. However, with high levels of renewable generation, excessively conservative approaches would result in a high reserve cost and to a waste of renewable generation‐ related benefits. For example, operators not comfortable with the possibility of high forecast errors might define high levels of reserve to compensate potential deviations, thus reducing the economic attractiveness of RES. On the other hand, because risk is not really monitored in deterministic approaches, conservative rules may sometimes fail in specific circumstances not included in their original rationale. The alternative consists of using probabilistic methods for setting the reserve requirements. In fact, the use of probabilistic methods for setting the reserve in the operating domain is not new. A classic example is the Pennsylvania–New Jersey–Maryland (PJM) interconnection method that evaluates the risk of the committed generation units considering generation outages and load forecast errors[58]. The First Step: Hybrid of Probabilistic–Deterministic Rules Presently, SOs are starting to include information about forecast errors and uncertainty when defining the reserve requirements. The first step consisted of including a probabilistic component in the former

deterministic rules. For example, Electricity Reliability Council of Texas (ERCOT) (Texas Independent SO) defines the nonspinning reserve (the one that handles forecast errors) as the 95th percentile of the observed hourly net load error from the previous 30 days, plus the size of the largest unit[59]. The Spanish SO (REE – Red Eléctrica de España) defines the balancing reserve requirements as the sum of the generation shortage/surplus due to load and wind generation historical forecast errors, and unplanned outages[60]. The European Network of Transmission System Operators for Electricity (ENTSO‐E) revised the rules from the former Union for the Coordination of Transmission of Electricity (UCTE), and a new probabilistic criterion for setting the total reserve (secondary and tertiary) was included[54], but without explicitly mentioning renewable generation. The Next Step: Probabilistic Methods Gouveia and Matos[61] extended the PJM method by including the wind power uncertainty modeled with a Markov model. The main limitation (common to other methods) is that the model does not include forecasting information, but only the a priori probability distribution of a set of wind power levels. Thus, the next step is to use probabilistic tools with uncertainty forecasts as input. The first set of methods assumed that the WPF error follows a normal distribution. Several authors[62, 63] compute the level of imbalance between load and generation by summing the variances of the load and WPF errors (σ2LW  =  σ2L + σ2W), assuming that both random variables are normal and independent. The reserve requirement needed to deal with the estimated imbalances is given by the variation contained within three standard deviations (3·σLW) of the overall system imbalance. Doherty and O’Malley[64] presented an analytical method for defining the reserve requirements. The model includes generators’ unplanned outages (both full and partial outages), load, and WPF errors modeled by a normal distribution. Ortega‐Vazquez and Kirschen[65] described a method for minimizing the sum of expected cost of energy not served (i.e. expected energy not served [EENS] multiplied by the value of lost load) and the operating cost. The wind power and load uncertainties are normal distributions and combined by summing their variances. However, the normal distribution assumption is highly questionable because the WPF errors’ distribution presents high skewness[66] and kurtosis[67], even when several wind farms are aggregated[25]. This assumption, when setting the reserve requirements, may result in a underestimation of the risk[68].

12  Advances in Energy Systems

Excluding this normality assumption, Matos and Bessa[69] proposed a probabilistic method that takes as input a probabilistic WPF. The load probabilistic forecast and generation uncertainties (i.e. unplanned outages of conventional generation units and wind turbines plus WPF probabilistic forecast) are convolved with the fast Fourier transform. From the convolution result, risk attributes related with a generation shortage (upward reserve) and surplus (downward reserve) are computed for each reserve level. The results of this exercise are risk/reserve and risk/cost curves for each lead time. As an example, Figure 1.4 depicts the EENS (one of the possible reliability indices) against the reserve level. Establishing this relation allows the SO to make a decision with different methods: (i) set a reference value for the risk and get the reserve level directly from the curve; (ii) establish a trade‐off between reserve cost and risk and find the corresponding optimal reserve level; and (iii) build a value function to find the best compromise between cost and risk. Bessa et al.[70], in the framework of the European project ANEMOS.plus, reported results from an operational demonstration of this method for the Portuguese SO. The demonstration showed that the probabilistic method outperformed the deterministic rules in use at that time. Menemenlis et  al.[71] describe a method based on convolution which also constructs risk/reserve curves for each lead time. The wind power is

­ odeled with gamma functions fitted with time‐ m varying parameters for each class of generation level. The decision method consists of establishing a reference value for the risk. The main limitation of the method is that the gamma functions for each wind farm are combined, assuming independence of the uncertainties. This assumption does not consider the spatial dependency of uncertainties, which cannot be neglected. Maurer et al.[72] also proposed a method based on convolution, but the authors did not provide details about how the uncertainties are estimated. An interesting contribution is a two‐step approach for separating the total reserve in secondary and tertiary reserve. The use of ensembles of wind power for setting the reserve was proposed by Pahlow et  al.[73] The authors, using ensembles from the multischeme ensemble prediction system[74], compared several rules, such as reserve equal to the difference between the minimum and maximum of the ensemble in each hour. The results showed that the rules based on the ensemble are cost efficient and cover more hours when compared with the deterministic rule (i.e. 11% of installed capacity). The reviewed methods can be used for purchasing reserve services in a sequential electricity market (i.e. reserve market after the energy market) or included in the UC market clearing process. This second possibility is reviewed in the UC section.

350

Conv. + load Unc. Conv. + load + wind Unc.

300

EENS (MWh)

250 200 150 100 50 0 0

500

1000

1500

Reserve (MW) Figure 1.4  Example of the expected energy not supplied as a function of the reserve level for a look‐ahead time and taking into account different sources of uncertainty: conventional and load uncertainty; conventional, load and wind power uncertainty.

Handling Renewable Energy Variability and Uncertainty in Power System Operation  13

Reserve for Extreme Events The additional value/information of ramp forecasting in decision‐making problems (e.g. UC) is currently under discussion. Some authors believe that information about ramps should mainly be used as supporting information for the operators, and not as a direct input to an optimization problem (e.g. UC)[39, 75]. A ramp forecast tool can be used to generate alarms to the operator and quantify the risk, supporting preventive measures against rapid variations of wind power. A particular type of ramp occurs in situations with high wind speed. The wind turbines have a cut‐out wind speed value (generally around 25 m s−1), and above that value the turbine starts to trip. In this case, the SO must deal with an unforeseen rapid drop of wind generation that must be forecasted in order to take preventive measures. Figure  1.5 depicts a wind power reduction of around 600 MW in 1400 MW telemetered by the SO of Portugal, and originated by the cyclone Klaus on 23 January 2009. The forecasting system did not predict this extreme event, and the incident was handled with fast starting hydro units and importing energy from Spain. With solar power generation, an equivalent event can be triggered by fast‐moving clouds. Lin et  al.[76] describe a model for estimating the wind power generation reduction under extreme wind conditions. A high‐resolution tool, with calculations in the frequency domain and considering the spatial distribution of the wind farms, simulates power reduction trajectories in a minute‐to‐minute time resolution. From the simulation results, a reserve

requirements curve informing how much reserve is required for different wind speed values (close to cut‐ out speed) is constructed. In the operational domain, the tool combined with NWP enables the calculation of reduction trajectories samples and a corresponding reserve requirements curve. The main advantage of the described model is that it does not require extensive historical data of the wind farms in the region under analysis. UC and ED with RES The UC problem consists of scheduling generation units to minimize the cost of supplying the forecasted load with a set of constraints related to power system security and operation (e.g. ramping rates, start‐up times). The economic dispatch (ED) is based on the UC generation schedule and computes the generation levels of each unit that lead to the lowest possible cost. In summary, the UC decides which units should be running (e.g. on or off) and the ED determines the generation levels of the committed (or online) units. It is common to find these two processes in noninterconnected power systems and in the United States as a market clearing instrument. The UC and ED, in the United States, are used in the following steps[15]: in the day‐ahead market clearing, a security‐constrained UC and ED is used to schedule and dispatch the generation units, using bids (price and quantity pairs in a stepwise increasing function) from market agents; after the day‐ahead market ­ clearing, a reliability assessment commitment (RAC) is performed using security‐constrained

1400 1200

MW

1000 800 600 400 200

Generated power Forecast generation Available power

0 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 Hour Figure 1.5  Wind power reduction in Portugal due to extreme wind speed conditions originated by the cyclone Klaus in January 2009 (the wind power generation values are only from telemetered wind farms, available in real time, and not the total wind power generation in Portugal). Source: Reproduced with permission from Ref. 131. Copyright 2009, REN.

14  Advances in Energy Systems

UC; a  ­security‐constrained ED is performed in the real‐time market to dispatch the units scheduled by the day‐ahead market and RAC. Presently, SOs only use deterministic UC/ED, and the load and conventional generation uncertainty is covered in a constraint for guaranteeing a fixed amount of reserve. The state‐of‐the‐art for solving the UC problem is mixed‐integer programming (MIP), which is available in most commercial optimization packages. New UC/ED policies that include probabilistic information can be promising alternatives with increasing levels of RES. For example, ERCOT, when making the day‐ahead plan of the generation resources, uses as a WPF the 20% quantile (i.e. 80% probability of surplus)[77]. This can be understood as a conservative approach that guarantees a safety margin, but increases the cost due to the overcommitment of conventional generation units and leads to a greater need for downward reserve. Thus, an alternative approach is ­stochastic UC/ED (SUC/SED). The state‐of‐the‐art for the SUC and SED is a two‐stage stochastic programming with recourse[78]. The UC decisions, excluding fast‐starting generation units, are taken in the first stage (modeled as here‐ and‐now variables). The dispatch decisions are made in stage 2, modeled as wait‐and‐see variables, when realized values are known (i.e. real‐time). Generally, this stochastic problem is converted into a deterministic equivalent that is solved with MIP. The objective function in the published formulations is the minimization of the expected cost[30, 78]. However, subjects such as uncertainty modeling, strategies for including the reserve requirements, and whether or not transmission network constraints are included differ in the literature. Uncertainty Modeling Uncertainty is generally modeled through a set of scenarios. However, the characteristics of these scenarios differ from author to author. Some scenarios are generated with techniques that are only suitable for planning studies[79]. For the operational domain, several authors generated scenarios assuming that the wind power follows a normal distribution and without considering the temporal dependency of errors[80–83]. Further UC formulations use different principles. Constantinescu et  al.[84] describe a computational framework that combines a NWP model with a sampling technique for generating physical‐based scenarios able to capture the spatial–temporal evolution of forecast errors. Wang et  al.[30] and Zhou et  al.[85] modeled wind power uncertainty with  the method from Pinson et  al.[31] Sturt and Strbac[86]

use a quantile‐based scenario tree for wind power ­uncertainty. The scenario tree is constructed from a user‐defined topology, and forecast errors at each node are determined from quantiles of the forecast error distribution. The use of a large set of scenarios increases the computational requirements. A common practice is to reduce the number of scenarios with a technique that aggregates similar scenarios and eliminates scenarios low probability. The most‐used techniques are backward reduction and forward selection based on a family of Kantorovich or transportation probability metrics[87]. Pappala et  al.[82] described a scenario reduction method based on particle swarm optimization, and Sumaili et al.[88] described another algorithm based on clustering techniques. Modeling the Reserve Requirements In the SUC, uncertainty is implicitly modeled by using a set of scenarios. According to Bouffard and Galiana[81], the reserves in the SUC are defined internally and there is no need for specifying a priori a reserve requirement. The quantile‐based scenario tree proposed by Sturt and Strbac[86] for modeling wind power uncertainty avoids the need to consider additional reserves for the SUC. Because the scenarios capture the worst‐case tail, it allows the SUC to model dynamic levels of reserve that weigh the cost of providing them compared with the load shedding cost or the cost from committing expensive generation. In the SUC of Morales et  al.[89], the reserve requirements are determined considering the value of lost load, electrical energy cost and reserve bids, and also a set of scenarios. However, other authors also studied the inclusion of a constraint related with reserve requirements, for modeling uncertainty explicitly. Ruiz et  al.[78], although not considering RES uncertainty, showed that the inclusion of a suitable amount of reserve requirements in the SUC formulation resulted in better performance in terms of total cost and unserved energy. Restrepo and Galiana[90] proposed a deterministic UC that includes a constraint that limits the probability of the forecasted residual load exceeding the schedule upward and downward reserve. Wang et  al.[30] showed that an SUC with additional reserve for covering the impossibility of comprising all events in a limited set of scenarios performs better than an SUC without additional reserve. Pappala et al.[82] in addition to a set of scenarios, considered additional spinning reserve for dealing with wind power ramps between two consecutive periods.

Handling Renewable Energy Variability and Uncertainty in Power System Operation  15

Stochastic Versus Deterministic UC/ED Several authors conducted studies comparing deterministic UC and SUC. Tuohy et  al.[79] compared both approaches and the SUC decreased the total cost around 0.25%. Interesting differences were: the deterministic UC commits more expensive mid‐merit gas and peaking units; the SUC imports more energy because of the scenarios with low wind power generation; deterministic UC increases the number of startups in the generation units; the SUC presents a better performance in meeting the spinning and replacement reserve requirements. Wang et al.[30] compared different deterministic and SUC strategies for dealing with wind power uncertainty. Results for a three‐month period showed that a deterministic UC without any WPF (i.e. the wind generation is assumed to be zero) resulted in an overcommitment of conventional units, and when only point forecasts for wind power are included, it leads to the highest cost and load curtailment. The SUC strategies have a relatively low cost, but a deterministic UC with a reserve rule (point forecast minus 10% quantile and 5% of the load) obtained a similar cost. Zhou et al.[85] extended this comparison to the RAC, where fast‐starting units are committed or decommitted by using very‐short‐term wind power probabilistic forecasts and scenarios in the SUC. Both articles show that a dynamic rule for setting the reserve combined with deterministic UC presents a better performance than a fixed reserve value and can yield similar results to SUC, without the need to change considerably the current operational procedures or increasing the computational effort. On other hand, the SUC, as it uses scenarios with temporal correlation, can capture with more detail the ramps between hours and, consequently, compute a suitable reserve for dealing with temporal variability. Another advantage is that it includes the uncertainty impact in the objective function. Papavasiliou et  al.[91] compared an SUC that uses scenarios with a deterministic UC with two alternative rules for setting the reserve: 20% of the forecast peak load, 3% of the forecasted load plus 5% of the forecasted wind power. The SUC improves the daily operational cost over the deterministic UC: 0.39% for the first rule, and 0.54% for the second rule with 7.1% of wind power penetration; 1.33% and 1.09% correspondingly with 14% of wind power penetration. Lowery and O’Malley[92] studied the impact of WPF errors statistics (i.e. variance, skewness, and kurtosis) on the UC and system operation, using a test system a portfolio from the all Island Grid study (Ireland and Northern Ireland). The authors evaluated which

statistical properties of the error distribution contribute to the system operation performance. The results showed that variance has the most impact on the total cost, and skewness only decreases the system cost if complemented by kurtosis. For instance, skewness combined with variance results in an overcommitment of expensive base‐load units due to an incorrect estimation of the tail probabilities. The introduction of these error statistics in the UC changes the generation levels of flexible mid‐merit and peak units, while base‐ load units are almost stable. Furthermore, the number of unit startup increases with the forecast uncertainty, with the exception of coal‐based power plants. The authors also evaluated the impact of over‐ and underestimation of these statistics, and several conclusions were obtained. For example, inaccuracy in kurtosis will aggravate the effects (i.e. system cost) of an underestimation of variance because an overestimation of kurtosis increases the dependency from flexible units, and underestimation overcommits base‐load units. The overestimation of kurtosis when variance is accurate results in a decrease of the total cost because it reduces the use of slower generation units. The lowest number of require startups when variance is underestimated is when skewness is accurate. If variance is overestimated, inaccurate skewness reduce the number of startups of flexible units due to an incorrect estimation of the expected wind power. Constrained SUC/SED The constrained UC/ED is particularly important for countries (e.g. United States) with nodal prices. When the transmission network constraints are included, new optimization techniques are needed for solving the problem. Wu et  al.[93] use Lagrangian relaxation to decompose the optimization problem in tractable deterministic constrained SUC subproblems that can be solved with MIP. Wang et  al.[80] divided a constrained SUC in a master problem according to the wind power scenario. The security of an initial dispatch is checked in a subproblem, and a redispatch is made for mitigating violations. If the redispatch is insufficient, Benders decomposition is used to revise the generation ­commitment of the master problem. Morales et  al.[89] present a network‐constrained market‐clearing mechanism for energy and reserve (spinning and nonspinning) optimization, considering wind power uncertainty. A two‐stage stochastic programming problem is proposed. The first optimization stage comprises the constraints and rules of the electricity market, whereas the second stage represents the power system operation and physical limitations.

16  Advances in Energy Systems

Managing Network Congestion A “weak” transmission network with a high penetration of RES is prone to congestion situations, which limits the benefits arising from RES integration. For instance, network bottlenecks bind the reduction in operational costs promoted by wind power and increase the amount of curtailed wind generation during low load periods[77, 89]. The use of power flow tools capable of including renewable energy uncertainty can help detect and manage situations with congestion, avoiding curtailment of renewable energy, and in the medium term can defer network investments. Two approaches can be found for including uncertainty in power flow calculations: probabilistic power flow (PPF) and fuzzy power flow (FPF). Both methods calculate, under steady‐state conditions, the probability distributions or fuzzy values (possibility distributions) for voltages (angle and magnitude), active and reactive power flows, and active losses. In PPF studies, wind speed uncertainty is frequently modeled by a Weibull distribution[94]. However, this modeling assumption is not suitable for the operational domain because the goal is to include forecast uncertainty. Moreover, the spatial dependency between the random variables is also important for power flow calculations. The first PPF considering WPF uncertainty was developed by Hatziargyriou et al.[95], for radial distribution networks. The main limitations are the assumption of a normal distribution for the WPF error, and dependencies are neglected. Morales et  al.[96] presented a PPF that includes uncertainty through the statistical moments (i.e. mean, variance, skewness, and kurtosis) of the distribution and also includes the spatial dependency of uncertainties. The authors proposed an analytical method based on a modified point‐estimate method (PEM). The dependent input random variables are transformed into independent variables, using an orthogonal transformation. Furthermore, the authors also described a Monte Carlo method for solving the PPF through simulation, using spatial dependent scenarios. The comparison between the two methods showed that the Monte Carlo method takes about 350 times longer than PEM, and with a 95% confidence level, the solutions provided by both methods are the same. Furthermore, it was shown that the dependency between WPF affects more the active power flow, and an increase in the correlation coefficient results in an increase of the power flow standard deviation. Usaola[97] describes an analytical method for PPF based on the cumulant method and the Cornish‐Fisher expansion series. A small number of convolution

o­ perations are conducted for multimodal distributions. The method accepts dependent/independent, continuous/discrete random variables. WPF uncertainty is modeled with beta functions conditioned to the level of injected power, and dependent scenarios are generated with a sampling process similar to others in the literature[31]. The author compared the proposed method with a PEM with independent random variables. The results showed that the PEM attains a good performance (i.e. average error between moments of the power flows) for the independent case up to the second‐order moment, whereas the proposed method behaves better in the dependent case. Bessa et  al.[70] adapted the classical FPF for including WPF uncertainty. A set of forecasted quantiles is converted into a (triangular) fuzzy number. The FPF, in the framework of the ANEMOS. plus project, was operationally demonstrated for the Portuguese SO, and the goal was to detect possible congestions and voltage violations in the market dispatch for the next day. The demonstration showed that in hours where wind power was underforecasted, leading to branch congestion, the deterministic power forecast (PF) was unable to detect congestion, whereas the FPF, depending on the selected cutoff, was capable of detecting possible congestion. Evaluation analysis for the whole demonstration period (six months) showed that deterministic PF has the lowest percentage value of false alarms, but at the same time, has the highest percentage of overlooked congestions. FPF, with a cutoff properly defined by the user, offers a more balanced solution between the percentage of false alarms and the percentage of overlooked congestions. All reviewed methods were applied only for detecting congestion situations under normal steady‐state operation. Some SOs also run power flow calculations with branch contingencies (N‐1 regime) as the ENTSO‐E Policy 3 suggests[98]. Moreover, the methods do not define (at least directly) any preventive/ corrective action or perform a redispatch for solving the ­problem. Vlachogiannis[99] proposed a constrained PPF that includes uncertainties from both wind power and electric vehicles (EVs). The method uses a learning algorithm based on learning automata systems for determining the values of continuous and discrete control variables. The goal is to determine robust values for the control variables that maintain the risk of constraints’ violation below a predefined level. This method has the following limitations: WPF are not included (i.e. wind speed uncertainty is modeled with a Weibull distribution) and spatial and temporal dependencies are neglected.

Handling Renewable Energy Variability and Uncertainty in Power System Operation  17

An alternative approach is an optimal power flow (OPF). Jabr and Pal[100] described an OPF algorithm that minimizes the generation costs, including a cost term for wind power spilling and another for using reserves due to a wind power shortage. Wind power uncertainty is represented by a probability mass function. However, the main limitation of the method is that uncertainty is only included in the objective function. Capitanescu et al.[101] describes an optimization algorithm for day‐ ahead steady‐state security assessment. The algorithm based on power injection scenarios defines preventive and corrective control actions that guarantee power system security for any possible contingency. The algorithm is divided into three stages: day‐ahead decisions; preventive control actions; and correction post‐contingency actions. These stages involve OPF and security constrained OPF. The authors did not consider RES uncertainty explicitly. However, the model can be extended to include this uncertainty source (e.g. using physical or statistical scenarios). The method also presents a high computational time, but parallel computing and scenario reduction techniques can improve the performance.

­THE ROLE OF STORAGE DEVICES Interconnected Power Systems Storage units such as pumped hydro storage (PHS) or compressed air energy storage (CAES) can play an important role in handling RES variability and uncertainty[102]. Hedegaard and Meibom[103] examined the application of different storage technologies for balancing power at different timescales. The authors analyzed several technologies (i.e. lead‐acid batteries, flow batteries, EVs, CAES, PHS, electrolysis combined with fuel cells) and concluded that all are suitable for intrahour and intraday/day‐ahead balancing. CAES, PHS, and electrolysis combined with fuel cells showed potential for several days ahead and seasonal horizons. Two different business models can be envisioned for exploring the economic and technical potential of storage devices. The first model consists of having an electrical utility aggregating small or large storage facilities. This utility can combine storage with RES for smoothing forecast errors, and/or price arbitrage (i.e. store in low‐price hours and sell during high‐ price hours), and/or participate in the ancillary services market. In the second business model, storage is an asset of the distribution or transmission SO. In this model, storage can be used for balancing services, solving network congestions, voltage support (in the distribution grid), and decreasing operational costs.

In the literature, it is common to find optimization algorithms for coordinating storage units with RES plants[29]. Long‐term evaluations of different storage technologies[104, 105], considering capital costs, show that RES combined with storage can lead to a negative net present value, even when selling reserve services. This economic assessment depends on the price differential between peak and valley hours and reserve prices. Economics are also important for assessing the value of storage units operated by a SO. Nyamdash et  al.[105] show that storage offsets the cost introduced by wind power, as it reduces the participation of mid‐merit units and increases the participation of base‐load units. Nevertheless, this conclusion is only valid for small storage power rating (200–600 MW), as large storage power rating (1200–1800 MW) can shift the load peak to the night period. The authors also evaluated the impact on the net load under different operational strategies. A strategy called mid‐merit (i.e. 12 hours of night charge and 12 hours of day discharge) leads to less wind power curtailment and decreases the net load variability. Meibom et al.[106] tested an SUC in the Irish power system and showed that there is no general change in the daily operation of pumped hydro units with the increased wind power capacity in the system. This means that storage is mainly driven by load variation. Tuohy and O’Malley[107], using a similar SUC model, studied the operation of the Irish power system with high penetrations of wind power and PHS. The results show that the main advantage offered by PHS is to decrease wind power curtailment, but only at high wind power installed levels. It is also shown that storage provides flexibility for dealing with WPF uncertainty, and the need for storage decreases with the forecast error. Economic analysis (from the SO perspective) also lead to conclusions similar to the abovementioned authors. The main economic savings from storage are from avoiding wind curtailment, but this is only significant for high wind penetration levels (e.g. between 42% and 55%). The authors also show that with increasing wind power, the interconnections are used more for ­exporting energy, decreasing the need for storage. Connolly et al.[108] make a similar study and reach similar conclusions, considering the complete energy system (e.g. electricity, transport, and heat) using the EnergyPLAN tool, but without simulating with detail power system operation on an hourly resolution. Isolated Power Systems Fast output power variability occurring in timescales less than a minute are likely to induce significant network frequency variations in isolated power

18  Advances in Energy Systems

systems. The mitigation of such unwanted effects that may lead to system instability can be achieved through the use of energy storage systems that are able to provide high cycling capabilities and high ramp rates as a result of the continuous operation that is required for effective and fast power modulation. Short‐timescale energy storage solutions based on flywheels, supercapacitors, or superconducting magnetic energy storage systems (SMES) are the preeminent technologies for this specific application[55, 102]. The use of supercapacitors in the DC link of the back‐to‐back AC/DC/AC converter of individual doubly fed induction generators (DFIG) in a wind farm is proposed by Qu and Qiao[109] as a way to mitigate global wind farm active power fluctuations resulting from wind variation in very short timescale. Similarly, the use of flywheel storage systems to mitigate wind power fluctuations is presented and discussed by Cimuca et al.[110] The combined use of different energy storage technologies such as flywheels, supercapacitors, or batteries in hybrid systems with offshore wind generation, diesel, and photovoltaic generation is presented by Ray et al.[111] Nomura et al.[112] proposed the use of SMES to smooth the power fluctuations of a 100 MW wind farm. The SMES system is used in a wind farm that is interconnected with a grid through a back‐to‐back DC link. The authors propose a strategy that enables the use of the stored energy in the SMES in order to compensate for the inertia of the turbines, allowing fast wind turbine speed control, ­depending on the wind condition. In small islands with low grid inertia, the installation of energy storage solutions can significantly increase system security and planning flexibility for renewable integration, while providing adequate regulation capabilities to fulfill load‐frequency control requirements. The general operation philosophy of an energy storage system in an isolated system is to provide a continuous balance between demand and supply, thus absorbing the excess of power or injecting power into the system. As a result, a significant improvement in global system stability can be achieved, contributing to higher RES penetration levels in isolated power systems. The main objectives of the study consider optimal sizing and management of the batteries, aiming to obtain the maximum economic benefit while providing adequate spinning reserves. Oudalov et al.[113] formulate a numerical optimization problem in order to optimize the economical profitability, taking into account investment and operational costs and the difference between the revenues from availability of frequency control reserve and payments resulting from selling excess energy on the spot market. Lee

and Wang[114] developed specific strategies in order to coordinate the control of different energy storage technologies such as batteries and flywheels in a hybrid power system containing wind and solar power generation in ­order to maintain system frequency in a very tight control band. ­ACTIVE AND REACTIVE POWER CONTROL OF RES As previously mentioned, the increasing integration levels of RES into electric power systems has been inducing SOs to request online control requirements over active and reactive power generation levels in order to maintain system operation within acceptable security levels. The main objective is to assure that RES generators provide additional support for global system operation, providing services for reactive power/voltage control as well as active power control, enabling an appropriate dispatch in case of severe situations such as line overloading. It is important to notice that the flexibility of DFIG and variable speed synchronous generators equipped with a full converter allows implementation of specific controllers for active and reactive power control at the turbine level, in coordination with the wind farm supervisory control system. Delfino et al.[107] have proposed an advanced control scheme enabling photovoltaic units to independently control their active and reactive power injection to the grid. The effectiveness of the control algorithm has been tested through simulation. The effects of reactive power control on a medium‐voltage (MV) grid have been highlighted, showing benefits in the voltage profile. Turitsyn et al.[108] have investigated the opportunity to control reactive power injection from a PV inverter to promote voltage regulation under stressed transient conditions. The proposed control solution presents an opportunity for distribution utilities to optimize the performance of distribution units and even minimize thermal losses. This control strategy installed at the PV inverter level proposes as future work reactive power control even during night periods when sunshine is not available, allowing the inverter to perform as a FACTS device. Rodríguez‐Amenedo et al.[115] proposed a hierarchical control architecture headed by a supervisory control system for active and reactive power control at the wind farm substation level in coordination with specific control actions implemented at the turbine level for local control of active and reactive power. The proposed structure allows the wind power to follow a given reference provided by the SO or

Handling Renewable Energy Variability and Uncertainty in Power System Operation  19

Distribution system operator

Transmission system operator

Wind power dispatch center

Wind farm manager #1

(1)

(2)

(n)

Wind farm manager #2

(1)

(n)

Wind farm manager #k

(1)

(2)

(n)

Figure 1.6  Architecture of a wind power dispatch center.

to  follow a strategy of maximization of power production if no reference is provided. Moursi et al.[116] described strategies for secondary voltage control in a remote grid bus through appropriate reactive power management in a wind farm equipped with a DFIG. System parameters such as short‐circuit ratio at the wind farm connection point, as well as communication requirements and inherent delays, are taken into consideration. Additionally, the authors developed a strategy for reactive power allocation to individual wind turbines within the wind farm. Almeida et  al.[117] proposed the development of a wind turbine dispatch strategy in a wind farm following SO requests either for active and/or reactive power. The authors consider that wind turbines are operated on a deloaded maximum power extraction

curve and will respond to a supervisory wind farm control after a request from a SO to adjust the outputs of the wind park. The authors also suggest the development of a wind power dispatch center concept, illustrated in Figure 1.6, that will interact with individual wind farms and with the SO to which they are connected in order to implement a concept similar to a virtual power plant that will allow for the global coordination of power injections from different wind farms. In cases with network congestion, a supervisory control strategy for dynamic redispatching of variable speed wind generators and conventional generators is proposed by Vergnol et al.[118] Under these conditions, grid integration of RES generation is increased while ensuring network security criteria.

20  Advances in Energy Systems

­MARKET RULES AND PRODUCTS FOR DEALING WITH VARIABILITY AND UNCERTAINTY Some changes in the market rules and new products can help the SO in managing variability and uncertainty. For example, the participation of RES generation in the electricity market through an intermediary agent or their integration in a portfolio with different generation technologies decreases the forecast error[119, 120]. Another possibility is very‐short‐term trading of RES. This can reduce regulation costs, compared with a day‐ahead horizon, by 30% for a horizon between 6 and 12 hours ahead, and 70% for 2 hours ahead[121]. Several authors tested a rolling‐window approach for SUC, with hourly updates of load and generation forecasts (i.e. simulating intraday trading). Tuohy et al.[79] found that the change in base‐load units is small, the use of mid‐merit gas turbines, and the total number of startups all decrease when the uncertainty decreases. Moreover, the number of hours where reserve requirements are not met increases when changing from an hourly to 6‐hours update. Meibom et al.[106] concluded that the demand for replacement reserve increases with wind power penetration and forecast horizon. An interesting conclusion is that the cost reduction due to perfect forecasts changes with the generation portfolio under study. For instance, the rescheduling cost is high in portfolios with a high share of less‐flexible power plants such as coal‐fired thermal plants or CCGTs. Weber[122] examined different market designs for facilitating operation with a large share of renewable sources. The most favorable design is based on intraday markets similar to the six voluntary intraday sessions of the Iberian market (Portugal and Spain). According to the author, this provides a good balance between the possibility of adjusting the RES bids by using very short‐term forecasts and sufficient liquidity (which is an essential property for market efficiency). The pricing system of balancing markets should be ­revised and harmonized across countries/control areas. Vandezande et  al.[123] compared single‐ and double‐ price systems for the balancing market. In the former, there is a single price for positive and negative imbalances, whereas in the second, there are asymmetric prices for imbalances. The authors show that the two‐ price system encourages trading agents to sell energy according to the lowest imbalance price. For example, if the energy shortage price is lower, the agent can develop a strategy for overestimating its generation. Conversely, the one‐price system encourages the

agent to offer its expected (or forecasted) value of generation. Bessa et  al.[124] also showed that asymmetric prices create an opportunity for introducing a bias in the WPFs, and that this can lead to over‐ and undercommitment situations. In the United States, the penalization schemes for wind power imbalances are more flexible[120]. For example, the Midwest SOs use an 8% tolerance band, and the California SOs settle imbalances on a monthly basis. In some cases, there are no penalizations for deviation[15]. Some market protocols designed for conventional units may not be appropriate for renewable generation, and can create unfair situations or even windfall profits. Sioshansi and Hurlbut[77] explained how a zonal market in ERCOT system created an incentive for overestimating wind power generation in order to receive a curtailment payment for solving network congestion. This allowed a payment even in situations when the wind was insufficient to meet the scheduled quantity. A nodal pricing system was adopted to avoid this behavior. Weijde and Hobbs[125] have shown that nodal prices can decrease startup and variable generation costs, resulting from full modeling of the transmission network (in contrast to approaches based on net transfer capacity). Furthermore, the authors also showed that the coordination of international balancing markets could lower the costs of UC and ED.

­EMERGENT APPROACHES Across the world, different initiatives with a strong involvement of the distribution system operator (DSO) aim to identify solutions toward the implementation of the smart grid concept[126]. Smart Grids provide additional capabilities for the observability and controllability at the distribution network level (e.g. distributed storage, distributed generation, microgeneration, controllable loads). The implementation of a smart grid concept is strongly supported by Information and Communications Technology (ICT), and its synergy with power system engineering enables new features, such as self‐healing and islanding operation, which improve system resilience in the presence of distributed RES generation. This smart grid concept creates conditions for demand management[127], either by sending direct control signals to the load or sending price signals to induce a certain reaction from the load. This changes the paradigm, moving it from “generation following load” to “load following generation.” The controllable loads can participate (or bid) in reserve markets primarily designed to solve RES forecast errors and

Handling Renewable Energy Variability and Uncertainty in Power System Operation  21

ramps. In fact, with increased wind power penetration, a flexible load such as a domestic heat pump can offer similar savings than a PHS, and is less sensitive to changes in fuel prices, interest rates, and the level of storage needs[108, 128]. A very flexible load for demand dispatch is the EV. An EV operating in vehicle‐to‐grid mode (V2G, i.e. bidirectional power injection) can play the same role as storage devices[129], but acting just as a controllable load (i.e. unidirectional controllable injection) can also provide reserve services[127, 129] that can be used for mitigating forecast errors and rapid ramps. Furthermore, an EV with a proper electronic interface control can provide primary frequency control in isolated and micro‐grid systems, allowing the integration of additional wind power without harming power system dynamic security[129]. The smart grid concept is also being transferred to the transmission network and power system operation[130]. The idea consists of developing smart dispatch centers with new functions for monitoring, analyzing, and controlling the power system. These new functions will use, for example, the monitoring capability of phasor measurement units, online algorithms for voltage and small‐signal stability analysis, and algorithms capable of estimating a snapshot of the system operating conditions for the next minutes/hour. Technological advances, such as FACTS, HVDC, and the smart substation concept, complement the new software functions.

­CONCLUSIONS Handling RES (mainly wind and solar power) variability and uncertainty in power system operation can be achieved through forecasting and management tools, and technological advances. Current forecasting algorithms offer an acceptable forecast error, and provide different approaches for modeling uncertainty. Uncertainty forecasts are an important input for stochastic management tools designed for several problems, such as setting the reserve requirements or the UC, that allow an accurate characterization of the risk associated with different decisions. Technological advances in storage devices and the possibility of controlling active/ reactive generation of RES power plants offer more flexibility from RES generation. A relevant issue is that the additional value of new management methods depends upon the characteristics of the power system under analysis. These characteristics are minimum generation levels, transmission and interconnections capacity, wind power penetra-

tion, maximum ramping, start‐up times, market gate closure, and so on. For example, a system with a high capacity of fast‐starting units (e.g. hydropower plants) has lower requirements for ramp forecasting algorithms or specific reserve for ramping events, and the impact on the reliability of generation variability is low. In a power system with a long period between day‐ahead scheduling and the operating day, and without intraday scheduling (or trading), good forecasts will be more valued. In interconnected systems, storage units should be used when the level of curtailed RES becomes significant in the power system. Nevertheless, this will depend on the generation portfolio, interconnection capacity, renewable energy forecast errors and variability. A “good” forecast[124] combined with suitable operational procedures (e.g. stochastic optimization, intraday markets) can avoid the need to invest in storage units. In isolated systems, the use of storage for timescales less than a minute is essential when the penetration levels of renewable generation are high. The adoption of probabilistic approaches for setting the reserve requirements and UC represent a radical departure from established procedures. Research concerning these methods provides important insights that could be used to improve the deterministic processes or to a full transition to stochastic methods. Finally, the electricity market role cannot be neglected, and the development of the smart grid paradigm should lead to new market rules capable of extracting the full potential of the new resources (i.e. renewable generation, active loads, and storage units).

­ACKNOWLEDGMENTS This work is funded in part by the ERDF – European Regional Development Fund through the COMPETE Programme (operational programme for competitiveness) and by national funds through the FCT  – Fundação para a Ciência e a Tecnologia (Portuguese Foundation for Science and Technology) within project ≪FCOMP – 01–0124‐FEDER‐022701≫. Ricardo J. Bessa and Bernardo Silva want to e­ xpress their gratitude to FCT for supporting this work under PhD Scholarships SFRH/BD/33738/2009 and SFRH/ BD/61600/2009, cofunded by the European Social Fund through the POPH program. The authors graciously thank the partners from the European projects ANEMOS.plus (contract no. 038692) and TWENTIES (contract no. 249812) funded in part by the European Commission under the sixth and seventh RTD Framework Programme.

22  Advances in Energy Systems

REFERENCES 1. Holttinen, H. (2012). Wind integration: experience, issues, and challenges. WIREs Energy Environ. 1: 243–255. 2. Abbad, J.R. (2010). Electricity market participation of wind farms: the success story of the Spanish pragmatism. Energy Policy 38: 3174–3179. 3. Monteiro C, Bessa R, Miranda V, et  al. Wind power forecasting: state‐of‐the‐art 2009. Argonne National Laboratory Report ANL/DIS‐10‐1, 2009. Available at:  http://www.dis.anl.gov/pubs/65613.pdf. (Accessed January 2013). 4. Ela, E. and O’Malley, M. (2012). Studying the variability and uncertainty impacts of variable generation at ­multiple timescales. IEEE Trans. Power Syst. 27: 1324–1333. 5. Lalor, G., Mullane, A., and O’Malley, M. (2005). Frequency control and wind turbine technologies. IEEE Trans. Power Syst. 20: 1905–1913. 6. Mohseni, M. and Islam, S.M. (2012). Review of international grid codes for wind power integration: diversity, technology and a case for global standard. Renewable Sustainable Energy Rev. 16: 3876–3890. 7. Carrasco, J.M., Franquelo, L.G., Bialasiewicz, J.T. et al. (2006). Power‐electronic systems for the grid integration of renewable energy sources: a survey. IEEE Trans. Ind. Electron. 53: 1002–1016. 8. de Almeida, R.G. and Lopes, J.A.P. (2007). Participation of  doubly fed induction wind generators in system ­frequency regulation. IEEE Trans. Power Syst. 22: 944–950. 9. Tsili, M. and Papathanassiou, S. (2009). A review of grid code technical requirements for wind farms. IET Renew. Power Gener. 3: 308–332. 10. Alatrash H, Mensah A, Mark E, et al. Generator emulation controls for photovoltaic inverters. In IEEE 8th International Conference on Power Electronics and ECCE Asia (ICPE & ECCE). Korea: The Shilla Jeju; August 30‐September 1, 2011. 11. Ramtharan, G., Ekanayake, J.B., and Jenkins, N. (2007). Frequency support from doubly fed induction generator wind turbines. IET Renew. Power Gener. 1: 3–9. 12. Silva, B., Moreira, C.L., Seca, L. et al. (2012). Provision of inertial and Sprimary frequency control services using offshore multiterminal HVDC networks. IEEE Trans. Sustainable Energy 3: 800–808. 13. Kottick, D., Blau, M., and Edelstein, D. (1993). Battery energy storage for frequency regulation in an island power system. IEEE Trans. Energy Convers. 8: 455–459. 14. Aditya, S.K. and Das, D. (2001). Battery energy storage for load frequency control of an interconnected power system. Electr. Power Syst. Res. 58: 179–185. 15. Botterud, A., Wang, J., Miranda, V. et al. (2010). Wind power forecasting in U.S. electricity markets. Electr. J. 23: 1040–6190. 16. Reikard, G., Pinson, P., and Bidlot, J.‐R. (2011). Forecasting ocean wave energy: the ECMWF wave

model and time series methods. Ocean Eng. 38: 1089–1099. 17. Lorenz, E., Scheidsteger, T., Hurka, J. et  al. (2011). Regional PV power prediction for improved grid integration. Prog. Photovoltaics Res. Appl. 19: 757–771. 18. Taylor, J.W. and McSharry, P.E. (2007). Short‐term load forecasting methods: an evaluation based on European data. IEEE Trans. Power Syst. 22: 2213–2219. 19. Contaxis, G.C. and Kabouris, J. (1991). Short term scheduling in a wind/diesel autonomous energy system. IEEE Trans. Power Syst. 6: 1161–1167. 20. Pinson, P. and Madsen, H. (2012). Adaptive modelling and forecasting of offshore wind power fluctuations with Markov‐switching autoregressive models. J. Forecasting 31: 281–313. 21. Giebel G, Badger J, Louka P, et  al. Description of NWP,  mesoscale and CFD models. Deliverable 4.1, EU  Project ANEMOS; 2006. Available at: http://130.226.56.153/zephyr/publ/ANEMOS_D4.1_ ModelDescription.pdf. (Accessed January 2013). 22. Larson, K.A. and Westrick, K. (2006). Short‐term wind forecasting using off‐site observations. Wind Energy 9: 55–62. 23. Bessa, R.J., Miranda, V., and Gama, J. (2009). Entropy and correntropy against minimum square error in offline and online three‐day ahead wind power forecasting. IEEE Trans. Power Syst. 24: 1657–1666. 24. Thordarson, F.Ö., Madsen, H., Nielsen, H.A. et  al. (2010). Conditional weighted combination of wind power forecasts. Wind Energy 13: 751–763. 25. Siebert N. Development of methods for regional wind power forecasting. PhD Thesis. Centre Energétique et Procédés, Mines ParisTech; 2008. 26. Focken, U., Lange, M., Mönnich, K. et al. (2002). Short‐ term prediction of the aggregated power output of wind farms—a statistical analysis of the reduction of the ­prediction error by spatial smoothing effects. J. Wind Eng. Ind. Aerodyn. 90: 231–246. 27. Bremnes, J.B. (2004). Probabilistic wind power forecasts using local quantile regression. Wind Energy 7: 47–54. 28. Bessa, R.J., Miranda, V., Botterud, A. et  al. (2012). Time‐adaptive quantile‐copula for wind power probabilistic forecasting. Renewable Energy 40: 29–39. 29. Castronuovo, E.D. and Lopes, J.A.P. (2004). On the optimization of the daily operation of a wind‐hydro power plant. IEEE Trans. Power Syst. 19: 1599–1606. 30. Wang, J., Botterud, A., Bessa, R. et  al. (2011). Wind power forecasting uncertainty and unit commitment. Appl. Energy 88: 4014–4023. 31. Pinson, P., Papaefthymiou, G., Klockl, B. et al. (2009). From probabilistic forecasts to statistical scenarios of short‐term wind power production. Wind Energy 12: 51–62. 32. Papaefthymiou G, Pinson P. Modeling of spatial dependence in wind power forecast uncertainty. In PMAPS 2008: 10th International Conference on Probabilistic Methods Applied to Power Systems. IEEE: Puerto Rico; May 25–29, 2008.

Handling Renewable Energy Variability and Uncertainty in Power System Operation  23

33. Leutbecher, M. and Palmer, T.N. (2008). Ensemble forecasting. J. Comput. Phys. 227: 3515–3539. 34. Ziehmann, C. (2000). Comparison of a single‐model EPS with a multi‐model ensemble consisting of a few operational models. Tellus Ser. A 52: 280–299. 35. Nielsen HA, Madsen H, Nielsen TS, et al. Wind power ensemble forecasting using wind speed and direction ensembles from ECMWF or NCEP. Denmark Project PSO, Technical Report, 2005. Available at: http://www. risoe.dtu.dk/rispubl/NEI/nei‐dk‐4552.pdf. (Accessed January 2013). 36. Cutler, N.J., Outhred, H.R., MacGill, I.F. et al. (2008). Characterizing future large, rapid changes in aggregated wind power using numerical weather prediction spatial fields. Wind Energy 12: 542–555. 37. Cutler, N.J., Outhred, H.R., MacGill, I.F. et al. (2012). Predicting and presenting plausible future scenarios of wind power production from numerical weather prediction systems: a qualitative ex ante evaluation for decision making. Wind Energy 15: 473–488. 38. Bossavy, A., Girard, R., and Kariniotakis, G. (2013). Forecasting ramps of wind power production with numerical weather prediction ensembles. Wind Energy 16: 51–63. 39. Ferreira C, Gama J, Miranda V, et al. Probabilistic ramp detection and forecasting for wind power prediction. In  PMAPS 2012—International Conference on Probabilistic Methods Applied to Power Systems. IEEE: Istanbul, Turkey; June 10–14, 2012. 40. Mellit, A. and Pavan, A.M. (2010). A 24‐h forecast of solar irradiance using artificial neural network: application for performance prediction of a grid‐connected PV plant at Trieste, Italy. Sol. Energy 84: 807–821. 41. Ahlstrom M, Kankiewicz A. Solar power forecasting. In  Solar Power: Making High Penetration Possible Workshop. Iowa: Cedar Rapids; October 15, 2010. 42. Heinemann, D., Lorenz, E., and Girodo, M. (2006). Forecasting of solar radiation. In: Solar Energy Resource Management for Electricity Generation from Local Level to Global Scale (ed. E.D. Dunlop, L. Wald and M. Suri), 83–94. New York: Nova Science Publishers Inc. 43. Bannehr, L., Rohn, M., and Warnecke, G. (1996). A  functional analytic method to derive displacement vector fields from satellite image sequences. Int. J. Remote Sens. 17: 383–392. 44. Wittmann, M., Breitkreuz, H., Schroedter‐Homscheidt, M. et al. (2008). Case studies on the use of solar irradiance forecast for optimized operation strategies of solar thermal power plants. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 1: 18–27. 45. Lorenz, E., Hurka, J., Heinemann, D. et  al. (2009). Irradiance forecasting for the power prediction of grid‐ connected photovoltaic systems. IEEE J. Spec. Top Earth Obs. Remote Sens. 2: 2–10. 46. Lannoye, E., Flynn, D., and O’Malley, M. (2012). Evaluation of power system flexibility. IEEE Trans. Power Syst. 27: 922–931. 47. Harnessing Variable Renewables  –  A Guide to the Balancing Challenge. Paris, France: OECD/IEA; 2011.

48. Lokhov A. Technical and economic aspects of load following with nuclear power plants. Nuclear Energy Agency; 2011. Available at: http://www.oecd‐nea.org/ ndd/reports/2011/load‐following‐npp.pdf. (Accessed January 2013). 49. Auer J, Keil J. State‐of‐the‐art electricity storage ­systems. Deutsche Bank Research, 2012. Available at: http://www.dbresearch.com/PROD/DBR_INTERNET_ EN‐PROD/PROD0000000000286166/State‐of‐the‐art+ electricity+storage+systems%3A+Indispensable+ elements+of+the+energy+revolution.PDF. (Accessed January 2013). 50. Nicolosi, M. (2010). Wind power integration and power system flexibility—an empirical analysis of extreme events in Germany under the new negative price regime. Energy Policy 38: 7257–7268. 51. Soder, L., Abildgaard, H., Estanqueiro, A. et al. (2012). Experience and challenges with short‐term balancing in European systems with large share of wind power. IEEE Trans. Sustainable Energy 3: 853–861. 52. Gül T, Stenzel T. Variability of wind power and other renewables—management options strategies. International Energy Agency Report, 2005. Available at: http://www.uwig.org/IEA_Report_on_variability.pdf. (Accessed January 2013). 53. European wind integration study (EWIS) towards a successful integration of wind power into European ­ electricity grids. Phase I Final Report; 2007. Available at: http://www.wind‐integration.eu/downloads/library/ EWIS_Final_Report.pdf. (Accessed January 2013). 54. ENTSO‐E. Operating Handbook—Policies P1: Load‐ Frequency Control and Performance; 2009. Available at: https://www.entsoe.eu/publications/system ‐ operations ‐ reports / operation ‐ handbook/. (Accessed January 2013). 55. Banakar, H., Luo, C., and Ooi, B.T. (2008). Impacts of wind power minute‐to‐minute variations on power system operation. IEEE Trans. Power Syst. 23: 150–160. 56. Holttinen, H., Milligan, M., Ela, E. et  al. (2012). Methodologies to determine operating reserves due to increased wind power. IEEE Trans. Sustainable Energy 3: 713–723. 57. Rebours Y. A comprehensive assessment of markets for frequency and voltage control ancillary services. PhD Thesis. University of Manchester, Faculty of Engineering and Physical Sciences; 2008. 58. Anstine, L.T., Burke, R.E., Casey, J.E. et  al. (1963). Application of probability methods to the determination of spinning reserve requirements for the Pennsylvania– New Jersey–Maryland interconnection. IEEE Trans. Power Apparatus Syst. 82: 720–735. 59. ERCOT. ERCOT methodologies for determining ancillary service requirements; 2011. 60. Secretaría General de Energía. REE P.O. 1.5— Establecimiento de la reserva para la regulación frecuencia‐potencia. BOE núm. 2006, 173:27473–27474. 61. Gouveia, E. and Matos, M.A. (2009). Evaluating operational risk in a power system with a large amount of wind power. Electr. Power Syst. Res. 79: 734–739.

24  Advances in Energy Systems

62. Soder, L. (1993). Reserve margin planning in a wind‐ hydrothermal power system. IEEE Trans. Power Syst. 8: 564–571. 63. Holttinen, H. (2005). Impact of hourly wind power variations on the system operation in the Nordic countries. Wind Energy 8: 197–218. 64. Doherty, R. and O’Malley, M. (2005). New approach to quantify reserve demand in systems with significant installed wind capacity. IEEE Trans. Power Syst. 20: 587–595. 65. Ortega‐Vazquez, M.A. and Kirschen, D.S. (2009). Estimating the spinning reserve requirements in systems with significant wind power generation penetration. IEEE Trans. Power Syst. 24: 114–124. 66. Lange, M. (2005). On the uncertainty of wind power predictions—analysis of the forecast accuracy and statistical distribution of errors. J. Sol. Energy Eng. 127: 177–194. 67. Bludszuweit, H., Dominguez‐Navarro, J.A., and Llombart, A. (2008). Statistical analysis of wind power forecast error. IEEE Trans. Power Syst. 23: 983–991. 68. Bessa RJ, Matos MA. Comparison of probabilistic and deterministic approaches for setting operating reserve in systems with high penetration of wind power. In 7th Mediterranean Conference and Exhibition on Power Generation, Transmission, Distribution and Energy Conversion (Medpower). Agia Napa: Cyprus; November 7–10, 2010. 69. Matos, M.A. and Bessa, R.J. (2011). Setting the operating reserve using probabilistic wind power forecasts. IEEE Trans. Power Syst. 26: 594–603. 70. Bessa, R.J., Matos, M.A., Costa, I.C. et  al. (2012). Reserve setting and steady‐state security assessment using wind power uncertainty forecast: a case study. IEEE Trans. Sustainable Energy 3: 827–836. 71. Menemenlis, N., Huneault, M., and Robitaille, A. (2012). Computation of dynamic operating balancing reserve for wind power integration for the time‐horizon 1–48 hours. IEEE Trans. Sustainable Energy 3: 692–702. 72. Maurer, C., Krahl, S., and Weber, H. (2009). Dimensioning of secondary and tertiary control reserve by probabilistic methods. Eur. Trans. Electr. Power 19: 544–552. 73. Pahlow, M., Möhrlen, C., and Jørgensen, J.U. (2009). Application of cost functions for large scale integration of wind power using a multi‐scheme ensemble prediction technique. In: Optimization Advances in Electric Power Systems (ed. E.D. Castronuovo). New York: Nova Science Publishers Inc. 74. Lang S, Möhrlen J, Jørgensen J, et al. Application of a multi‐scheme ensemble prediction system for wind power forecasting in Ireland and comparison with validation results from Denmark and Germany. In European Wind Energy Conference EWEC’06. Athens, Greece; February 27‐March 2, 2006. 75. Otterson S, Pinson P. Are ramp forecasts really useful? In EWEA Annual Event. Copenhagen, Denmark; April 16–19, 2012.

76. Lin, J., Sun, Y.‐Z., Cheng, L. et al. (2012). Assessment of the power reduction of wind farms under extreme wind condition by a high resolution simulation model. Appl. Energy 96: 21–32. 77. Sioshansi, R. and Hurlbut, D. (2010). Market protocols in ERCOT and their effect on wind generation. Energy Policy 38: 3192–3197. 78. Ruiz, P.A., Philbrick, C.R., Zak, E. et  al. (2009). Uncertainty management in the unit commitment problem. IEEE Trans. Power Syst. 24: 642–651. 79. Tuohy, A., Meibom, P., Denny, E. et  al. (2009). Unit commitment for systems with significant wind penetration. IEEE Trans. Power Syst. 24: 592–601. 80. Wang, J., Shahidehpour, M., and Li, Z. (2008). Security‐ constrained unit commitment with volatile wind power generation. IEEE Trans. Power Syst. 23: 1319–1327. 81. Bouffard, F. and Galiana, F.D. (2008). Stochastic security for operations planning with significant wind power generation. IEEE Trans. Power Syst. 23: 306–316. 82. Pappala, V.S., Erlich, I., Rohrig, K. et  al. (2009). Stochastic model for the optimal operation of a wind‐ thermal power system. IEEE Trans. Power Syst. 24: 940–950. 83. Xiao, J., Hodge, B.‐M.S., Pekny, J.F. et  al. (2011). Operating reserve policies with high wind power penetration. Comput. Chem. Eng. 35: 1876–1885. 84. Constantinescu, E.M., Zavala, V.M., Rocklin, M. et al. (2011). A computational framework for uncertainty quantification and stochastic optimization in unit commitment with wind power generation. IEEE Trans. Power Syst. 26: 431–441. 85. Zhou, Z., Botterud, A., Wang, J. et al. (2012). Application of probabilistic wind power forecasting in electricity markets. Wind Energy. 86. Sturt, A. and Strbac, G. (2012). Efficient stochastic scheduling for simulation of wind‐integrated power systems. IEEE Trans. Power Syst. 27: 323–334. 87. Growe‐Kuska N, Heitsch H, Romisch W. Scenario reduction and scenario tree construction for power management problems. In IEEE Bologna PowerTech ­ Conference. Bologna, Italy; June 23–26, 2003. 88. Sumaili J, Keko H, Miranda V, et  al. Finding representative wind power scenarios and their probabilities for stochastic models. In ISAP 2011  –  16th International Conference on Intelligent System Applications to Power Systems. Hersonissos, Greece; September 25–28, 2011. 89. Morales, J.M., Conejo, A.J., and Pérez‐Ruiz, J. (2009). Economic valuation of reserves in power systems with high penetration of wind power. IEEE Trans. Power Syst. 24: 900–910. 90. Restrepo, J.F. and Galiana, F.D. (2011). Assessing the yearly impact of wind power through a new hybrid deterministic/stochastic unit commitment. IEEE Trans. Power Syst. 26: 401–410. 91. Papavasiliou, A., Oren, S.S., and O’Neill, R.P. (2011). Reserve requirements for wind power integration: a scenario‐based stochastic programming framework. ­ IEEE Trans. Power Syst. 26: 2197–2206.

Handling Renewable Energy Variability and Uncertainty in Power System Operation  25

92. Lowery, C. and O’Malley, M. (2012). Impact of wind forecast error statistics upon unit commitment. IEEE Trans. Sustainable Energy 3: 760–768. 93. Wu, L., Shahidehpour, M., and Li, T. (2007). Stochastic security‐constrained unit commitment. IEEE Trans. Power Syst. 22: 800–811. 94. Villanueva, D., Pazos, J.L., and Feijóo, A. (2011). Probabilistic load flow including wind power generation. IEEE Trans. Power Syst. 26: 1659–1667. 95. Hatziargyriou, N.D., Karakatsanis, T.S., and Papadopoulos, M. (1993). Probabilistic load flow in distribution systems containing dispersed wind power generation. IEEE Trans. Power Syst. 8: 159–165. 96. Morales, J.M., Baringo, L., Conejo, A.J. et al. (2010). Probabilistic power flow with correlated wind sources. IET Gener. Transm. Distrib. 4: 641–651. 97. Usaola, J. (2009). Probabilistic load flow in systems with wind generation. IET Gener. Transm. Distrib. 3: 1031–1041. 98. ENTSO‐E. Operating Handbook  –  Policy 3: Operational Security; 2009. Available at: https://www. entsoe.eu/publications/system‐operations‐reports/ operation‐handbook/. (Accessed January 2013). 99. Vlachogiannis, J.G. (2009). Probabilistic constrained load flow considering integration of wind power generation and electric vehicles. IEEE Trans. Power Syst. 24: 1808–1817. 100. Jabr, R.A. and Pal, B.C. (2009). Intermittent wind generation in optimal power flow dispatching. IET Gener. Transm. Distrib. 3: 66–74. 101. Capitanescu, F., Fliscounakis, S., Panciatici, P. et  al. (2012). Cautious operation planning under uncertainties. IEEE Trans. Power Syst. 27: 1859–1869. 102. Díaz‐González, F., Sumper, A., Gomis‐Bellmunt, O. et al. (2012). A review of energy storage technologies for wind power applications. Renewable Sustainable Energy Rev. 16: 2154–2171. 103. Hedegaard, K. and Meibom, P. (2012). Wind power impacts and electricity storage—a time scale perspective. Renewable Energy 37: 318–324. 104. Loisel, R., Mercier, A., Gatzen, C. et al. (2011). Market evaluation of hybrid wind‐storage power systems in  case of balancing responsibilities. Renewable Sustainable Energy Rev. 15: 5003–5012. 105. Nyamdash, B., Denny, E., and O’Malley, M. (2010).  The viability of balancing wind generation with large scale energy storage. Energy Policy 38: 7200–7208. 106. Meibom, P., Barth, R., Hasche, B. et  al. (2011). Stochastic optimization model to study the operational impacts of high wind penetrations in Ireland. IEEE Trans. Power Syst. 26: 1367–1379. 107. Tuohy, A. and O’Malley, M. (2011). Pumped storage in systems with very high wind penetration. Energy Policy 39: 1965–1974. 108. Connolly, D., Lund, H., Mathiesen, B.V. et al. (2012). The technical and economic implications of integrating fluctuating renewable energy using energy storage. Renewable Energy 43: 47–60.

109. Qu, L. and Qiao, W. (2011). Constant power control of DFIG wind turbines with supercapacitor energy storage. IEEE Trans. Ind. Appl. 47: 359–367. 110. Cimuca, G.O., Saudemont, C., Robyns, B. et al. (2006). Control and performance evaluation of a flywheel energy‐ storage system associated to a variable‐speed wind generator. IEEE Trans. Ind. Electron. 53: 1074–1085. 111. Ray, P.K., Mohanty, S.R., and Kishor, N. (2011). Proportional–integral controller based small‐signal analysis of hybrid distributed generation systems. Energy Convers. Manage. 52: 1943–1954. 112. Nomura, S., Ohata, Y., Hagita, T. et  al. (2005). Wind farms linked by SMES systems. IEEE Trans. Appl. Supercond. 15: 1951–1954. 113. Oudalov, A., Chartouni, D., and Ohler, C. (2007). Optimizing a battery energy storage system for primary frequency control. IEEE Trans. Power Syst. 22: 1259–1266. 114. Lee, D.‐J. and Wang, L. (2008). Small‐signal stability analysis of an autonomous hybrid renewable energy power generation/energy storage system part I: time‐domain simulations. IEEE Trans. Energy Convers. 23: 311–320. 115. Rodríguez‐Amenedo, J.L., Arnalte, S., and Burgos, J.C. (2002). Automatic generation control of a wind farm with variable speed wind turbines. IEEE Trans. Energy Convers. 17: 279–284. 116. Moursi, M.E., Joos, G., and Abbey, C. (2008). A secondary voltage control strategy for transmission level interconnection of wind generation. IEEE Trans. Power Electron. 23: 1178–1190. 117. de Almeida, R.G., Castronuovo, E.D., and Lopes, J.A.P. (2006). Optimum generation control in wind parks when carrying out system operator requests. IEEE Trans. Power Syst. 21: 718–725. 118. Vergnol, A., Sprooten, J., Robyns, B. et al. (2011). Line overload alleviation through corrective control in presence of wind energy. Electr. Power Syst. Res. 81: 1583–1591. 119. Ummels, B.C., Gibescu, M., Kling, W.L. et al. (2006). Integration of wind power in the liberalized Dutch electricity market. Wind Energy 9: 579–590. 120. Aparicio, N., MacGill, I., Rivier Abbad, J. et al. (2012). Comparison of wind energy support policy and electricity market design in Europe, the United States, and Australia. IEEE Trans. Sustainable Energy 3: 809–818. 121. Holttinen, H. (2005). Optimal electricity market for wind power. Energy Policy 33: 2052–2063. 122. Weber, C. (2010). Adequate intraday market design to enable the integration of wind energy into the European power systems. Energy Policy 38: 3155–3163. 123. Vandezande, L., Meeus, L., Belmans, R. et al. (2010). Well‐functioning balancing markets: a prerequisite for  wind power integration. Energy Policy 38: 3146–3154. 124. Bessa, R.J., Miranda, V., Botterud, A. et  al. (2011). ‘Good’ or ‘bad’ wind power forecasts: a relative ­concept. Wind Energy 14: 625–636. 125. Weijde, A.H. and Hobbs, B.F. (2011). Locational‐ based coupling of electricity markets: benefits from

26  Advances in Energy Systems

coordinating unit commitment and balancing markets. J. Regul. Econ. 39: 223–251. 126. Farhangi, H. (2010). The path of the smart grid. IEEE Power Energ. Mag. 8: 18–28. 127. Brooks, A., Lu, E., Reicher, D. et al. (2010). Demand dispatch. IEEE Power Energ. Mag. 8: 20–29. 128. Papaefthymiou, G., Hasche, B., and Nabe, C. (2012). Potential of heat pumps for demand side management and wind power integration in the German electricity market. IEEE Trans. Sustainable Energy 3: 636–642. 129. Lopes, J.A.P., Soares, F.J., and Almeida, P.R. (2011). Integration of electric vehicles in the electric power system. Proc. IEEE 99: 168–183. 130. Fangxing, L., Wei, Q., Hongbin, S. et al. (2010). Smart transmission grid: vision and framework. IEEE Trans. Smart Grid 1: 168–177. 131. Redes Energéticas Nacionais website. Available at: http://www.centrodeinformacao.ren.pt/EN/Informacao

Exploracao/Pages/DiagramadeProdu%C3%A7%C3% A3oE%C3%B3lica.aspx. (Accessed January 2013).

FURTHER READING Birge, J.R. and Louveaux, F. (2011). Introduction to ­Stochastic Programming, 2e. New York: Springer. Ela E, Milligan M, Kirby B. Operating reserves and ­variable generation. Technical Report, 2011, NREL/ TP‐5500–51978. Available at: http://www.nrel.gov/docs/ fy11osti/51978.pdf. Holttinen H, Meibom P, Orths A, et al. Design and operation of power systems with large amounts of wind power. Final Report. Phase one 2006–2008, IEA Wind Task 25. Available at: http://www.vtt.fi/inf/pdf/tiedotteet/2009/T2493.pdf. Matos, M.A. (2007). Decision under risk as a multicriteria problem. Eur. J. Oper. Res. 138: 1516–1529.

2

Short‐Term Frequency Response of Power Systems with High Nonsynchronous Penetration Levels Lisa Ruttledge and Damian Flynn School of Electrical and Electronic Engineering, University College Dublin, Dublin, Ireland

In addition to the generation and transmission of electricity, the reliable and secure operation of power systems also relies on various ancillary support services supplied by power plants and loads. The portfolio of required support services has historically been well understood. However, as conventional power plant are displaced by modern variable renewable generation, such as wind and solar, so too are those traditional ancillary support services. As a result, new support services are required and changes in operational behavior are necessary. The impact of increased variable generation levels on one such support service category, short‐term system frequency response, is considered here. The evolution of power system behavior with respect to this service is assessed, with potential solutions to the associated challenges explored, and pertinent future questions in this area highlighted. ­INTRODUCTION Electricity is one of the most widely used commodities in the world today, with its share of global final energy consumption doubling over the past four

decades[1]. Electric power systems form the backbone of the ­production, delivery, management, and use of electricity. Therefore, efficient and reliable system operation is of paramount importance to ensure its effective delivery from generators to loads. While the associated technical, economic, and market challenges have traditionally been very well understood, the introduction of modern variable renewable sources, at scale, heralds an evolution of operational challenges. Electrical demand generally follows certain daily and seasonal patterns. For example, the daily demand profile begins at a low level in the morning, ramps up as people wake up and activity increases, and exhibits peaks and troughs as people follow daily routines, e.g. high demand when cooking evening meals or low demand during the night[2]. Today, electricity cannot be economically or conveniently stored at scale and so must be generated as needed. Large power plant can take several hours to start up, depending on the plant technology employed, and so system operators must plan ahead to ensure that sufficient generation is online to meet forecasted demand at all times. Unit commitment is the process of optimizing plant scheduling (on/off decisions) for a time horizon of generally one

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

28  Advances in Energy Systems

day to one week to meet the predicted total demand at lowest cost, while respecting constraints, such as thermal limits of transmission lines, minimum/maximum generation levels of units, etc. Nearer real‐time, that is from tens of minutes to a few hours, more accurate forecasts may become available, depending on the system in question, and using knowledge on the status of individual plant, unit power outputs are dispatched to meet demand through a process called economic dispatch. However, while daily and seasonal patterns can be observed, the actual electrical demand, for a given point in time, exhibits a degree of variability and uncertainty. In real‐time, as fluctuations in demand occur, governor controls allow generators to respond to second‐by‐second variations automatically through varying the volume of working fuel (e.g. steam), and hence mechanical input power, in the turbines of conventional plant, thus allowing them to ramp their outputs up and down. Over longer time frames (seconds to minutes) central dispatching, such as automatic generation control (AGC) in many systems, adjusts the power set‐point of individual plant in response to gradual demand fluctuations (load following), to maintain system balance[3]. Power System Frequency Response Following Contingency In general, fossil fuel power plant burn fuel to generate steam, to drive a multi‐stage turbine, which in turn powers a generator to convert mechanical energy into AC (alternating current) electricity. The generator rotational speed is directly coupled to the frequency of the AC, and so such a machine is known as a synchronous generator. The rotating masses of the generator and associated turbine act as a short‐term store of energy and can be quantified by the plant’s inertial (H) constant: a measure of how long (seconds) the unit could operate at maximum output using its stored rotational energy alone. Following an imbalance between generation and demand, and in the absence of other energy storage on the system, energy is inherently absorbed or released (to counteract the imbalance) by the rotating masses coupled to the system speed (frequency) slowing down or speeding up. For example, for a deficit in systemwide generation compared to demand, the loading on the generating plant must increase. As the combined mechanical input to the generators is less than the electrical output (demand), the rotating masses associated with those machines slow down to provide some of their stored energy as a short‐term energy boost, resulting in a drop in system frequency. Similarly, if the system demand is less than the combined generation output,

the loading is reduced and the synchronous machines accelerate, absorbing surplus energy, resulting in an increase in system frequency. Similarly, synchronous motor loads, in the form of pumps, fans, drives, etc. scattered across the power system, which are also coupled to the system frequency, add to the inertial storage[4]. As a result, any power imbalance on the system appears as a deviation in system frequency from nominal with the severity of the deviation, i.e. its speed and magnitude, dependent on the size of the power imbalance and the stored energy available on the system at the time of the event. While the nominal electrical frequency in a jurisdiction depends on that adopted by individual synchronous areas (e.g. 60 Hz in North America and parts of South America, 50 Hz in most other jurisdictions), the frequency is ideally kept constant across large power systems. Tight frequency control is important for maintaining generating plant performance, which is dependent on the correct operation of auxiliary drives associated with fuel, feedwater and combustion air supply systems: as the frequency falls, the drives slow down, reducing their ability to sustain performance. In addition, for mechanical structures which have natural frequencies close to the nominal system frequency, resonant vibration of machine components is also a risk, with steam turbine blade failure due to fatigue from over‐ or under‐frequency operation a possibility[5]. However, such mechanical component failure is not generally a concern unless the frequency deviates by more than about 5%. In the past, the extensive use of electric clocks based on synchronous motors and the use of system frequency for other timing purposes required accurate maintenance of synchronous time[3]. Furthermore, a considerable drop in frequency could result in high magnetizing currents in induction motors and transformers. As a result of the significant influences that the system frequency has on the above factors, among others, its variability is an important stability metric, and it must remain within an acceptable range of operation for all connected equipment. Operating Standards Power systems around the world have defined standards of frequency behavior that all connected plant must be capable of riding through[2, 6, 7]. For example, Table  2.1 outlines the frequency response ranges for the combined Ireland and Northern Ireland system under various conditions[8]. If the frequency approaches, or exceeds, the thresholds defined in a ­jurisdiction’s operating standards, emergency controls

Short‐Term Frequency Response of Power Systems with High Nonsynchronous Penetration Levels  29

are activated, such as the activation of static reserves and, if necessary, the tripping of load to restore balance on the system. Imbalances Under normal conditions, a balance between g­eneration and demand is maintained; however, in light of the large number of constituent parts, power systems are subject to sudden faults or contingency events, which may act to compromise the generation‐demand balance. Frequency issues can also arise due to unintended consequences of system operating procedures, such as frequency wobbles experienced within the main European synchronous area of European Network of Transmission System Operators for Electricity (ENTSO‐E) at hour boundaries, up to 2011, caused by the standardized time interval for cross‐border schedule changes[9]. Frequency thresholds, such as those in Table 2.1, must be maintained by ensuring sufficient reserves (additional capacity to cope with unexpected events and forecast errors) across all time frames or other responses from power plant during contingency events. In order to hedge against any imbalance or other undesirable effects of a contingency event, support services are required in the form of short‐term plant responses and reserve, as illustrated in Figure  2.1. The system frequency response can generally be categorized into noninstantaneous

events, such as an error between forecasted and realized demand, and instantaneous events, such as the unexpected tripping of a generator or large load. Power plant are dispatched such that there is flexibility available to ramp their outputs up or down to regain system balance for noninstantaneous disturbances, through regulation of small real‐time imbalances, load‐following due to errors associated with forecast load and (variable) renewable generation production, and ramping capabilities due to significant changes in net load. Governor frequency droop, whereby a generator increases its output for a fall (droop) in frequency, is based on the generator’s nameplate capacity and generally has a value of 4–5%, that is for a 4–5% decrease in system frequency the generator output would vary from 0 to Table 2.1  Ireland and Northern Ireland system frequency ranges. Description

Range

Normal operating range During Transmission System disturbances During exceptional Transmission System disturbances, not exceeding 60 minutes duration for frequency in the range 47.5– 48.8 Hz and 50.2–52 Hz and not exceeding 20 s for frequency in the range 47–47.5 Hz Maximum rate of change of frequency

49.8–50.2 Hz 48.2–52 Hz 47–52 Hz

0.5 Hz s−1

Operating reserve Noninstantaneous

Regulation

Load following

Instantaneous

Ramping

Synchronous inertial response

Contingency

Fast-acting response

Short-term frequency response

Figure 2.1  Frequency response categories.

Primary reserve (Droop)

Secondary reserve

Tertiary reserve

Return frequency to nominal

Replace primary and secondary reserves

30  Advances in Energy Systems

1 pu, based on the unit rating. Over longer periods (minutes) the power set‐point of individual plant can be changed either through a central AGC mechanism or through communication with the system operator. The response time depends on the lags associated with aspects such as fuel supply and steam pressure within the power plant system. For example, in hydroelectric plant, when the governor gate is initially opened, due to water inertia and water being an incompressible liquid, water flow does not immediately increase, causing the pressure across the turbine to reduce, and so the power output also reduces initially before later increasing[3]. For sudden imbalances following a contingency event, the emergency control can be divided into a number of different phases, comprising inertial response, fast‐acting frequency response, and primary, secondary, and tertiary reserves based on the time frame over which they act. Inertial response, fast frequency response (FFR), and primary operating reserve act in the seconds following an event to arrest the initial imbalance and establish a steady‐state recovery frequency, while secondary and tertiary actions provide more sustained responses following the event. Following the inherent temporary energy boost provided by the synchronous inertial response, the fast‐acting response from various sources, or the primary frequency response based on active controls, provide a controlled response to arrest the falling frequency and initiate recovery toward nominal frequency by reducing the generation‐demand imbalance. The inertial response and FFR operate over similar time frames, and in many systems the terms may be used interchangeably. In particular, a fast‐ acting frequency response could be described as an inertial response. However, in some systems, typically smaller systems, such as Ireland and Northern Ireland, a clear demarcation is made between the two, whereby an inertial response can only be provided by synchronous machine‐based generators (and hence there is no delay in the activation of the response following an event occurring) and a fast‐ acting frequency response which can be provided by synchronous or nonsynchronous generators (and hence a delay is permissible before the response is provided). Similar to noninstantaneous events, the predominant source of primary frequency response is through an automatic governor droop response of generators operating below rated output. While governor action should arrest the drop in frequency following a sudden large imbalance, it will not return the frequency to the pre‐disturbance value, which is instead achieved though AGC or secondary/tertiary reserve,

through central control. Due to their coupling with the system frequency, all synchronous units will return to nominal frequency in sympathy with each other, while recovering their lost rotational energy. Generally, conventional generators act to restore balance within acceptable frequency operational and reliability limits as per Table 2.1. For the purpose of this review, the response of interest is the short‐term dynamic response to significant instantaneous events, which is defined here as those responses up to and including the primary frequency response, as highlighted in Figure 2.1. Over the past couple of decades, concerns over the environmental impact of fossil fuels and security of energy supply have driven the large‐scale deployment of renewable sources. Targets for electricity from renewable sources include 20% across Europe as a whole by 2020, with higher targets in many individual jurisdictions[10]. Ambitious renewable targets are also being proposed in regions such as China (20%), Alaska (50%), and Texas (10 GW renewable energy) to be achieved over the next decade[11]. A number of power systems already have high renewable penetrations, such as Sweden (58%) and Hydro Quebec (98%), although this is largely based on hydro‐electric generation[12, 13]. In some smaller systems significantly higher penetration levels are being pursued, such as 90% in New Zealand by 2025, where again much will be achieved using current hydro‐electric sources[14]. In 2012, the global renewable energy capacity grew by over 8% to 1560 GW, with solar photovoltaic (PV) and hydropower each accounting for about one‐third of new capacity. In 2013, renewables accounted for more than 56% of net additions to global power capacity, while representing far higher proportions of new capacity in several countries[12]. By their nature, many renewable sources of electricity generation, such as wind and solar, are variable, that is their available output is dependent upon weather patterns, rather than being under direct operator control. While modern variable generation (VG) may be dispatched to a level below their available power (curtailed), their output is limited by the renewable resource available at that point in time, unlike the instantaneous fuel supply available to conventional plants. For economic and governmental policy reasons, generally variable renewable plants have priority dispatch, with conventional plants serving the remaining system load, often called the net load. With the introduction of such variable generation sources, the variability and uncertainty of the net load increases such that the challenge of retaining power balance, also known as system flexibility, is likely to increase[15].

Short‐Term Frequency Response of Power Systems with High Nonsynchronous Penetration Levels  31

­ REQUENCY RESPONSE EVOLUTION F WITH INCREASED VARIABLE GENERATION Hydroelectric plant, biomass, and concentrated solar plant (CSP) technologies are based on synchronous generators, similar to conventional fossil‐fueled and nuclear plants. However, most variable renewable technologies, including large‐scale wind generation and PV solar generation, are not synchronously connected to the system, due to the decoupling of their AC frequency from any rotating masses. Similarly, high voltage interconnection technologies such as high‐voltage direct current (HVDC) are also not synchronized to the system. Consequently, these nonsynchronous technologies, which employ power electronic controls, do not produce the same natural response to deviations in system frequency as conventional machines would. With increasing penetration levels of nonsynchronous generation and loads on power systems worldwide, the dynamic behavior of power systems is changing[16]. Solar generation, which involves directly harnessing energy from the sun, exists in two main forms. PV panels use semiconductor solar cells to convert the electromagnetic energy in sunlight into direct current (DC) electricity. In contrast, CSP plants mirror concentrate solar thermal energy to raise steam and drive a conventional steam turbine generator. In 2013, solar generation was the second‐highest source of new capacity in the United States, second only to natural gas. Cumulative installed capacity of solar PV in 2011 was above 65 GW, with Germany and Italy accounting for approximately half of the global cumulative capacity, followed by Japan, Spain, the United States, and China. At 290 MW, the Agua Caliente Solar PV project, completed in Arizona in April 2014, is currently the largest solar installation in the world. Concentrating solar enjoys a smaller market share and is currently the most prevalent in the United States and Spain (approximately 1.5 GW). Due to its use of thermal energy (steam) and traditional turbines, and energy storage due to the associated thermal inertia, CSP can offer a certain degree of control over its available output. In contrast, PV solar generation does not involve any large rotating masses, meaning that there is no inherent stored energy coupled with the system frequency, as with conventional plants. Large‐scale wind turbines can be divided into two broad categories, which govern the behavior of wind power generation: fixed‐speed and variable‐speed technologies. Fixed‐speed wind turbines are based around induction generators, such that their rotational speed is loosely coupled to the system frequency, and, therefore, allowing them to provide an inertial

response[17]. To maximize the efficiency of the energy extraction, wind turbines should rotate faster at higher wind speeds, subject to minimum and maximum operational speed limits. Consequently, in variable‐speed wind turbines, the rotor speed depends on the wind speed rather than the system synchronous speed (frequency), and thus the rotor speed is decoupled from the system frequency. Such an approach also reduces the stresses on the turbine mechanical structures. The majority of modern wind plants, both onshore and offshore, are based on variable‐speed technology. While increasingly some variable‐speed (direct drive) turbines employ synchronous generators, the generator output passes through an AC‐DC‐AC converter, ­resulting in a decoupling of the synchronous generator’s output power frequency (and the rotating generator and ­turbine rotor masses) from the system frequency, thus resulting in the elimination of the inherent inertial frequency response behavior from the turbine. In addition to the inclusion of nonsynchronous generation, the aggregation benefits associated with large interconnections has heralded increased interconnection between adjacent systems in recent decades, predominantly through HVDC, which allows long distance submarine connections between countries. Similar to variable renewable technologies, HVDC interconnection is an inverter‐based technology, ­resulting in a decoupling of the interconnected systems. Short‐Term Frequency Response Impact As a result of the electrical decoupling of variable renewable technologies from the system frequency through power electronic controls, those frequency response services (inertial response, primary reserve, etc.) traditionally provided through conventional plants are being displaced. Consequently, power system dynamic studies have become a focus of many utilities, original equipment manufacturers (OEMs), and academic institutions around the world as the changing system response to imbalances is recognized[18–24]. When considering the impact of variable generation on short‐term power system frequency response, a number of elements of the frequency characteristic, as illustrated in Figure 2.2, can be considered. Rate of Change of Frequency (RoCoF) The initial rate of change of frequency (RoCoF) following a significant disturbance is an important metric for the system frequency response. The  greater the initial RoCoF following an event, the greater the risk of the system frequency dropping

32  Advances in Energy Systems

Disturbance (Imbalance) Occurs

System frequency (Hz)

50

Short-term frequency response

49.8

49.6

49.4

Inertial response Fast frequency response

49.2

Primary frequency response Secondary & tertiary responses

49

Time

Figure 2.2  Illustrative frequency trace following event.

Synchronous inertia (GWs)

60 50

Inertia 2009 Inertia 2020

40 30 20 10 0

0

10

20

30

40

50

60

70

80

90

100

Cumulative probability (% hours) Figure 2.3  Ireland and Northern Ireland system stored energy duration curve.

below load‐shedding thresholds and heading toward system collapse. As variable generation (VG) displaces conventional plants, it also displaces the inherent synchronous inertial response provided by such plants[25]. As a result, there is less stored rotational energy available to arrest the frequency fall following a contingency event. Figure  2.3 illustrates the stored energy duration curve of the combined Ireland and Northern Ireland system in 2020 compared to that in 2009. It can be observed that the stored energy synchronized to the system is set to reduce significantly, thus potentially leading to higher RoCoFs. In Ireland and Northern Ireland a minimum stored energy level of 25 000 MWs has been recommended[26]. This is currently an area of concern for the system where all units are required to ride through frequency variations of 0.5 Hz s−1 ­following

an ­imbalance. Similarly, in Hawaii high R ­ oCoFs have been associated with gas turbines tripping, leading to modifications to their temperature control systems[21]. Many systems employ anti‐islanding protection schemes in the form of RoCoF relays, aimed at disconnecting distribution‐connected generation in the event of part of the network being isolated from the rest of the grid[27]. In the event of high RoCoF conditions, the relay assumes that an “electrical island” condition has occurred and trips the distribution‐ connected generation for safety reasons. The relays are unable to distinguish between a high RoCoF associated with an islanding event and that due to a generator tripping, potentially leading to false triggering. Approximately half of the wind generation in Ireland is connected at the distribution level, and enabled with anti‐islanding RoCoF relay protection with a

Short‐Term Frequency Response of Power Systems with High Nonsynchronous Penetration Levels  33

0.6

RoCoF (Hz/s)

0.4 0.2 0 –0.2 –0.4 –0.6 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Wind generation level (pu) Figure 2.4  Ireland and Northern Ireland system RoCoF 2020.

trigger setting of 0.55 Hz s−1. As the stored energy of the system reduces, higher RoCoFs are likely to be experienced as a result of contingency events, which could increase the risk of tripping large quantities of wind generation from the distribution level, potentially leading to a cascade event. Figure 2.4 illustrates the maximum RoCoF that would be experienced by relays on the Ireland and Northern Ireland system following the loss of the largest single infeed (causing negative RoCoF values) or the loss of an exporting interconnector (causing positive RoCoF values) as a function of the wind generation level, for every hour, according to simulations carried out on a model of the 2020 system. It can be observed that, without intervention measures, the maximum RoCoF experienced exceeds the 0.5 Hz s−1 and even the 0.55 Hz s−1 threshold numerous times in the year, thus increasing the risk of cascade events. However, the initial RoCoF is dependent upon the stored inertial energy, as well as the size of the infeed trip. A number of solution measures have been proposed, including selectively upgrading or replacing the RoCoF relays, with, for example vector shift based relays, and increasing the ride‐through RoCoF for all generators from 0.5 to 1.0 Hz s−1[28]. With increasing levels of variable generation, the size of the largest infeed is likely to reduce, counteracting the effect of reduced system inertia, with more part‐loaded units. It is likely that as the penetration of variable, nonsynchronous technologies increases on systems worldwide, systems may develop requirements for a minimum stored energy level available on the system, or implicitly in the form of minimum numbers of conventional generating units online.

Frequency Nadir In addition to concerns relating to the initial RoCoF, a key metric of the frequency stability of a power system following an imbalance is the most extreme frequency reached. For the sudden loss of a generating infeed, this constitutes the lowest frequency before the system begins to recover, also known as the frequency nadir. For extremely low frequencies, systems have emergency safeguards that trigger load tripping, shedding customers in order to regain balance. In ­order to maintain reliability standards, a system must ensure sufficient short‐term frequency response to avoid such load shedding. Helpfully, nonsynchronous VG does not generally replace conventional plant MW for MW. For example, in the California Independent System Operator (CAISO) frequency response study, it was found that on average, for every 3 MW of wind generation production, 2 MW of conventional generation was decommitted and 1 MW of conventional generation was dispatched down[29], though this ratio may vary from system to system. In such cases, the increase in nonsynchronous generation across a system may act to improve the system frequency response capability as conventional plants, which are dispatched down, have more headroom to respond, while the system inertia level has not been severely impacted. However, as variable generation is increased further, more conventional plants are dispatched off, resulting in less system stored energy and traditional governor droop response[30]. The impact of nonsynchronous generation on the frequency response is most pronounced for smaller systems with high nonsynchronous generation penetration levels. On the New Zealand system, for

34  Advances in Energy Systems

example, it was found that future high wind penetration levels displacing conventional, high‐inertia plants could cause a significant increase in the spinning reserve requirement, resulting in increased system operational costs[31]. Similarly, lower frequency nadirs and higher RoCoFs would be expected on the combined Ireland and Northern Ireland system in 2020 as a result of increased nonsynchronous generation levels (wind production and HVDC imports) approaching 75% of instantaneous demand[32, 33]. It should be noted, however, that the issue of primary frequency response on modern power systems is not solely a result of increased nonsynchronous generation. In the Eastern interconnection in the United States, one of the two major synchronous areas in North America, there has been a notable decline in frequency response capability in recent years[34]. To date, the deterioration in frequency response has been predominantly due to wide governor dead bands, generators operating in modes that do not offer a frequency responsive reserve (e.g. sliding pressure mode), governors that are not enabled, and a reduced percentage of direct drive motor load, rather than increased nonsynchronous generation levels. A potential driving factor behind these issues has been the electricity market design, which in some areas does not incentivize frequency response, or indeed offers disincentives (whereby generators are penalized if deviating from their previously agreed schedule)[35]. Furthermore, in the CAISO system, governor withdrawal, whereby the response from conventional plants does not operate as normal or is disabled, was found to cause an approximately 20% degradation in system frequency, as measured by the North American Electric Reliability Corporation (NERC) frequency metric[29]. As with the Eastern interconnection, it was found that governor action presented a more significant impact compared to a reduction in system inertia associated with a higher penetration of renewable generation. Furthermore, given predicted levels of nonsynchronous penetration for California’s near future the frequency response of the system is generally expected to be above the NERC requirements. High Frequency Under‐frequency events, caused by the sudden tripping of a large infeed, have traditionally presented the most probable contingency event on systems. However, with increased levels of HVDC interconnection between systems, and with raised nonsynchronous penetration levels, high‐frequency events are likely to pose an increasing challenge to frequency control.

Reduced system inertia, coupled with conventional units operating at reduced outputs, results in fewer traditional sources of high‐frequency response services, that is an ability to reduce generation output. In the case of a high‐frequency imbalance, a conventional plant would have less margin for ramping back due to their minimum generation level constraints. As nonsynchronous penetration levels increase further, more conventional plants are displaced by renewable sources and the synchronous inertia of the system falls, which may result in higher rates of change of frequency following the loss of large loads. Although often seen as an issue when exporting between systems, high‐­ frequency events have also been experienced with system splitting, where subareas are exporting large amounts of power to other subareas[36, 37]. For example, on 4 November 2006, following a fault on the European interconnected transmission system (initially due to the switching out of a double circuit 380 kV line to enable a ship to travel along the Ems River to the North Sea), several high‐voltage lines were tripped, splitting the grid into three separate areas (west, northeast, and southeast). As a result, the west and southeast regions both experienced under‐ frequency events while the northeast experienced a high‐frequency event, reaching 51.4 Hz at its peak. While approximately 6200 MW of wind generation initially tripped due to the high frequency, helping to reduce the system frequency to normal levels, they soon automatically reconnected, counteracting measures to reduce the imbalance, including reducing the output or switching off thermal units and starting a pumped storage plant in pumping mode (additional load)[36]. As a further example, on 26 February 2008, mal‐ operation of protection led to a sustained 138 kV short‐circuit fault in southern Florida, and the disconnection of approximately 1350 MW of load in the proximity to the fault, the tripping of 17 generators, including two nuclear units, and under‐frequency load shedding of a further 2300 MW in the wider southern Florida peninsula. The long duration of the fault (1.7 seconds) led to frequency swings of ±0.6 Hz in SE Florida and ±0.3 Hz in NE Florida. Subsequently, six of the combined cycle gas turbine (CCGT) units in the region tripped out due to turbine combustor lean blowout: as the units accelerated in response to the (high) frequency excursion, reaching 60.64 Hz in the southeast, the directly coupled compressors forced more air into their combustion chambers, while at the same time the governor control reduced gas flow in response to the increase in frequency. The resulting loss of flame, or blowout, caused the units to trip ­offline[38].

Short‐Term Frequency Response of Power Systems with High Nonsynchronous Penetration Levels  35

­ OTENTIAL FREQUENCY RESPONSE P SOLUTIONS In response to the changing frequency response capability of power systems, support services are required, and are increasingly available, from sources other than conventional plants. In general, potential alternative sources of short‐term frequency response encompass the following.

Demand‐Side Response, Including Electric Vehicles While the traditional frequency response paradigm was designed to procure balancing services from large generating plants, many electrical loads can also respond without compromising customer energy service. Utilities around the world have operated frequency‐dependent demand management schemes for large industrial loads for many years, mainly for contingency events or for peak lopping during high demand periods[2, 39, 40]. However, an aggregation of a large number of dynamically controlled loads, e.g. refrigerators and freezers, also has the potential to provide significant added frequency support following contingency events[41]. On a similar theme, other household appliances, such as water and space heaters, making use of their thermal inertia, as well as heating, ventilating, and air conditioning (HVAC) equipment and water pumps, provide effective and cheap sources of primary frequency control[42, 43]. Decentralized approaches can provide primary frequency control, with demand being capable of responding directly to frequency errors in a manner similar to conventional generators[44]. Approximately 8 GW of the United States’ over 60 GW of demand response is managed in the Pennsylvania–New Jersey–Maryland (PJM) interconnection, in which retail electricity customers may earn a revenue stream for reducing electricity consumption in emergency situations[2, 45]. Similarly, the frequency response for the Electricity Reliability Council of Texas (ERCOT) is provided through interruptible load during emergency events. Also, in Great Britain, mechanisms exist for residential loads to provide a frequency response service: provided within two seconds of instruction and sustained for a minimum of 30 minutes, subject to a minimum capacity of 3 MW. In such cases, smaller loads can be aggregated to form more significant responses, so that the response capabilities of individual loads are not ignored. Among the benefits associated with frequency response from demand over traditional (conventional generator) approaches are its speed (can respond in very short time frames, often for less cost

than frequency response from conventional sources), potentially lower costs, its distributed nature, and it low contribution to pollution. One challenge, however, is the number of potential contributors distributed across the system, which requires consideration of the measurement and subsequent payment for the (expected/actual) service provided. With increased electrification of multiple industries, including transport, electric vehicles (EVs) could represent a significant portion of electrical demand in the future, as many countries pursue high EV penetrations to achieve various emission targets. Germany, for example, plans to increase its EV production from 2000 units currently to 1 million units by 2020 to achieve a 40% emissions reduction. In 2012, the China State Council pledged to invest over $16 billion to achieve a target of 5 million battery‐ powered EVs and plug‐in hybrids by 2020[46, 47]. EV introduction was subsidized in 10 selected cities (later increased to 25 cities) with Shenzhen, Hefei, Beijing, and Hanzhou among those demonstrating the most progress to date. The optimal charging profiles of EVs, whereby use is made of low‐electricity tariffs at times of low demand to flatten the demand profile, has been an area of considerable interest. EVs could provide a potential source of emergency frequency response following a sudden imbalance, although the response would depend on the distribution of EV load at different times[48, 49]. In the Northwest Power Pool (NWPP) in the United States, the additional balancing required for a 10 GW wind generation scenario in 2019 could be met through EVs if about 13% of the existing light‐duty vehicle stock were plug‐in hybrid EVs[50]. Storage Devices – Batteries and Flywheels Energy storage devices also provide a potential frequency response capability. Storage has traditionally been realized using fixed‐speed pumped hydro units. However, new storage technologies such as batteries, flow batteries, and compressed air energy storage (CAES) are of increasing interest, as well as emerging technologies such as high‐speed flywheels, ­ultracapacitors, and superconducting magnetic energy storage (SMES). The latter technologies have been successfully deployed or demonstrated to varying degrees in both distributed and centralized applications. There are currently two CAES plants in the world today, one in Alabama (110 MW) and the other in Germany (290 MW), while in contrast there is over 100 GW of pumped‐hydro installed. Both CAES and pumped hydro have specific geographic requirements, making their installation site dependent. In 2004, a

36  Advances in Energy Systems

2700 MW plant was proposed for Norton, Ohio, consisting of 9 × 300 MW generating units and an existing limestone mine 700 m beneath the surface with a cavern volume of 120×, providing a storage capacity of 43 GWh. These plans are still in place in 2014. Of the various storage options available to systems, batteries, and flywheels provide the most promising opportunities for short‐term frequency response due to their speed of response and suitability to shorter term rather than duration of response. In addition, these technologies are not generally subject to site‐specific requirements, allowing for development in more varied locations. Battery energy storage systems, capable of changing their output in the order of milliseconds, have been employed for frequency response in PJM in North America, Japan, and Chile, among others, allowing generator capacity to be offloaded for frequency response obligations, and responding to both high‐ and low‐frequency deviations[51]. Given the ability of battery systems to be more easily incorporated into the network it is also seen to help with network congestion, particularly in countries such as Italy, where there has been a rapid growth in PV installations in recent years[52]. Similarly, in Great Britain, battery‐based energy storage systems are proposed to be connected to the 11 kV grid, assessed on the basis of their ability to overcome variable output and frequency fluctuations caused by wind and solar farms. The conversion of electrical energy into rotational energy through a large rotating flywheel accelerated through a motor provides another source of short‐term storage. The flywheel, which is decoupled from the system frequency to operate at very high speeds (up to approx. 40 000 rpm), can be sped up or slowed down to rapidly shift energy to or from the grid, which ensures steady power delivery[53]. Again, flywheel storage is operational in PJM, with a number of projects having been developed since 2008. Conventional Plants Upgrading and adapting conventional plants can also provide support services[54]. In fact, for many systems incorporating large volumes of variable renewable technologies, operators, and regulators are also exploring the potential for existing and new conventional plants to improve their response capabilities. Furthermore, by identifying the potential for conventional plants to ride through more significant frequency excursions, the frequency requirements to maintain system reliability standards may be reexamined, as explored in Ireland. The issue of reduced system inertia due to the displacement of conventional plants can be somewhat

addressed by lowering the minimum generation level of conventional plants. For a given variable, renewable generation level system inertia can be increased by allowing more conventional units to be scheduled online. Such an approach also provides increased margins for high‐frequency responses at times of low loading of conventional plants. While the efficiency of the generator may be compromised by the reduced minimum loading level, with implications for emissions, the turbine is likely to be the limiting factor at lower loads such as turbine blades fluttering, economizer misting, or issues of mixed flow[55]. As an example of what might be achievable, for a test plant in North America an owner experimented with reducing output below 19% of net rating as opposed to 40–50% normally. By dispatching plants at lower outputs instead of decommitting them, the synchronous inertia of the system was maintained, which had the added effect of reducing the size of the largest generating infeed, thus lessening the risk of increased RoCoFs. CCGTs have been adopted at scale on many systems due to their fast start‐up times and resulting flexibility, amongst many reasons. However, particularly if operating at or near maximum output, a reduction in system frequency results in a decrease in turbine speed, which, in turn, reduces the air flow produced by the compressor, thus decreasing the gas turbine output if appropriate compensatory measures are not taken. Options include introducing fast‐acting inlet guide vanes on the compressor to rapidly increase the air flow, or spraying demineralized water into the compressor inlet to increase the mass flow rate[56]. Similarly, for steam‐based plants, condensate throttling utilizes fast valving on the steam extraction lines to the low‐pressure (LP) feedwater heaters so that an increase in power output can be sustained. The much slower responding coal pulverizers can then meet the increased firing demand without the resulting delays creating severe temperature overshoots and metalwork stress[57]. Synchronous compensators represent a traditional option, which is experiencing a revival of interest. They may be formed using decommitted or offline conventional plants, or may instead be directly made for purpose. Synchronous compensators, also known as synchronous condensers, are synchronous generators that are not attached to a prime mover and run without any load. They are typically used to generate or absorb reactive power, supporting the system voltage. However, due to their synchronized rotational speed, they can also be used to increase the stored inertial energy of the system, thus improving the initial RoCoF following an imbalance. As an

Short‐Term Frequency Response of Power Systems with High Nonsynchronous Penetration Levels  37

example, a 270 MVA synchronous compensator was installed in Denmark in 2013, the first in the country for 40 years. Ostensibly intended to provide voltage support in a weak part of the network, it was designed with a flange connected to the rotor shaft such that a flywheel could be later added, if required, to further increase the system inertia. Emulated/Synthetic Inertia Although modern variable‐speed wind turbines are mechanically decoupled from the system frequency, supplementary controls can enable them to harness the inertial energy associated with their rotating blades following a frequency transient. Temporary power injections in the order of 5–10% of rated capacity for 5–10 seconds are achievable. While the provided synthetic inertial response can be very fast (tens to hundreds of milliseconds), it is different from the inherent inertial response provided by a synchronous plant, as it relies on active controls comprising various tunable parameters, as well as the wind turbine operating conditions, such as wind speed. For turbines operating below their rated speed, harnessing their stored rotational energy through synthetic inertial controls results in a slowing down of the rotor and hence a deviation from the optimal tip‐speed ratio and a reduction in the aerodynamic energy captured by the turbine. As a result, the initial power injection from the turbine is followed by an energy recovery phase, whereby the turbine output is less than its set‐point, as the turbine tracks back to optimal efficiency by restoring the blade rotational speed. While the inertial response from a conventional plant is also followed by an energy recovery phase, synchronous plants regain

the rotational energy delivered in sympathy with the system frequency recovery. For turbines operating above rated wind speed prior to the event, their blades can be pitched back into the wind to reduce or avoid the energy recovery phase. Meanwhile, turbines operating close to their minimum operational speed should not provide a synthetic inertial response in order to avoid stall conditions. Following the above principles, Figure 2.5 illustrates frequency injection field tests for a GE 1.5 MW wind turbine. Consequently, at low wind speeds, negligible response is provided to avoid stall conditions, with larger responses and associated energy recovery periods with increasing wind speeds. Above rated wind speed, the energy recovery is avoided as the turbine blades are pitched to harness previously untapped energy. Reliance on active power controls, whether it be for wind turbines, HVDC interconnections, or demand‐ side measures, for the detection of disturbances and triggering of plant responses introduces inherent delays into the provision of the response. Another key difference between conventional and nonsynchronous plants is the manner in which the response is triggered. While a synchronous inertial response is proportional to the RoCoF, the trigger signal for the synthetic inertial response from wind generation is not limited to this method. Several control approaches have been proposed by both industry and academia, which can broadly be separated into three main categories. The first is based on the rate of change of system frequency, similar to conventional plant. However, the presence of noise implies that measurement of the RoCoF through electronic monitoring equipment can be difficult in reality, and such an approach is not widely employed. Consequently, usage of the

2000

Power output (kW)

14 m/s 11.5 m/s

1500

10 m/s 1000

8 m/s

500

5 m/s 0

0

10

20

30

40

50

Time (seconds) Figure 2.5  Emulated inertia response shape and variation with wind speed (GE field tests).

60

38  Advances in Energy Systems

term emulated inertia or synthetic inertia is somewhat of a misnomer. The second control option provides a fixed power injection signal following a deviation in frequency beyond a certain threshold. While such an approach implies a degree of certainty in its shape, the presence of such a static, invariant response at high penetration levels may result in a response larger or smaller than required, and an extended energy recovery phase, depending on the size of the event and the number of turbines responding[58]. The third control approach is based on a response proportional to the deviation in frequency from nominal, considering a certain deadband. The optimal control approach that is most likely to be adopted at scale is not yet known, but is an area of active research[59]. A PV solar plant does not possess any stored rotational energy and so would require an external storage device in order to achieve an emulated inertial response, while CSP can provide inertial energy through its steam turbine and synchronous generator. Some OEMs have also begun to promote “hybrid” wind turbines that come with an integral battery. The turbine’s battery can store the equivalent energy of less than one minute of the turbine operating at full power, but by pairing the battery with advanced wind‐ forecasting algorithms, wind farm operators could stabilize a certain amount of power output for up to an hour[60]. Furthermore, such storage could allow for the avoidance of the energy recovery phase associated with synthetic inertial controls, which could dampen the impact of a large number of wind turbines recovering concurrently following an event.

4% and 40%). There may well be knock on impacts for the quantification of the systemwide droop response capability. Furthermore, the droop response provided by nonsynchronous sources normally incorporates a dead band to avoid false triggering. However, despite the above complications, due to the highly tunable nature of power electronic controls, and the absence of delays associated with fuel and air supply in conventional plant, the droop response from nonsynchronous generation can offer a considerably faster response to that from conventional plant. As well as increasing their power output in response to a loss of generation, wind and solar generation can also provide a droop response which reduces plant output during a high‐frequency event. Again, due to the tunability of the response the droop characteristic of nonsynchronous generation (choosing points A–D) can be asymmetrical, that is having a different slope (droop) for high‐ and low‐frequency events, as illustrated in Figure 2.6. Furthermore, the droop calculation may be based upon the boundary of the deadband or from the center‐point of the deadband, resulting in a stepped response when the threshold is exceeded. The inclusion of a droop response from wind generation can significantly improve the frequency response of systems[63, 64]. Figure  2.7 illustrates the improved simulated frequency response of the 2010 Californian (CAISO) system incorporating droop response and emulated inertia from wind plant, and, in particular, the fast‐acting synthetic inertial response significantly raises the frequency nadir, although the recovery to steady‐state is slightly slower.

Nonsynchronous Generator Droop Response Nonsynchronous variable generation, such as wind turbines and PV solar generation, can also provide a droop response similar to that from conventional plants. However, in order to respond to a low‐frequency excursion, VG plants must be dispatched down from their maximum available power to allow headroom for an upward response. While the droop response from conventional plant is generally calculated based on the rated power of the unit, and is assumed linear across the operational range, the droop response defined for wind turbines or solar plant may vary significantly with system requirements[61, 62]. A number of droop responses from nonsynchronous generation have been proposed for different systems, including those in which the droop is based on the available generating capacity at the time of the event. Such an approach implies that the droop, as normally calculated, varies significantly across the operating range of the wind turbine (perhaps varying between

­ RID CODE REQUIREMENTS G AND ANCILLARY SERVICE MARKETS With the continued displacement of the frequency response capabilities from conventional plants, there is a shift from considering such services as a natural response toward an ancillary service which may have to be procured in the future[66]. Potential mechanisms for procuring such services are currently of significant interest in a number of jurisdictions, with options ranging from additional grid code requirements and ancillary service market mechanisms to ensure sufficient short‐term frequency response capability. Table  2.2 highlights the state of play of frequency response requirements and incentives in New Zealand, Ireland, and Northern Ireland, Hydro Quebec and ERCOT, e.g. Refs.[67–70]. Grid codes, which mandate the response behavior of all grid participants, have undergone considerable revision in systems with high variable renewable

Short‐Term Frequency Response of Power Systems with High Nonsynchronous Penetration Levels  39

Power output (pu)

1

A

0.6

C

B

0.8 Stepped response

0.4 0.2 0 0.99

1

1.01

System frequency (pu) Figure 2.6  Illustrative droop characteristic. 60.1 No wind response 60

Wind droop response Wind droop and emulated inertia response

Frequency (Hz)

59.9 59.8 59.7 59.6 59.5 59.4

0

5

10

15

20

25

30

35

40

Time (seconds) Figure 2.7  Simulated frequency response of the CAISO system incorporating droop response and emulated inertia from a wind plant[65].

p­ enetrations in recent years. Many have adapted to require frequency response capability from variable renewable technology similar to that of a conventional plant[71]. Hydro‐Quebec has already defined an inertial response capability requirement for large wind power plants, stating that they must have an equivalent inertial capability to that of a conventional plant with an inertial constant, H, of 3.5 seconds, with an amendment to this requirement currently under review[72]. In Puerto Rico, minimum technical requirements include a droop response from wind generation, and it has been further recommended that specification of the type of droop required by wind and PV plant be defined more clearly, in recognition of the fact that the

droop response from such plants can be varied from that provided by a conventional plant[73]. ERCOT include primary frequency response requirements from their inverter‐based (variable renewable) generation, including a droop response. In many other systems, including China, Ireland, and Great Britain, active power control requirements from wind plants have been included in the grid code[61, 74–76]. As part of ongoing grid code updates, transmission system operators (TSOs) are also reviewing the requirements placed on, and compliance of, conventional plants to ensure equality of the burden for system reliability. In addition to grid codes, which may only  require a  minimum required capability, market mechanisms

40  Advances in Energy Systems

Table 2.2 Frequency response products.

Peak load (2012/2013) Renewable penetration Variable generation (VG) penetration VG primary frequency control

New Zealand

Ireland and N. Ireland

Hydro Quebec

ERCOT

6000 MW 60% 4%

6500 MW 18% 17%

38 800 MW 98% 5%

67 245 MW 8% 8%

Response within 1 s and sustained for 60 s. Called fast instantaneous reserve

Capability required in grid code

All generating units above 10 MW must have capability

Required for over‐frequency events and under frequency if not operating at max available capacity. 5% droop SIR product proposed (future concept)

Capability required from wind plant in grid code. Equivalent to inertia response of synchronous generation with H constant 3.5 s

Proposed (future concept)

Synchronous inertia product (from synchronous units) Fast‐acting frequency response

Over frequency response

Short‐term frequency response in ancillary service market

SIR product proposed Product proposed between 2 and 10 s

Over frequency reserve implemented by over frequency tripping scheme Yes

Over frequency generation shedding

Yes

Short‐Term Frequency Response of Power Systems with High Nonsynchronous Penetration Levels  41

play a significant role in frequency response behavior. While most systems require that all conventional plants provide a governor droop response, in a number of interconnections in North America it has been shown that a governor droop response is actually disincentivized with current market operations. Units that deviate significantly from their dispatch in the settlement market are penalized for doing so[77]. Possible solutions include amendments to those market mechanism which act as a disincentive for frequency response or further, to actually incentivize frequency response from plant[35]. Further to new grid code requirements, market mechanisms and incentive schemes, in which short‐ term frequency response services are traded, are also being developed[78]. Few regions in the world have ancillary service markets for such services, likely due to the fact that such services were traditionally provided by conventional plant without the need for in‐market payment. Large interconnections such as the Eastern and Western interconnections in North America, in their current form, do not urgently require the introduction of ancillary service markets for short‐ term frequency response because sufficient volumes of these resources already exist[79]. However, such incentives and markets may become more important with increasing nonsynchronous penetration levels and further displacement of conventional frequency response resources. In fact, the procurement of ancillary services through market mechanisms has already proven effective in a number of systems. In Australia, frequency control ancillary service (FCAS) market mechanisms, introduced in 2001, have generally been regarded as a positive intervention, as they have resulted in an overall reduction in ancillary service costs compared to non‐market arrangements, while maintaining frequency standards. The ancillary services introduced include a market for primary response, requiring full response within 6 seconds and to be sustained for 60 seconds. The inertial response is normally considered as being inherent to a synchronous machine based ­ generator, and hence typically is not actively recognized as part of unit commitment and reserve allocation procedures. However, in smaller systems, such as Ireland and Northern Ireland, proposals exist for it to be linked with the minimum generation levels of online units, that is if generators are encouraged (through ancillary service payments) to reduce their minimum stable output then, particularly for low system demand levels, more units can be scheduled online to meet the demand, raising the rotational stored energy of the system, and improving the overall inertial response. In recognition of the

fundamental changes that variable renewable generation has introduced to system frequency behavior, a number of systems have also designed new tradable products to ensure and enhance frequency stability[80]. In Ireland and Northern Ireland, and ERCOT, new suites of ancillary services, including synchronous inertia response (SIR) and FFR have been proposed. Similarly, New Zealand has developed a market for fast‐acting response (within 1 second and sustained for 60 seconds). In most systems employing new products there are no barriers to the provision of the newly defined frequency control ancillary services by renewables which can meet the technical requirements[81].

I­ SSUES RESULTING FROM NONSYNCHRONOUS FREQUENCY RESPONSE The availability of a dynamic capability from nonsynchronous technologies, such as those outlined in the previous section may have a significant impact on the behavior of future systems. The extent of their influence must be fully understood to ensure that stability and reliability standards are maintained. In particular, the impact of deadbands, static responses, response tuning, active power response forecasting, and coordination should be considered. The integration of responses which have been tuned by active controls has the potential to alter the operation of systems considerably from that traditionally experienced, as the contribution from conventional generation gradually reduces. While the provision of a droop response through nonsynchronous generation potentially allows for a faster response, how it is employed will have a significant impact on the overall system frequency response capability. As a result, extreme care should be taken to ensure that the technology is implemented effectively and robustly, and that any undesirable consequences of this shift in operational behavior are predicted and avoided. For example, the inclusion of a stepped droop response at the deadband boundary, from either a generation‐ or demand‐based response, as illustrated for a simulated high‐frequency event in Figure  2.8, could result in oscillatory behavior as the response is activated and deactivated near the deadband threshold[82]. While such responses from individual variable renewable plant may not have a considerable systemwide impact, policies must recognize that the aggregate impact with increased variable renewable penetration levels may not be negligible. Fast transient frequency support, via a controlled inertial response from wind turbines, fast acting load

42  Advances in Energy Systems

50.3 50.25

Frequency (Hz)

50.2 50.15 50.1 50.05 50 49.95 49.9

0

4

8

12

16

20

24

28

32

36

40

Time (s) Figure 2.8  Simulated frequency response following disturbance with units having a stepped droop curve governor response.

60.05 60 59.95

Frequency (Hz)

59.9 59.85 59.8 59.75 59.7

No wind response

59.65

Synth. inertia only Droop only

59.6 59.55

Synth inertia & droop 0

10

20

30

40

50

60

Time (sec) Figure 2.9  Impact of large aggregate energy recovery from active controls on system frequency response. Western interconnection response for 40% wind power penetration[83].

response, or the injection of power from energy storage devices all help to improve the under‐frequency load shedding margin[29, 58]. In the case that a fast‐acting frequency response is provided through synthetic inertia or demand, the impact of the aggregate recovery of this short‐term response must be considered as well as the initial power increase, as illustrated in Refs.[59, 83]. Too large a short‐term response, caused by overly aggressive tuning, coupled with a high p­ enetration

of the technology could result in a significant energy recovery as energy is paid back to variable speed wind turbines or solar plant, as illustrated in F ­ igure  2.9. This significant draw of power could result in a further power imbalance, potentially causing a double dip in the system frequency. As the actively controlled responses from nonsynchronous generation do not necessarily recover in sympathy with the system frequency, the tuning of active ­controls and the scale

Short‐Term Frequency Response of Power Systems with High Nonsynchronous Penetration Levels  43

at which this technology is rolled out must consider the full implications of their impact. Similarly, care must be taken with the widespread inclusion of static or invariant frequency responses, subject to defined activation thresholds. According to rules put in place in Germany for generators connected to the low‐ voltage distribution network, an immediate shutdown was required following the system frequency reaching 50.2 Hz. As the installed power of PV systems has increased significantly in recent years, concerns arose, particularly at times of high solar generation, that the system was at risk of losing several GW of generation at once in extreme cases. The resulting power imbalance could have exceeded the primary frequency control reserve available, thus leaving the system vulnerable to instability. In January 2012, an obligatory measure mandating new frequency settings was introduced and retrofitting is due to be completed in 2014 at a cost of over €175 million. When considering a response from wind generation, it should be noted that it will vary as a function of the wind speed, which is unlikely to be deterministic during the transient period following a significant frequency deviation. As a result, the prediction of the frequency response of the system becomes less certain. Furthermore, priority is given to the load management functions of the wind turbine control and thus the synthetic inertial response is also dependent on the wind turbulence and mechanical states of the drive train and tower. For example, an individual turbine might transiently be running at a speed greater than the steady‐state condition associated with a particular power level, because of a gust coinciding with the grid event[84]. While the inertial performance of an individual wind turbine is stochastic in practice due to short‐term variability, the aggregation of the response across a system filters much of this variability. However, the stochasticity of individual turbine responses is likely to cause difficulty in the implementation of “availability of service” payment mechanisms used in ancillary service markets. While variable renewable generators and loads can indeed provide a frequency response, the actual value of this service must be identified in order for it to be accommodated in normal system operation and future ancillary service markets. At the outset, it is important to recognize the limitations of various technologies. For example, without additional expense, sustained responses cannot be provided by nonsynchronous generation operating below rated conditions in normal (not curtailed) operation. As such, the value of fast‐acting temporary responses such as synthetic inertia lies in arresting the RoCoF in the time frame subsequent to the initial RoCoF, as well as i­mproving

the nadir. In light of these response characteristics, it is important that any incentive or requirement for synthetic inertial controls from wind generation, for example, is designed for the rapid delivery of the ­service, in recognition of its true value to the system.

­CONCLUSIONS The short‐term frequency response of power systems with high nonsynchronous penetration levels is an area of increasing interest. As nonsynchronous generation levels increase the nature of the response to sudden power imbalances is changing. Dynamic frequency behavior in the future will change as traditional sources from conventional generation are displaced by variable renewable technologies and other mechanically decoupled elements. Nonsynchronous generation can provide frequency responsive services such as synthetic inertia and a droop response, however, such responses cannot be treated as a direct replacement for synchronous inertia or a governor response. They are inherently different to traditional plant capabilities due to their reliance on active controls and a variable and uncertain fuel source and so may result in new frequency response behavior. With this in mind, it is important that the portfolio of support services from all contributing technologies be considered in their totality to ensure continued high levels of system reliability. The characteristics of individual frequency responsive technologies are relatively well understood through field testing and real world experience. However, there are a number of outstanding, pertinent challenges associated with the composition of future systemwide frequency response portfolios. While it is understood that a variety of technologies can provide supportive services following imbalances, the management and coordination of these services sourced from a variety of devices with distinct individual characteristics requires further investigation. Related to this issue is the potential for the frequency response capability to be forecast, as system operators (and ancillary service markets) must ensure sufficient response capability to maintain reliability standards. While active power responses have traditionally been viewed as a deterministic resource, its provision through variable renewable technologies may incur some stochasticity, which must be managed to maintain system reliability, that is ensure sufficient reserve in the event of a contingency at least cost. Online prediction of such a dynamic response from the portfolio of sources available, if achievable, would avoid such uncertainty.

44  Advances in Energy Systems

Another outstanding question is whether services should be procured through grid codes or market mechanisms, or whether different approaches are appropriate for different systems, perhaps due to the amount of response capability already available from conventional plant, or the structure of existing electricity markets. While it is generally accepted that the frequency response is changing, the definition of  what constitutes “acceptable” behavior requires clarification: given the configurability of modern (nonsynchronous generation) controls, simple metrics such as improving the frequency nadir are no longer fit for purpose. Suitable definition of the objective is important with respect to the value that may be placed upon various support services. It must also be noted that short‐term frequency response issues cannot be answered in isolation. The increased nonsynchronous penetration levels are having impacts on many aspects of modern power systems, including voltage and transient stability, fault levels, and conventional plant cycling, to name but a few. When considering the future technology portfolio for ancillary services, a broad range of requirements and constraints should thus be considered to ensure the optimal solution.

­REFERENCES 1. International Energy Agency (2012). Key World Energy Statistics 2012. Paris, France: IEA. 2. Fox, B., Flynn, D., Bryans, L. et al. (2014). Wind Power Integration: Connection and System Operational Aspects. Herts, UK: IET. 3. Kundur, P. Power System Stability and Control, Electric Power Research Institute Power System Engineering Series. New York: MacGraw‐Hill. 4. Tavakoli, R., Power, M. Ruttledge, L., and Flynn, D. (2012). Load inertia estimation using white and grey‐box estimators for power systems with high wind penetration. IFAC Proc 45(21), 399–404. 5. Moura, F., Camacho, J., Guimaraes, G., and Chaves, M. (2012). Steam turbines under abnormal frequency conditions in distributed generation systems. In: Mechanical Engineering, 387–388. INTECH. 6. ERCOT. ERCOT nodal operating guidelines. Section 2: system operations and control requirements, 2014. 7. ENTSO‐E. Load‐frequency control and reserves, 2013. Available at https://www.entsoe.eu/fileadmin/user_ upload/_library/resources/LCFR/130628‐NC_LFCR‐ Issue1.pdf. (Accessed August 1, 2014). 8. EirGrid. Operating security standards, 2010. Available at:http://www.eirgrid.com/media/Operational%20 Security%20Standards.pdf. (Accessed May 1, 2014). 9. EURELECTRIC; ENTSO‐E. Deterministic frequency deviations: root causes and proposals for potential solutions, 2011. 10. European Renewable Energy Council. EU tracking roadmap: keeping track of renewable energy targets

towards 2020, 2013. Available at: http://keepontrack.eu/ contents/publicationseutrackingroadmap/roadmap_ finalversion3.pdf. (Accessed July 7, 2014). 11. GlobalData. Asia‐Pacific Renewable Energy Policy Handbook, 2014. Available at: http://store.globaldata. com/market‐reports/power/asiapacific‐renewable‐energy‐ policy‐handbook‐2014#.VFeuBvmsV8E. (Accessed June 1, 2014). 12. Renewable Energy Policy Network for the 21st Century (REN21). Renewables 2014: Global Status Report, 2014. Available at: http://www.ren21.net/portals/ 0Zdocuments/resources/gsr/2014/gsr2014_full%20 report_low%20res.pdf. (Accessed August 1, 2014). 13. IEA (2013). Energy policies of IEA countries. Sweden. 14. New Zealand Government. Renewable energy in New  Zealand. Available at: http://www.med.govt.nz. (Accessed July 7, 2014). 15. Soder, L., Abildgaard, H., Estanqueiro, A. et al. (2012). Experience and challenges with short‐term balancing on European systems with large share of wind power. IEEE Trans. Sustainable Energy 3: 853–861. 16. Ackermann, T. (2012). Wind Power in Power Systems, 2e. Chichester, UK: Wiley. 17. Kennedy, J., Fox, B., Littler, T., and Flynn, D. (2011). Validation of fixed speed induction generator models for inertial response using wind farm measurements. IEEE Trans. Power Syst. 26: 1454–1461. 18. Baily M. Provision of frequency responsive power reserve from disconnectable load. In: IEEE Colloq Economic Provision of a Frequency Responsive Power System Service, London, 1998. 19. Doherty, R., Mulane, A., Lalor, G. et  al. (2010). An assessment of the impact of wind generation on system frequency control. IEEE Trans. Power Syst. 25: 452–460. 20. Dudurych I. Statistical analysis of frequency response of island power system under increasing wind penetration. In: IEEE Power and Energy Society General Meeting, Minneapolis, 2010. 21. Dangelmaier L. System frequency performance of the  Hawaii Electric Light system. In: IEEE Power and Energy Society General Meeting, San Diego, CA, 2011. 22. Miller, N., Loutan, C., Shao, M., and Clark, K. (2013). Emergency response US system frequency with high wind penetration. IEEE Power Energ. Mag. 63–71. 23. NERC. Accommodating high levels of variable generation, 2009. Available at: http://www.nerc.com/files/ ivgtf_report_041609.pdf. (Accessed April 1, 2014). 24. GE Energy. Western wind and solar integration study, 2010. Available at: http://www.nrel.gov/docs/fy10osti/ 47434.pdf. (Accessed May 1, 2014). 25. Sharma S, Huang S, Sarma N. System inertial frequency response estimation and impact of renewable resources in ERCOT interconnection. In: IEEE Power and Energy Society General Meeting, San Diego, CA, 2011. 26. EirGrid and SONI. All Island TSO Facilitation of Renewables Studies, 2010. Available at: http://www. eirgrid.com/renewables/facilitationofrenewables. (Accessed January 5, 2014).

Short‐Term Frequency Response of Power Systems with High Nonsynchronous Penetration Levels  45

27. ESB International. ESB Networks embedded generation interface protection. Rate of change of frequency setting for embedded generators, 2012. 28. Commission for Energy Regulation.Rate of change of frequency modification to the grid code: decision paper, 2014. Available at: http://www.cer.ie/docs/000745/ CER14081%20R0C0F%20Decision%20Paper%20‐%20 FINAL%20FOR%20PUBLICATION.pdf. (Accessed May 1, 2014). 29. Miller N, Shao M, Venkataraman S. California ISO frequency response study, 2011. Available at: http://www. caiso.com/documents/report‐frequencyresponsestudy. pdf. (Accessed May 1, 2014). 30. Tielens P, Ergun H, Van Hertem D. Techno‐economic analysis of large‐scale integration of solar power plants in the European grid. In: 2nd Solar Integration Workshop, Lisbon, Portugal, 2012. 31. Pelletier M, Phethean M, Nutt S. Grid code requirements for artificial inertia control systems in the New Zealand Power System. In: IEEE Power and Energy Society General Meeting, San Diego, 2012. 32. EirGrid and SONI. All island TSO facilitation of renewables studies, 2010. Available at: http://www.eirgrid. com/media/Renewable%20Studies%20V3.pdf. (Accessed May 1, 2014). 33. O’Sullivan, J., Rogers, A., Flynn, D. et  al. (2014). Studying the maximum instantaneous non‐synchronous generation in an island system: frequency stability challenges in Ireland. IEEE Trans. Power Syst. 29: ­ 2943–2951. 34. Ingleson J, Allen E. Tracking the Eastern Interconnection frequency governing characteristic. In IEEE PES General Meeting, Minneapolis, MN, 2010. 35. Ela E, Milligan M, Kirby B, Tuohy A, Brooks D. Alternative approaches for incentivising the frequency responsive reserve ancillary service, 2012. Available at: http://www.nrel.gov/docs/fy12osti/54393.pdf. (Accessed June 1, 2014). 36. UCTE. System disturbance on 4 November 2006: Final report. Available at: https://www.entsoe.eu/fileadmin/ user_upload/_library/publications/ce/otherreports/ Final‐Report‐20070130.pdf. (Accessed June 1, 2014). 37. EirGrid and SONI. Northern Ireland system separation studies, 2012. Available at: http://www.eirgrid.com/ media/Northern%20Ireland%20System%20Seperation%20 Studies%202012.pdf. (Accessed June 1, 2014). 38. NERC. Industry Advisory: turbine combustor lean blowout, 2008. Available at: www.nerc.com. (Accessed July 5, 2014). 39. Xu, Z., Ostergaard, J., and Togeby, M. (2011). Demand as frequency controlled reserve. IEEE Trans. Power Syst. 26: 1062–1071. 40. NREL. Western wind and solar integration study: phase 1, 2010. Available on: http://www.nrel.gov/docs/fy10osti/ 47434.pdf. (Accessed May 1, 2014). 41. Short, J., Infield, D., and Freris, L. (2007). Stabilization of grid frequency through dynamic demand control. IEEE Trans. Power Syst. 22: 1284–1293. 42. Samsarakoon, K., Ekanayake, J., and Jenkins, N. (2012). Investigation of domestic load control to provide

p­ rimary frequency response using smart meters. IEEE Trans. Smart Grid 3: 282–292. 43. Qazi H, Flynn D. Power balance provision through coordinated control of modern storage heater load. In: 12th Wind Integration Workshop, London, UK, 2013. 44. Molina‐Garcia, A., Bouffard, F., and Kirschen, D. (2011). Decentralized demand‐side contribution to primary frequency control. IEEE Trans. Power Syst. 26: 411–419. 45. De Martini P. DR 2.0  –  a future of customer response, 2013. Available at: http://www.demandresponsesmartgrid. org/Resources/Documents/FINAL_DR%202.0_13. 07.08.pdf. (Accessed May 1, 2014). 46. Cheung K. Integration of renewables: status and challenges in China, 2011. Available at: http://www.iea.org/ publications/freepublications/publication/Integration_ of_Renewables.pdf. (Accessed May 1, 2014). 47. Tao W. Recharging China’s electric vehicle policy. In: Carnegie‐Tsinghua Center for Global Policy, Policy Outlook, 2013. 48. Mu, Y., Wu, J., Ekanayake, J. et al. (2013). Primary frequency response from electric vehicles in the Great Britain power system. IEEE Trans. Smart Grid 4: 1142–1150. 49. Keane E, Flynn D. Potential for electric vehicles to provide power system reserve. In: IEEE PES Innovation Smart Grid Technologies, Washington DC, 2012. 50. Tuffner F, Kintner‐Meyer M. Using electric vehicles to meet balancing requirements associated with wind power, US Department of Energy, 2011. 51. Hsieh E, Johnson R. Frequency response from autonomous battery energy storage, CIGRE US National Committee, 2012. 52. European Photovoltaic Industry Association. Connecting the sun: solar photovoltaics on the road to large‐scale grid integration, 2012. Available at: http://www.epia. org/news/publications. (Accessed May 1, 2014). 53. U.S. Department of Energy. Grid energy storage, 2013. Available at: http://energy.gov/sites/prod/files/2013/12/ f5/Grid%20Energy%20Storage%20December%20 2013.pdf. (Accessed May 1, 2014). 54. RWE. The need for smart megawatts: power generation in Europe: facts and trends, 2009. Available at: http:// www.rwe.com/web/cms/mediablob/en/1031648/data/ 213096/2/rwe/investor‐relations/events/archive‐2009/ Charts.pdf. (Accessed May 1, 2014). 55. Cochran J, Lew D, Kumar N. Flexible coal: evolution from baseload to peaking plant, 2013. Available at: http://www.nrel.gov/docs/fy14osti/60575.pdf. (Accessed August 1, 2014). 56. Balling L, Siemens. Fast cycling and rapid start‐up: new generation of plants achieves impressive results, 2010. Available at: http://www. http://modernpowersystems. com. Available at: http://m.energy.siemens.com/hq/pool/ hq/power‐generation/power‐plants/gas‐fired‐power‐ plants/combined‐cycle‐powerplants/Fast_cycling_and_ rapid_start‐up_US.pdf. (Accessed May 1, 2014). 57. Weitzel P. A steam generator for 700C to 760C. Advanced ultra‐supercritical design and plant arrangement: what stays the same and what needs to be

46  Advances in Energy Systems

­different. In: Seventh International Conference on Advances in Materials Technology for Fossil Fuel Plants, Waikaloa, 2013. 58. Ruttledge, L., Miller, N., O’Sullivan, J., and Flynn, D. (2012). Frequency response of power systems with variable speed wind turbines. IEEE Trans. Sustainable Energy 3: 683–691. 59. Ruttledge L, Flynn D. Emulated inertial response from wind turbines: the case for bespoke power system optimisation. In: Wind Integration Workshop, Lisbon, Portugal, 2012. 60. Bullis K. Wind turbines, battery included, can keep power supplies stable, 2013. Available at: http://www. technologyreview.com/news/514331/wind‐turbines‐ battery‐included‐can‐keep‐power‐supplies‐stable. (Accessed July 7, 2014). 61. EirGrid. The Grid Code v5, 2013. Available at: http:// www.eirgrid.com/media/GridCodeVersion5.pdf. (Accessed August 1, 2014). 62. Siemens. Active power control in Siemens wind turbines, 2011. Available at: http://www.nrel.gov/ ­ electricity/transmission/pdfs/nelson.pdf. (Accessed May 1, 2014). 63. Ruttledge L, Flynn D. System‐wide contribution to frequency response from variable speed wind turbines. In: IEEE Power and Energy Society General Meeting, San Diego, CA, 2012. 64. Miller N, Clark K. Advanced controls enable wind plants to provide ancillary services. In: IEEE Power and Energy Society General Meeting, Minneapolis, MN, 2010. 65. Miller N, Clark K, Shao M. Frequency responsive wind plant controls: impact on grid performance. In: PES General Meeting, San Diego, CA, 2011. 66. Milligan M, Donohoo P. Operating reserves and wind power integration: an international comparison. In: 9th Wind Integration Workshop, Quebec, Canada, 2010. 67. ERCOT. ERCOT concept paper: future ancillary ­services in ERCOT, 2013. 68. Hydro‐Quebec TransEnergie. Technical requirements for the connection of power plants to the Hydro Quebec transmission system, 2009. Available at: http://www. hydroquebec.com/transenergie/fr/commerce/pdf/exigence_ raccordement_fev_09_en.pdf. (Accessed May 1, 2014). 69. EirGrid. DS3: system services review TSO recommendations, 2012. Available at: http://www.eirgrid.com/ media/SS_May_2013_TSO_Recommendations_ Summary_Paper.pdf. (Accessed May 1, 2014). 70. KEMA. System services international review: market update, 2011. Available at: http://www.eirgrid.com/ media/System%20Services%20International%20 Review%20‐%20Final.v2.pdf. (Accessed May 1, 2014). 71. ENTSO‐E network code for grid connection applicable to all generators, 2012. Available at: https://www.entsoe.eu/major‐ projects/network‐code‐development/requirements‐for‐ generators/Pages/default.aspx. (Accessed August 1, 2014).

72. Brisbois J, Aubut N. Wind farm inertia emulation to ­fulfill Hydro‐Quebec’s specific need. In: IEEE Power and Energy Society General Meeting, Detroit, 2011. 73. Gevorgian V, Booth S. Review of PREPA technical requirements for interconnecting wind and solar generation, 2012. Available at: http://www.nrel.gov/docs/ fy14osti/57089.pdf. (Accessed August 1, 2014). 74. National Grid Electricity Transmission plc. The grid code Issue 5, 2014. Available at: http://www2. nationalgrid.com/UK/Industry‐information/Electricity‐ codes/Grid‐code/The‐Grid‐code. (Accessed November 1, 2014). 75. General Administration of Quality Supervision, Inspection and Quarantine of the People’s Republic of China. Technical rule for connecting wind farm into power system, 2011. Standard number: GB/T‐19963. 76. Jiang, L., Chi, Y., Qin, H. et al. (2011). Wind energy in China: status and prospects. IEEE Power Energ. Mag. 9: 36–46. 77. Ela, E., Tuohy, A., Milligan, M., Kirby, B., and Brooks, D. (2012). Alternative approaches for incentivising ­frequency responsive reserve. Electr. J. 25: 88–102. 78. Rebours, Y., Kirschen, D., and Trotignon, M. (2007). Fundamental design issues in markets for ancillary ­services. Electr. J. 20: 26–34. 79. Ela, E., Gevorigan, V., Tuohy, A. et al. (2013). Market designs for the primary frequency response ancillary service ‐part II: case studies. IEEE Trans. Power Syst. 29: 432–440. 80. Thorncraft S, Outhred H. Experience with market‐based ancillary services in the Australian national electricity market. In: IEEE PES General Meeting, Tampa, FL, 2007. 81. Transpower New Zealand Limited. Electricity industry participation code: ancillary services procurement plan, 2013. Available at: http://www.ea.govt.nz/code‐and‐ compliance/the‐code/documents‐incorporated‐into‐the‐ code‐by‐reference. (Accessed May 1, 2014). 82. Niemeyer S. Comparison of governor deadband & droop settings of a single 600  MW unit, 2010. Available  at: http://www.texasre.org/Lists/Calendar/ Attachments/243/3%20‐%20Comparison%20of%20 Governor%20Deadband%20Settings%20Feb%202010. pdf. (Aaccessed 1 June 1, 2014). 83. National Renewable Energy Lab, University of Colorado, Electric Power Research Institute. Active Power Controls from Wind Power: Bridging the Gaps, 2014. Available at: http://www.nrel.gov/docs/fy14osti/ 60574.pdf. (Accessed June 1, 2014). 84. Wu, L. and Infield, D. (2013). Towards and assessment of power system frequency support from wind plant: modelling aggregate inertial response. IEEE Trans. Power Syst. 28: 2283–2291.

3

Technical Impacts of High Penetration Levels of Wind Power on Power System Stability Damian Flynn1, Zakir Rather1, Atle Rygg Årdal2, Salvatore D’Arco2, Anca D. Hansen3, Nicolaos A. Cutululis3, Poul Sorensen3, Ana Estanqueiro4, Emilio Gómez‐Lázaro5, Nickie Menemenlis6, Charles Smith7 and Ye Wang8 School of Electrical and Electronic Engineering, University College Dublin, Dublin, Ireland 2  SINTEF Energy Research, Trondheim, Norway 3  Department of Wind Energy, Technical University of Denmark, Roskilde, Denmark 4  Laboratorio Nacional de Energia e Geologia (LNEG), UESEO, Lisbon, Portugal 5 Renewable Energy Research Institute, University of Castilla la Mancha, Albacete, Spain 6  Electrical Systems and Mathematics, IREQ, Varennes, Canada 7  Utility Variable Generation Integration Group (UVIG), Kitty Hawk, NC, USA 8  EdF Research & Development, Paris, France

1 

With increasing penetrations of wind generation, based on power‐electronic converters, power systems are transitioning away from well‐understood synchronous generator‐based systems, with growing implications for their stability. Issues of concern will vary with system size, wind penetration level, geographical distribution and turbine type, network topology, electricity market structure, unit commitment procedures, and other factors. However, variable‐speed wind turbines, both onshore and connected ­offshore through direct current (DC) grids, offer many control opportunities to either replace or enhance existing capabilities. Achieving a complete under­ standing of future stability issues, and ensuring the effectiveness of new measures and policies, is an iterative procedure involving portfolio development and flexibility assessment, generation cost simulations, load flow, and security analysis, in addition to the ­stability analysis itself, while being supported by field demonstrations and real‐world model validation.

­INTRODUCTION Wind energy is being rapidly integrated into many power systems across the globe, with a total installed capacity of 370 GW, and with 51 GW added in 2014 alone[1]. As the penetration of wind generation increases, the impact on power system dynamics is becoming increasingly apparent, and will become a more integral part of system planning and renewables integration studies[2]. Historically, power systems have been based on large, synchronous generators connected to a strongly meshed transmission network, with the dynamic characteristics of such systems ­being well understood. However, renewable generation, ­particularly in the form of wind and solar generation, is increasingly universally connected via power electronics interfaces, may well be connected to the distribution network, or weaker parts of the network, may offer new control capabilities, and, of course, is subject to the variability and uncertainty associated

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

48  Advances in Energy Systems

with local and regional weather patterns[3, 4]. The time variability and nondispatchable nature of wind generation may pose substantial challenges, particularly at higher levels of penetration, including an increase in regulation costs and incremental operating reserves, but can also lead to increased opportunities for energy storage, demand‐side response, cross‐border interconnections, and other flexibility measures. In addition to onshore wind power installations, which are already saturating in some countries, such as Denmark, a large number of offshore wind power plants (WPPs) have been developed recently, and this trend is likely to continue into the future[5]. Increasingly, such plants will be sited further offshore, in the form of larger wind farms, and will be connected onshore either individually through a high‐voltage direct current (HVDC) connection, or as part of an interconnected direct current (DC) grid[6]. Wind generation, by its mere presence, does not necessarily worsen the stability of a system, but it does change its characteristics, and through intelligent coordination of power electronic‐based controls, system capabilities could even be enhanced in some situations[2, 7]. System stability issues range from the ability to maintain generator synchronism when subject to a large disturbance (transient stability); the ability to restore steady‐state conditions (voltage, current, power) after being subject to a small disturbance (small‐signal stability); the ability to recover and maintain system frequency following a major generation‐load imbalance (frequency stability); and the ability to maintain an acceptable voltage profile after being subjected to a disturbance (voltage stability)[8–10]. Those issues of concern for a particular system will depend on system size, wind distribution relative to the load and other generation, along with the unit commitment/economic dispatch (UC/ED) decisions and network configuration. However, they are likely to be first observed during the night or seasonal low‐ demand periods when instantaneous wind penetration may be high[11], say greater than 20%, or alternatively when wind exports across a region are high, even in cases when the annual (wind) energy contribution to the system is comparatively low. So, for example, during periods of high instantaneous wind penetration, with reduced numbers of conventional (synchronous) generators online, frequency stability may be affected due to the reduction in governor response[12, 13], and, particularly for smaller systems, by the reduction in synchronous inertia[14–17]. For example, the All‐Island Grid Study of Ireland would be insecure, without additional measures being taken, for approaching 30% of the year 2020 due to a lack of adequate synchronous inertia[18], as shown in Figure 3.1. Similarly, a study of

the Electricity Reliability Council of Texas ­(ERCOT) system observed a decline in its frequency response, based on frequency event records taken over a span of four years with increased wind penetration[17]. The transient stability of a system may also be reduced when synchronous units are de‐­committed and replaced with wind generation connected at lower voltage levels, and hidden behind a relatively large impedance[19, 20]. ­However, transient stability impacts are largely affected by the turbine technology. For example, a study performed by Transpower (New Zealand) reported that “old” technology, fixed‐speed induction generators (FSIGs), worsen the transient stability of the system, as they absorb reactive power during and after a fault, and are generally not voltage ride through (VRT) compliant[21]. However, variable‐ speed wind turbines (doubly fed induction generators and direct drive full converters) have VRT capability, and can improve the transient stability of the system. Alternatively, angle stability, both from a small‐signal and transient stability point of view, may be threatened due to large voltage angle differences, when a large wind power export occurs from one region to another[22]. In addition, the reverse power flow from former load feeders may have implications for associated protection systems[23]. During times of system stress, wind power curtailment can be seen as one solution to maintain system stability, and other security‐related concerns that have been foreseen[24]. Such a measure, though, should be seen as a last resort, and is more likely to be observed in small isolated systems, such as Hawaii, Ireland, New Zealand, and so on[15, 25, 26]. A simple metric, known as the instantaneous system nonsynchronous penetration (SNSP), which is essentially representing the proportion of nonsynchronous generation (wind and HVDC import) with respective to total load (including HVDC export), within a synchronous region, is being used by the All‐Island transmission system operators (TSOs) (EirGrid and SONI). It has been identified that higher SNSP levels (55 + %) may require wind curtailment, unless some alternative measures to maintain system security are considered[25, 27]. However, even for much larger systems, such as the European Continental synchronous area, critical instantaneous penetration rates can be associated with an anticipated growth in wind and solar generation, although such limits will vary with operational conditions[28, 29]. The primary objective of dynamic analysis of a future system is to identify areas of concern, before proposing measures that reduce the risk of wind curtailment due to the introduction of dynamic constraints[30]. Soft measures may include appropriate modification of controller settings, coordinated protection schemes,

Technical Impacts of High Penetration Levels of Wind Power on Power System Stability  49

60,000

Inertia 2010 Inertia 2020

Synchronous inertia (MW s)

50,000 40,000 30,000 20,000 10,000 0

Insecure region 0

10

20

30

40

50

60

70

80

90

100

Percentage of hours in year (%) Figure 3.1  Inertia duration curves for the all‐island system in Ireland.

es

seri

H

on

c

sta

ns

Disturbances—type, location, magnitude

ab ap

tra ls

cc mi

tro

na

Dynamic analyses

on

dy

Static analyses

Pc

PP

tio

ion

op

tat

ity

Network protection strategies

teg

ilit

ies

W

en

ibil

res

lex

Mitigation options

PP

(Scenario initialization)

rep

ef

ad

sid

c lo

nd

mi

Load flow

s

itie

a ap

tic

uti

na

ma

De

Dy

trib

eries

bil

Grid model

dis

time s

WP

ad

W

Lo

Hourly

Production cost simulation

ies

ime

yt ourl

Scenario selection Operational measures

Figure 3.2  Representative schematic of power system operational analysis.

and market‐based flexibility incentives, while hard measures may include network reinforcement and phasing in/retrofitting flexible generation plants.

­SYSTEM MODELING As a precursor to assessing the stability of a system for given future snapshot scenarios, production (generation) cost simulations will ensure that a given wind integration scenario is feasible. In addition, steady‐ state load flow[31, 32], N − 1 contingency[33, 34], and short

circuit analyses[8, 31] will have been performed in o­ rder to assess the steady‐state adequacy and utilization of the transmission system, and to assess if the plant portfolio and grid network are sufficiently strong to cope with a number of predefined disturbances, linking to significant system failure. Such analysis, represented in Figure  3.2 may already have implicitly addressed some dynamic issues for the system, e.g. ramping capabilities of conventional units, plant flexibility associated with wind forecast uncertainty, spinning reserve requirements, minimum number of, or locational, must‐run generation units[33]. It is also

50  Advances in Energy Systems

noted that the steady‐state and dynamic capabilities of WPPs, and their control capability, can be observed within the load flow and stability analysis blocks. Consideration of a future power system for production cost simulations will clearly include portfolio development, i.e. the retirement of existing plants and the introduction of new units, in specified locations and with particular dynamic characteristics, in order to support increased demand growth. The capacity value of wind generation[35] and overall system reliability may inform the required installed conventional generation portfolio, as part of a capacity expansion model[36]. An evaluation of the reserve requirements on various timescales, recognizing the variability and time‐varying nature associated with large‐scale wind penetrations[37–39], and associated wind and demand forecasting capabilities must also be considered[40, 41]. The resulting flexibility of the portfolio, including any plant retrofits, particularly at subhourly intervals, will need to be addressed to recognize unit ramping limits and startup/shutdown costs, as well as hydrological constraints in the case of hydropower[42–44]. It may also be appropriate to consider the cycling costs for existing thermal generation[45–48], or operational practices, along with transmission constraints and market structures that could affect the realizable flexibility of the system[49, 50]. Systemwide voltage control and reactive power management may also need to be more closely integrated with UC procedures at higher wind penetration levels, before any dynamic studies can begin[51, 52]. For example, Denmark is moving toward undergrounding its transmission system, to be compensated by switchable shunt reactors, by 2050[53], while at the same time aiming for a 100% electricity share from wind generation by 2035, with very few (or without) conventional central power plants in operation, thereby reducing overall short circuit capacity and the number of continuously acting automatic voltage regulators (AVRs) in the system. Under such circumstances, wind power variability may necessitate frequent switching of discrete controllers, particularly shunt reactors to regulate voltage, which will, in turn, so significantly impact their lifetime as to make this an economically infeasible practice[54]. Therefore, the Danish TSO, Energinet.dk, is considering a coordinated automatic voltage control (AVC) system to overcome future operational challenges[55]. Large‐scale wind integration will most probably also necessitate upgrades or expansions of the transmission (and distribution) network, particularly if offshore DC grids are incorporated. Measures such as dynamic line rating or high‐temperature low‐sag conductors[56–58], special protection schemes, and

local storage[59] may also enable increased network utilization, and/or ease (planning) delays in network expansion. Reactors, static var compensators (SVCs), flexible AC transmission system (FACTS) devices, and so on at particular locations may also be required to maintain steady‐state (and later dynamic) network performance, and to provide an acceptable voltage profile. Iterations may be required between production cost simulations and load flow studies, before obtaining a plant portfolio and network configuration that is deemed cost effective, relieves/reduces network congestion, and maintains security of supply. Dynamic simulations may introduce a further iteration to this process, by further requiring revisions in operational practice, network reinforcement, and plant portfolio characteristics. So, for example, line (congestion) limits may be set by transient stability concerns rather than steady‐state thermal limits[60–62]: maximum power throughput over a transmission line is normally limited by thermal limits for short transmission lines, voltage stability limits for medium‐ length lines, and rotor angle stability limits for long transmission lines. The load‐flow analysis and production cost simulations form initialization inputs to the stability assessment by defining acceptable generation dispatches and network configurations. Traditionally, system stability may have been assessed for particular snapshot cases, such as maximum/minimum system demand conditions amongst a number of cases, in order to reduce the computational burden. To ensure coherency with steady‐state (N − 1 security) assessments, as highlighted earlier, it can be valuable to perform dynamic analysis under similar conditions. However, since wind power production is typically weakly correlated with system demand, a much wider range of credible analysis cases should be considered, in order to fully appreciate the impact of high wind penetrations on system dynamics. Wind generation may be weakly correlated (diurnally and/or seasonally) with demand, but not uncorrelated with demand, so it is simplistic (and inaccurate) to focus on high wind production coupled with low/high demand scenarios alone[25, 45]. Indeed, the statistical likelihood of particular scenarios should be considered in the selection process and when evaluating consequent actions. Furthermore, different wind deployments should be considered for analysis, in terms of turbine technology and geographical spread. Where possible, wind and demand time series should be employed, in order to capture the underlying correlation[63]. Multiyear analysis should perhaps be performed in order to capture less common but threatening scenarios.

Technical Impacts of High Penetration Levels of Wind Power on Power System Stability  51

Validation of dynamic generator (conventional units and WPPs) models is of key importance, although an appropriate model complexity will be dependent on the study application. For example, an assumption of constant wind speeds may be appropriate for short‐ term stability studies of a few seconds duration[9], whereas long‐term stability studies over several minutes may need to consider the impact of varying wind speeds. It is generally suitable to employ generic models representing different wind turbine technologies, although the models used should not only consider the underlying physics of wind turbine dynamics but they should also recognize relevant (minimum) grid code requirements for the system under study that have a (significant) impact on the WPP controls. Originally, IEEE established a working group on Dynamic Performance of Wind Power Generation, which then merged with the Western Electricity Coordinating Council (WECC) Renewable Energy Modeling Task Force[64], which again works together with the IEC working group on Electrical Simulation Models for WPP[65]. In 2014, the WECC task force published specifications for second‐generation generic models[66], and in 2015, the IEC working group published the first edition of an IEC standard[67]. Sharing many of the same experts, these models are very similar, but there are some minor differences, with the main one being that the IEC models include options for more details, e.g. on reactive power capability[68]. The WECC and IEC generic models are intended for short‐term power system stability studies of 10–30 seconds duration, assuming that wind speeds are constant during the simulations. So far, there are no standard models available to investigate wind power variability in long‐term stability studies, and the need for such models will depend on the particular phenomenon being studied and the size of the synchronous area under study. The WECC document and the IEC standard specify models for each of the four main types of wind turbines, i.e. type 1 (a directly connected induction generator), type 2 (same as type 1, but with variable rotor resistance), type 3 (doubly fed asynchronous generator), and type 4 (fully sized power converter). Each model includes a set of parameters that can vary from one wind turbine manufacturer to another. A default parameter set may be used in studies where the specific wind turbines are not known, but the parameters are used to account for variations in the dynamic behavior of different wind turbines. Also, a specific wind turbine can be operated in different control modes depending on the requirements of a specific TSO. The existing WECC and IEC models cover several different reactive power/voltage control modes

during normal operation and during voltage dips. Emulated inertial responses are not directly implemented in the models, because this type of response is at an early stage of development and therefore not considered sufficiently mature for standardization, but the power reference points can be used to connect the generic models to a specific user defined emulated inertial response control model. So, for example, additional adjustments and extensions to the type 4 IEC generic model have been considered elsewhere[69], in order to reflect the dynamic features of wind turbines relevant for active power and grid frequency control capability studies. Wind turbine manufacturers covering the majority of the market (Enercon, Gamesa, General Electric [GE], Senvion, Siemens, and Vestas) have contributed to the model specifications, with internal model validation ensuring that the WECC[70] and IEC[71, 72] models are applicable to their specific wind turbines. The relation between individual turbine controllers and the centralized plant controller must also be addressed[73–75]. In particular, communication time delays can compromise the ability to perform fast‐responding services, such as emulated inertia controls. Furthermore, the response time of wind turbine inverters limits their ability to support the grid during the first 10–100 s of ms after a disturbance has occurred. A study variant may be to assess the advantages of enhanced wind turbine capabilities, coupled with coordinated setpoint controls across a network area. For offshore wind plants connected via HVDC transmission, the modeling requirements depend heavily on the study scope. In many cases, it is sufficient to limit the modeling to the onshore HVDC inverter, and use a simplified aggregated wind plant model. Such an approach is particularly valid when onshore voltage and reactive power issues are in focus, because the DC stage decouples reactive power flows in the offshore alternating current (AC) system from the onshore grid. However, when discussing active power control and system frequency support, the relation between the HVDC controller, the centralized plant controller, and the individual turbine controllers must be addressed[76]. Again, communication delays and response times are important when quantifying the response during the first few seconds after a disturbance has occurred. For fast transients in the millisecond range, the dynamics of the DC system are important, which will require detailed models to be simulated on shorter time steps[77]. Software packages that focus on the power system (electromechanical) dynamics of interest can accurately simulate WPPs connected through voltage source

52  Advances in Energy Systems

converters (VSC)‐HVDC[78]. Adopting a combined simulation strategy, i.e. stability simulation for AC grid dynamics, and electromagnetic transient simulations for DC grid dynamics, provides an acceptable simulation speed and accuracy. Finally, (dynamic) load modeling is a topic that historically has received limited attention, partly due to practical difficulties in obtaining widespread data for model validation, and the time‐varying nature of the models themselves, as the load composition varies diurnally and seasonally, as well as evolving annually[79–81]. With increased wind penetrations, however, leading to “lighter” systems, and with wind plants located at subtransmission voltage levels, load characteristics are likely to play a greater role in the dynamic performance of the system.

­FREQUENCY CONTROL AND INERTIAL ISSUES As wind penetration levels rise, conventional generation will gradually be displaced, with implications for the frequency regulation capacity. At lower wind penetration levels, system flexibility may actually be enhanced, as conventional units are backed off but remain online, enhancing the headroom or maneuverability of the system as a whole. However, if such generation is displaced offline the fraction of generation participating in governor control is likely to reduce, along with the inherent inertia of the system, 60.1

resulting in faster frequency dynamics following a major network fault or load‐generation imbalance[13, 82–84] . Wind turbines can, of course, provide a governor droop response, similar to conventional units, and such capability is mandated in many grid codes. However, while a high‐frequency response can be readily achieved, i.e. a sustained reduction in output, a low‐frequency response requires the turbines to have been curtailed in advance, i.e. a period of reduced production. In some jurisdictions, e.g. ERCOT, wind turbines that have been curtailed (for network reasons) can contribute to the frequency response capability[85], but, for many systems, wind governor controls remain an untapped resource, and the implications for UC and reserve policies are not resolved. The frequency response may also depend on the type (synchronous or wind generator) and location [transmission connected or distributed generation (DG)] of the generation loss. For example, a study on the US Western Interconnection identified that the frequency response for a DG outage was improved over that for a transmission connected outage[86], as shown in Figure 3.3. The difference follows from the fact that the DG loss results in a depressed (local) distribution voltage, such that a low load voltage reduces (local) power consumption, and hence, the postdisturbance system demand. Fixed‐speed wind turbines do naturally provide an inertial response akin to that provided by a synchronous machine[87]. However, variable‐speed wind turbines decouple the rotating mass of the turbine

Trip of two Polo Verde units Trip of equivalent DG

60.05 59.60

Frequency (Hz)

59.95 59.9 59.85 59.8 59.75 59.7 59.65 59.6

0

10

20

30

40

50

60

Time (seconds) Figure 3.3  Frequency response of Light Spring Hi‐Mix case – DG trip versus two Palo Verde unit trips. Source: Reprinted with permission from the Ref.[86]. Copyright 2014 National Renewable Energy Laboratory.

Technical Impacts of High Penetration Levels of Wind Power on Power System Stability  53

from the power system, which offers a number of operational and quality benefits to the turbine but removes any intrinsic inertial capability. At times of high nonsynchronous penetration, and the resulting displacement of synchronous generation, the online inertia will be reduced, altering the system response for both faults and contingencies[17, 29]. For smaller power systems, or those linked together by asynchronous HVDC links, the effect can be particularly important and may be of concern[25, 88]. The resulting high rate of change of frequency (RoCoF) may, for example, cause anti‐­ islanding RoCoF protection to mal‐operate, further increasing the generation‐demand imbalance in the system[89]. Low inertia has not, as yet, caused a problem for larger power systems, but is being investigated[63, 90]. For example, as previously shown in Figure  3.1, a significant reduction in synchronous inertia (stored energy) is estimated for the year 2020 compared to 2010 for the all‐island (Ireland) system, which presents a major bottleneck to higher levels of wind penetration, as all generating units are (only) required to withstand a maximum RoCoF of 0.5 Hz s−1 following a load‐­generation imbalance. RoCoF limits are also defined  as  part of anti‐islanding protection schemes for distribution‐connected generation, with approximately 50% of wind generation so connected in ­Ireland. Based on a UC for the year 2020, the maximum RoCoF, assuming that the largest infeed/outfeed is tripped in each hour, is shown in Figure 3.4. It can be observed that both the current maximum RoCoF generation limit (0.5 Hz s−1) and relay threshold (0.55 Hz s−1) value would be violated for numerous events, and hence increasing the risk of additional unit outages and

possible cascading events. Consequently, EirGrid and SONI (TSOs of Ireland and Northern Ireland) are currently exploring various measures, such as raising the RoCoF limit from 0.5 to 1 Hz s−1 for all units[91], modified and selective plant protection strategies, improved plant monitoring to ensure that conventional generators provide appropriate reserve in a timely manner following an energy imbalance, through alternative operational strategies (operational measures, load management, parking of machines, i.e. plant operation at low output but with reduced [or none] capability to provide system support services) or infrastructure reinforcement (synchronous condensers and construction of AC interconnections), or a combination of the above. Hawaii also faces the issue of high RoCoFs inducing gas unit tripping, which has led to a change in plant temperature control settings[14]. The same lean burnout phenomenon has also been seen in Florida, with multiple gas turbines tripping in one occasion[92, 93]. For larger systems, high RoCoFs are of less concern, although at high renewable (wind and solar) penetrations, low frequency nadirs can occur: a reference incident (3500 MW generation loss) on the European Continental synchronous area could result in the security level of 49.2 Hz being at risk for 18% of the time (Figure 3.5) assuming approximately 35% annual renewable energy contribution[28, 29]. The study conclusions were based on 814 680 “reference incident” dynamic simulations covering over 8760 hours of time steps and 93 annual scenarios representing different wind profile years[94]. It is also noted that the load self‐regulating effect, incorporated within the study, contributes strongly to frequency

0.6

RoCoF (Hz/second)

0.4 0.2 0 –0.2 –0.4 –0.6 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Wind generation level (pu) Figure 3.4  Maximum RoCoF for all‐island (Ireland) system for the year 2020 following tripping of the largest infeed/ outfeed.

54  Advances in Energy Systems

40

Probability of occurrence (%)

35

f0 = 49.9 Hz ΔPe = 3.5 GW

30

D = 1%/Hz

25 20 15 10 5 0 48.8

48.9

49

49.1

49.2

49.3

49.4

49.5

49.6

49.7

Frequency nadirs following an incident (Hz) Figure 3.5  Distribution of frequency nadirs of European Continental synchronous area following a reference incident.

stability, and is a key parameter for characterizing the “critical” variable renewable penetration rates. Modern wind turbines can provide a fast frequency (emulated inertial) response with its own characteristics. Due to the fast response time of wind farm controllers and the energy stored in wind turbine rotors, it is technically feasible to provide a rapid, but temporary, power injection, given that otherwise the wind turbines would lose too much rotational speed and therefore also aerodynamic torque[95]. The implementation strategies can differ between manufacturers and academic studies, however, in general, such controls cause the power output of an individual turbine, or farm, to temporarily increase in the range of 5–10% of the rated turbine power, following a significant under‐frequency excursion, for several seconds[96–98]. Typically, the response consists of a fixed power injection signal, triggered once the frequency deviation exceeds a defined threshold, or, alternatively, a power setpoint trajectory is defined based on the actual frequency deviation, again triggered after a deadband is exceeded[99–101]. Several studies based on meteorological or power measurement data[102–104] indicate that the aggregate supply of rotational energy at a national scale can assist the frequency response, but it is not always available and changes with turbine operating point, as well as being dependent on turbine electrical and mechanical constraints and controller tuning[105, 106]. Time delays associated with frequency measurement, activation deadbands and centralized farm communications may encourage local turbine controls. A good compromise should be

made for appropriate controller parameter settings, which allow wind turbines to satisfy the grid code but do not seriously impact their own stability and lifetime. However, the added instrumentation implies a trade‐off between increased cost and improved wind plant response. Particularly, in the Quebec system, operational experience has been gained for both type 3 and type 4 wind turbines[107, 108], which has raised confidence in the technology that a suitably shaped and sustained response can be  achieved[95], but also highlighted the opportunity to further improve the control strategies, enhance existing dynamic wind turbine (and farm) models, and also to revise system operator specifications and requirements (particularly considering the recovery phase of the response). Alternatively, fast‐acting response from various load categories (water/space heating, air‐conditioning, refrigeration systems, municipal water pumping, and swimming pools)[109–114], electric vehicles[115, 116], and storage devices, such as batteries and flywheels[117–119] are also beneficial options. One future alternative to currently proposed fast frequency responses may be a virtual synchronous machine approach, whereby power electronics converters are controlled in order to emulate, within a certain degree, the characteristics of a synchronous machine. Several implementations have been proposed in the last decade, e.g. VISMA (virtual synchronous machine), synchronverter[120], but the concept is relatively immature, and, in particular, control tuning and stability assessment for a power system containing several such units is particularly

Technical Impacts of High Penetration Levels of Wind Power on Power System Stability  55

challenging[121]. Additionally, with the emergence of large‐scale offshore wind farms, interconnected through offshore transnational DC grids, opportunities may exist for sharing of primary frequency reserves between asynchronous power systems, e.g. North Sea offshore grid, the mechanisms being similar to an onshore wind farm being connected through a power electronics converter. ­TRANSIENT STABILITY AND FAULT RIDE‐THROUGH Transient stability studies examine the operation of power systems during severe fault contingencies, e.g. a fault on a transmission line, and their ability to maintain synchronism, with times of high wind penetration being relevant here. Wind turbines can contribute to system restoration with low voltage ride through/high voltage ride through (LVRT/HVRT) capabilities[122, 123], as indicated by Figure  3.6, showing the LVRT grid code regulations for various countries. The priority given to active or reactive power recovery, as part of LVRT controls, may also be a system specific decision, with the former approach more likely to be appropriate for smaller systems[27, 124]. It should be noted, however, that current grid code requirements regarding wind turbine fault behavior do not represent a guarantee of transmission system stability. The level of support provided is network sensitive, and proper representation of the impedance connecting the wind farms is crucial. Transient stability performance also

strongly depends on the employed wind turbine technology and the grid code regulations in place, such that power systems with a significant proportion of older fixed‐speed (FSIG) technology, not equipped with fault ride through (FRT) capability, are likely to observe a large proportion of wind turbine outages during a fault‐induced voltage dip. For example, in Portugal and Spain, a significant share of the total installed wind turbines was previously not equipped with FRT capability, and therefore, large numbers of wind turbines were often tripped during a voltage dip[125]. This finding resulted in the introduction of FRT requirements, whereby decade old wind turbines were required to remain connected without reactive power support during a fault, while all new wind turbines were required to provide reactive power support during a fault. Figure  3.7 shows the FRT certified wind power in Spain, reaching 97% of the installed capacity, according to Spain’s TSO, Red Eléctrica de Españ a (REE). The number of (wind) power losses greater than 100 MW was approximately 50 times (2005, 2006, and 2008), 87 (2007), and 30 (2009), but falling to 0 after 2009. As a result of the FRT implementations, the problem of significant wind generation tripping has been solved; therefore, wind plant curtailment due to FRT requirements has not been required since 2008 in the Spanish system. A study on the New Zealand system, investigating the impact of wind power integration on transient stability, identified that stability‐related constraints ­ limit the power flow between different areas, due to

1 0.9

Retained voltage (pu)

0.8 0.7 0.6 0.5 0.4

Ireland, Alberta Germany borderline 2 Germany borderline 1 Quebec Denmark ENTSO-E, Type D upper bound ENTSO-E, Type D lower bound Spain UK

0.3 0.2 0.1 0 0

0.5

1

1.5

2

2.5

Time (seconds) Figure 3.6  Low voltage‐ride through requirement from wind power plants in various countries.

3

56  Advances in Energy Systems

25.000

Certified power each month (MW)

1.800

20.000

1.600 1.400 15.000

1.200 1000

10.000

800 600

5.000

400

Accumulated certified power (MW)

2.000

200 0

0

9 1 1 3 5 7 9 1 1 3 5 7 9 1 1 3 5 7 9 1 1 3 5 7 9 1 1 3 5 7 9 1 –0 –1 –0 –0 –0 –0 –0 –1 –0 –0 –0 –0 –0 –1 –0 –0 –0 –0 –0 –1 –0 –0 –0 –0 –0 –1 –0 –0 –0 –0 –0 –1 07 07 08 08 08 08 08 08 09 09 09 09 09 09 10 10 10 10 10 10 11 11 11 11 11 11 12 12 12 12 12 12 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20

Figure 3.7  Fault ride through certified wind power in Spain.

1.4 Heavy summer with standard load model Heavy summer with composite load model

1.2

Voltage (pu)

1 0.8 0.6 0.4 0.2 0

0

1

2

3

4

5

6

7

8

9

10

Time (seconds) Figure 3.8  Load‐induced voltage collapse in heavy summer base case, Midway‐Vincet (500 kV), California. Source: Reprinted with permission from the Ref.[86]. Copyright 2014 National Renewable Energy Laboratory.

FSIGs absorbing reactive power during and following a fault and eventually tripping off unless additional external dynamic reactive power support was available [21]. The Western Wind and Solar Integration Study (Phase 3) concluded that, with good system planning and power system engineering practices in place, transient stability should not be a bottleneck at higher levels of wind penetration [86]. However, a common-mode or sympathetic trip of DG during a disturbance could result in a slower recovery and a sustained lower voltage, which in worst-case undervoltage DG tripping could lead to a system collapse.

The challenges of simulating such, or other, behavior are illustrated by the response to a three‐phase fault at Midway‐Vincet (500 kV) in California using alternatively a standard WECC load model and a composite load model ­(Figure  3.8). The standard WECC load model assumes a 20% contribution from induction motors with the remainder represented by static components: the weighting of the static ZIP (constant impedance, Z, constant current, I, and constant power, P) terms vary by location, although in the majority of cases constant impedance is assumed for the reactive load and constant current is assumed for the active

Technical Impacts of High Penetration Levels of Wind Power on Power System Stability  57

load. The composite load model, also known as the WECC composite load model (CMPLDWG), integrates DG into the model with a variety of tripping characteristics, and also includes a much higher contribution from induction motors. It can be observed that the responses are completely dominated by the load model, and, not surprisingly, the study also highlighted the importance of better load modeling for transient‐stability analysis[86]. It was concluded that changing the load model had a greater impact on system performance over changing the level of renewable penetration. Scenarios of particular interest are associated with reverse power flow situations, where conventional units have been displaced to accommodate wind power. With wind farms connected at lower voltage levels, fault critical clearing times can be determined to see how they may be affected, with implications for network protection schemes and relay settings[126]. Indeed, protection relay settings may need to recognize changes in the dynamic response of the system, dependent on wind instantaneous levels and geographical dispersion. Wind turbines could trip due to a widely seen network fault, or the reduction in active power infeed could be significant, resulting in voltage depressions and frequency stability issues[127]. The operation of associated protection systems can, therefore, play a critical role, and its simulation may require sophisticated calculation methods[78]. Delayed active power recovery from (grid code compliant) wind turbines following a fault‐induced voltage dip may also result in a short‐term generation shortfall, resulting in frequency instability issues[128]. Offshore wind farms in DC grids can also pose FRT challenges, although plants consisting of mixed turbine types can improve the robustness to onshore faults with voltage source converters‐high voltage direct current VSC‐HVDC connection[129]. In order to mitigate stability problems, fast‐acting reactive power response devices during and following the disturbance are required. A variety of options, or combination of options, can be considered including synchronous condensers and FACTS devices, in addition to the response obtained from wind turbines and conventional generators[45, 130]. Coordinated wind turbine controls may also help to dampen oscillations, while VSC‐HVDC can, to some extent, also be used for system stabilization[76, 126]. Traditional system reinforcement (e.g. transformers, shunt capacitors, and line upgrades) may also be required to maintain adequate stability margins at higher levels of wind penetration. New wind turbine concepts, such as variable‐speed designs based on an electromagnetic coupler (synchronous generator directly connected to the grid, and able to generate reactive power up to three times rated) could also be considered[131].

­VOLTAGE STABILITY Voltage stability relates to maintaining an acceptable voltage profile in steady‐state and following a disturbance, such as an increase in load or a ­network fault. It is mainly associated with an inability to meet (local) reactive power requirements, and so is dependent on the reactive power capability of generators and the reactive demand of loads, but is also influenced by implemented voltage control strategies, such as interactions with transformer tap changers. It may be appropriate, particularly in network regions where (conventional) generation has been displaced, to introduce SVCs, static compensators (STATCOMs), synchronous condensers, or similar equipment, or even to make certain generators “must run” for voltage support reasons. Voltage instability may result in a loss of load, tripping of transmission lines and other elements, and so lead to cascading outages. Consequently, when assessing voltage stability at high wind penetrations, the potential to utilize the reactive power capabilities of the turbines is a key determining factor. In general, voltage stability is likely to be unaffected or enhanced by the presence of wind turbines[132], if the turbine reactive power control capabilities are deployed to manage voltage[63], and particularly if the turbines are connected at transmission level. However, this is largely true only for variable‐speed wind turbines, as FSIGs absorb reactive power during and following a fault, and, therefore, lower the voltage stability margin of the system. Moreover, wind‐driven displacement of conventional power plants reduces the overall systemwide dynamic reactive power and short‐circuit power capacity[133, 134]. The Irish All Island Facilitation of Renewables Studies investigated steady‐state voltage stability limits using PV and QV curves for a wide range of dispatches, with zero exchange across the HVDC interconnectors to Great Britain[25]. The analysis identified that distribution‐connected wind power tended to reduce voltage stability, as such installations were not equipped with additional reactive power compensation, while reactive power compensation at distribution level could not resolve reactive power issues at transmission level. It was also identified that voltage instability may occur even close to the nominal voltage level in the worst cases, even with grid code compliant wind turbines, as shown by the solid lines in Figure  3.9. However, with improved reactive power support deployment, voltage stability was found to be improved, as shown by the dashed lines. Subsequently, EirGrid has implemented an on‐ line wind stability analysis tool (WSAT)[30], which, amongst other functions, assesses the voltage stability for the current system condition and for a range

58  Advances in Energy Systems

QV curves for WMAX cases for S/S 4000 kV Woodland (T20TSA) + mitigation 1500

Injected reactive power (MVAr)

1000 500 0 0.500

0.600

0.700

0.800

0.900

1.000

1.100

–500 Wind 25% Wind 50% Wind 75% Wind 100% Wind 25%-mitigation Wind 50%-mitigation Wind 75%-mitigation Wind 100%-mitigation

–1000 –1500 –2000 –2500

Voltage (pu)

Figure 3.9  Reactive power versus voltage analysis (QV curves) for 400 kV busbar Woodland for winter maximum load and various wind power levels in the all‐island system.

120 CPP gen. (%)

of short‐term demand and wind production increase scenarios. A similar study using PV analysis to assess the wind integration impact on voltage stability was conducted by Transpower New Zealand[135]. One of the study assumptions was that wind generation, with limited voltage control ability, displaces other forms of generation. It was seen that at lower wind penetration levels, such that other generation was not displaced, the voltage stability of the system was improved due to the additional reactive power margin made available from conventional generators by relieving their active power output. However, at higher levels of wind penetration, with displacement of other forms of generation, the voltage stability limit was reduced by 10–34%. A case study on the Danish power system also suggested that the overall voltage security level was compromised at higher levels of wind penetration, considering current grid code compliance, but without deployment of any other reactive power sources, such as synchronous condensers and SVCs[136], as shown in Figure  3.10. Traditionally, PV curve analysis is used to estimate the maximum power transfer at a particular bus. However, a basic assumption is that the generation is dispatchable. Since wind energy is not, a new approach has been proposed to assess the voltage stability of wind integrated systems[137], whereby, in order to include wind variability, a PV surface for secure operation known as a voltage secure region of operation (VSROp) is proposed. The method assumes a constant wind generation level for each PV curve in the 3D surface.

100 80 60 20 CH

P

100

40 ge

n.

60 (%

)

80

50 0

en.

g ind

(%)

W

Figure 3.10  Voltage security boundary of the Western Danish power system, dependent on conventional power plant (CPP), combined heat and power (CHP), and wind generation mix.

­ MALL SIGNAL STABILITY AND S SUBSYNCHRONOUS INTERACTIONS In general terms, there are four major types of power oscillations[60, 138]: (i) intraplant oscillations, where machines within the same power station oscillate against each other at a frequency of 2–3 Hz, while the remaining system remains unaffected, (ii) local mode oscillations, where one generator oscillates against the rest of the system (1–2 Hz), (iii) interarea oscillations,

Technical Impacts of High Penetration Levels of Wind Power on Power System Stability  59

where a set of coherent machines oscillate against another group of coherent machines (1 Hz or less), and (iv) torsional mode oscillations, which are associated with the turbine generator shaft system (10–46 Hz). Variable‐speed wind turbines, similar to HVDC interconnection, do not generally introduce electromechanical oscillatory modes and hence do not directly contribute to system oscillations. However, depending on the wind penetration level and underlying wind turbine technology, system damping performance may change indirectly. For example, wind generation may displace individual synchronous units and hence impact the oscillatory modes and controller contributions: if a synchronous generator with an installed power system stabilizer is offline it cannot contribute to system damping. Similarly, increasing wind penetration levels may significantly alter the direction and magnitude of power flows within the transmission network, with implications for small signal stability. Furthermore, wind power interactions with synchronous machines may change the damping torque induced on their shafts[139]. A considerable number of studies have been performed to investigate the impact of wind penetration on power oscillations. However, no clear and generalized conclusions can be drawn as to whether wind integration improves or decreases power oscillation damping. One of the earlier studies[140], investigating the impact of fixed‐speed (type 1) and doubly fed (type 3) wind turbine technologies, concluded that type 1‐based wind farms improved damping more than type 3‐based wind plants. It was also shown that both wind plant categories improved system damping more than synchronous generators. Similar results were reported in other time domain simulation‐based studies considering only type 3 wind turbines[141–143]. Some other studies involving small signal stability analysis have also reported an improvement in system damping with wind integration[22, 144, 145]. However, other studies have concluded that increased wind penetration levels could have de‐stabilizing effects, and not necessarily at high wind penetration levels[22, 146–149]. For example, a voltage dip due to a fault at the wind turbine terminal is likely to excite torsional oscillations in the wind turbine shaft, with the frequency of such mechanical oscillations tending to be around 1.7 Hz. The wind turbine system acts as a low pass filter to such oscillations, and, as a result, the frequency of the oscillations introduced in the voltage and power output of the turbine is approximately 1 Hz, which is close to the natural frequency of power oscillations[127]. However, a third viewpoint reports that, depending on wind farm location, fault location, and turbine operating states, higher wind

penetrations could be detrimental or beneficial to the system response[150–155]. In general, increased wind penetrations will reduce the number of oscillatory modes and improve system damping, particularly if turbine (damping) controls are introduced[88]. The latter may be achieved by modulating either the active and/or reactive power output, with the effectiveness of different control input(s) and damping output(s) dependent on the network properties and on where the wind plant is connected. Finally, it should be noted that system damping can be made worse without careful coordination between WPPs providing power oscillation damping, i.e. when multiple WPPs are required to simultaneously contribute damping support[22]. The impact of subsynchronous interactions (SSIs) has also been investigated, incorporating subsynchronous resonance (SSR), subsynchronous torsional interaction (SSTI), and subsynchronous control interaction (SSCI), with the latter reported as the main concern for wind integration. For example, SSCI has been seen on the ERCOT system for a wind plant connected radially via a series compensated line, where, following a single‐line to ground fault on the transmission line, the wind plant experienced a buildup of subsynchronous oscillations, resulting in damage to both the series capacitor and the wind turbine[156] (Figure 3.11). Fixed‐speed (type 1) and wound rotor fixed‐speed (type 2) wind turbines, if operating close to synchronous frequency, generally do not see SSIs. However, due to their control response, doubly fed (type 3) wind turbine‐based wind plants can be sensitive to SSI, while direct‐drive (type 4) wind turbines are also reported to be SSI insensitive. A study

Figure 3.11  Measured quantities [(phase currents (blue) and voltages (magenta)] at receiving end of transmission line in ERCOT system. Source: Electranix.

60  Advances in Energy Systems

from Elforsk, Sweden, investigating the impact of variable‐speed wind turbines on SSI, obtained findings in agreement with the ERCOT study, suggesting that type 3 wind turbines were more susceptible to resonant conditions at low frequencies, while type 4 wind turbines have minimal impact due to the power electronic interface between the turbine and the transmission line[157]. A further study, investigating SSR in double‐cage induction generator based wind plants connected through series compensated transmission lines, concluded that although torsional interaction does not appear to occur for the considered range of wind farm sizes and series compensation levels, the IG effect, i.e. electrical mode becoming unstable, may be experienced with large wind plants in the range of 100–500 MW at series compensation levels of 50– 60%[158]. Analysis with various commercially available induction generators reveals similar potential for SSR oscillations. However, for any potential SSI found in a given system, adequate countermeasures, such as mitigation through FACTS devices, bypass filters, an appropriate level of series compensation or cost‐effective methods through auxiliary control (damping control) strategies[159] can help avoid any such interactions.

­CONCLUSIONS With increasing wind penetration, the stability of the power system will be affected. At low penetration levels, the effects will be limited and may indeed enhance system performance, with the existing population of conventional generators remaining online but operating at lower outputs. However, at higher wind penetrations, and as conventional (synchronous) generators are displaced, and offshore DC grids emerge, the nature of the power system will change from being largely synchronous to asynchronous. Stability issues are most likely to be seen first during low demand (and high wind) periods, but the nature of the stability challenge, i.e. frequency, voltage, transient, or small signal stability, will depend on the underlying characteristics of the system. It follows that, as part of a wind integration study for a particular system, that a stability assessment should be performed, particularly for wind annual energy penetrations beyond 10%, recognizing that the instantaneous wind penetration will at times be much higher. As part of such an exercise, adoption of operational (stability) tools embedded within energy management systems, coupling and interaction ­between distributed power electronic based (wind plant) controls, enhancements to grid codes and/or

promotion of new (flexibility‐based) ancillary services, changes in operational practice and electricity market structures, real‐time wind plant telemetry and control capability, and so on, may also need to be addressed. While stability analysis may well reveal new operational limits, following from existing load flow, UC, and so on, imposed limits, the operating boundary of a power system with high wind penetration may actually depend on other factors. For example, in systems such as Portugal, instantaneous variable energy generation (wind, run of river hydro and small‐scale combined heat and power), combined with reserves, can at times exceed demand. Similarly, the total generation in Denmark often exceeds the total demand, with wind generation alone exceeding the demand on a number of occasions, reaching 136% instantaneous penetration in December 2013. No technical difficulties have been identified, but the business case for wind generation is clearly affected, as it competes with other energy sources, some of them also renewable, while others are made mandatory to provide reserves and short‐term security of supply[125]. In the future, wind generation may well play a greater role in providing ancillary services to enhance wind energy value, while also reducing the technical (including dynamic) risks of operating power systems with a reduced share of dispatchable and synchronous energy sources. Several of these services, e.g. voltage control, FRT, and frequency support, have been available from wind plants for years and have made their way into different requirements in TSO connection codes. A number of TSOs have requested certain functionality to be incorporated as a future feature: Hydro‐Quebec and ERCOT request emulated inertia, ELIA (Belgium) and NGET (Great Britain) discuss power oscillation damping, while in Ireland a range of frequency and voltage support services are proposed. Key questions remain to be addressed for wide‐scale (commercial) adoption of such capabilities: is provision truly available when needed by the power system  –  how much and how fast and for how long – and can they be efficiently traded? Numerous efforts are underway around the world to better understand and address these issues. However, it is increasingly clear for many power systems across the world that while large‐scale wind integration can present system stability challenges (in addition to existing market‐related, environmental‐related, and other challenges), technical solutions, and commercial opportunities are eminently available, and the limits for further wind expansion aims remain to be reached.

Technical Impacts of High Penetration Levels of Wind Power on Power System Stability  61

­REFERENCES 1. Global Wind Energy Council (2015). Global Wind Statistics—2014. Brussels, Belgium: Global Wind Energy Council (GWEC). 2. IEA Wind Task 25 (2013). Expert Group Report on Recommended Practices: Wind Integration Studies. Paris, France: IEA. 3. Holttinen, H., Meibom, P., Orths, A. et  al. (2011). Impacts of large amounts of wind power on design and operation of power systems, results of IEA collaboration. Wind Energ. 14: 179–192. 4. Tielens P, Ergun H, Van Hertem D. Techno‐economic analysis of large‐scale integration of solar power plants in the European grid. In: 2nd Solar Integration Workshop, Lisbon, Portugal, 2012. 5. Wind Speed. Roadmap to the deployment of offshore wind energy in Central and Southern North Sea to 2030. EU‐IEE Project Wind Speed, March, 2011. Available at: www.windspeed.eu. 6. Chaudhary SK, Teodorescu R, Rodriguez P. Wind farm grid integration using VSC based HVDC transmission—an overview. In: IEEE Energy 2030 Conference, Atlanta, GA, 2008. 7. GE Energy (2005). The Effects of Integrating Wind Power on Transmission System Planning, Reliability, and Operations. Report on Phase 2. Schenectady, NY: New York State Energy Research and Development Authority. 8. Anderson, P.M. and Fouad, A.A. (2003). Power System Control and Stability, 2e. New York, NY: IEEE Press. 9. Kundur, P., Paserba, J., Ajjarapu, V. et  al. (2004). Definition and classification of power system stability: IEEE/CIGRE joint task force on stability terms and definitions. IEEE Trans. Power Syst. 19: 1387–1401. 10. Machowski, J., Bialek, J., and Bumby, J. (2008). Power System Dynamics: Stability and Control, 2e. New York, NY: Wiley. 11. Fox, B., Flynn, D., Bryans, L. et al. (2014). Wind Power Integration: Connection and System Operational Issues, IET Renewable Energy Series, 2e. London, UK: IET. 12. Ela, E., Gevorgian, V., Tuohy, A. et al. (2013). Market designs for the primary frequency response ancillary service—part II: case studies. IEEE Trans. Power Syst. 29: 432–440. 13. Miller, N., Shao, M., and Venkataraman, S. (2011). CAISO Frequency Response Study. Schenectady, NY: GE Energy. 14. Dangelmaier L. System frequency performance of the Hawaii Electric Light system. In: IEEE PES General Meeting, San Diego, CA, 2011. 15. Pelletier M, Phethean M, Nutt S. Grid code requirements for artificial inertia control systems in the New Zealand Power System. In: IEEE PES General Meeting, San Diego, 2012. 16. Ruttledge, L. and Flynn, D. (2015). Short‐term fre quency response of power systems with high non‐synchronous penetration levels. Wiley Interdiscip. Rev. Energy Environ. 4: 452–470.

17. Sharma S, Huang S, Sarma N. System inertial frequency response estimation and impact of renewable resources in ERCOT interconnection. In: IEEE PES General Meeting, San Diego, CA, 2011. 18. EirGrid. Delivering a secure sustainable electricity system DS3 programme, 2013. Available at: http://www. eirgridgroup.com/how‐the‐grid‐works/ds3‐programme. 19. El‐Klhy A, Iravani R. A review of the impacts of multiple wind power plants on large power systems dynamics. In: IEEE Electrical Power & Energy Conference (EPEC), Halifax, NS, 2013. 20. Meegahapola L, Flynn D. Impact on transient and frequency stability for a power system at very high wind penetration. In: IEEE PES General Meeting, Minneapolis, MN, 2010. 21. Transpower (2008). Wind Generation Investigation Project 7: Effect of Wind Generation on Transient Stability. Planning and Investigations. Wellington, New Zealand: Transpower. 22. Hansen AD, Altin M. Impact of advanced wind power ancillary services on power system. DTU Wind Energy E0081 Report, 2015. 23. Roy, N.K. and Pota, H.R. (2015). Current status and issues of concern for the integration of distributed generation into electricity networks. IEEE Syst. J. 9: 933–944. 24. Lew D, Bird L, Milligan M, et al. Wind and solar curtailment: international experience and practices. In: 12th International Workshop on Large‐Scale Integration of Wind Power into Power Systems as well as on Transmission Networks for Offshore Wind Power Plants, London, UK, October, 2013. 25. EirGrid and SONI (2010). All Island TSO Facilitation of Renewables Studies. Dublin, Ireland. 26. Hawaiian Electric Company (HELCO) (2011). Oahu Wind Integration and Transmission Study. Golden, CO: NREL, Hawaii Natural Energy Institute. 27. O’Sullivan, J., Rogers, A., Flynn, D. et  al. (2014). Studying the maximum instantaneous non‐synchronous generation in an Island system  –  frequency stability challenges in Ireland. IEEE Trans. Power Syst. 29: 2943–2951. 28. Burtin A, Silva V. Technical and economic analysis of the European electricity system with 60% RES: EDF R&D study, June, 2015. Available at: http://www. energypost.eu/edf‐study‐download‐15 29. Wang Y, Silva V, Winckels A. Impact of high penetration of wind and PV generation on frequency dynamics in the continental Europe interconnected system. In: 13th International Workshop on Large‐Scale Integration of Wind Power into Power Systems as well as on Transmission Networks for Offshore Wind Power Plants, Berlin, Germany, November, 2014. 30. Dudurych IM, O’Sullivan J, Rogers A, et al. Tools for handling high amounts of wind generation in National Control Centre in Ireland. In: IEEE PES General Meeting, San Diego, CA, July, 2012. 31. Bergen, A.R. and Vittal, V. (2000). Power Systems Analysis, 2e. Upper Saddle River, NJ: Prentice Hall.

62  Advances in Energy Systems

32. Powell, L. (2004). Power System Load Flow Analysis. London, UK: McGraw‐Hill. 33. Wood, A.J. and Wollenberg, B.F. (1996). Power Generation, Operation and Control, 2e. New York, NY: Wiley. 34. Lei Lai, L. (2002). Power System Restructuring and Deregulation. Chichester, UK: Wiley. 35. Keane, A., Milligan, M., Dent, C.J. et  al. (2011). Capacity value of wind power. IEEE Trans. Power Syst. 26: 564–572. 36. Shortt, A., Kiviluoma, J., and O’Malley, M. (2013). Accommodating variability in generation planning. IEEE Trans. Power Syst. 28: 158–169. 37. Holttinen, H., Milligan, M., Kirby, B. et  al. (2008). Using standard deviation as a measure of increased operational reserve requirement for wind power. Wind Energ. 32: 355–378. 38. Soder, L., Abildgaard, H., Estanqueiro, A. et al. (2012). Experience and challenges with short‐term balancing on European systems with large share of wind power. IEEE Trans. Sustain. Energy 3: 853–861. 39. Ela E, Milligan M, Kirby B. Operating reserves and variable generation. Technical Report No. NREL/ TP‐5500–51978, August, 2011. 40. Orwig, K.D., Ahlstrom, M., Banunarayanan, V. et  al. (2015). Recent trends in variable generation forecasting and its value to the power system. IEEE Trans. Sustain. Energy 6: 924–933. 41. Cutululis NA, Litong‐Palima M. Impact of offshore wind power variability on the frequency stability of European power system. In: Proceedings of the International Conference on Wind Energy Grid‐ Adaptive Technologies, Jeju, Korea; 2014. 42. Lannoye, E., Flynn, D., and O’Malley, M. (2012). Evaluation of power system flexibility. IEEE Trans. Power Syst. 27: 922–931. 43. O’Dwyer, C. and Flynn, D. (2015). Using energy storage to manage high net load variability at sub‐hourly timescales. IEEE Trans. Power Syst. 30: 2139–2148. 44. Matevosyan, J., Olsson, M., and Söder, L. (2009). Hydropower planning coordinated with wind power in areas with congestion problems for trading on the spot and the regulating market. Electr. Power Syst. Res. 79: 39–48. 45. GE Energy (2014). Minnesota Renewable Energy Integration and Transmission Study. Schenectady, NY: Minnesota Utilities & Transmission Companies and Minnesota Department of Commerce. 46. Lew D, Brinkman G, Ibanez E, et al. The western wind and solar integration study—phase 2. Report NREL/ TP‐5500–55588, NREL, 2013. 47. Troy, N., Flynn, D., Milligan, M. et  al. (2012). Unit commitment with dynamic cycling costs. IEEE Trans. Power Syst. 27: 2196–2205. 48. Turconi, R., O’Dwyer, C., Flynn, D. et  al. (2014). Emissions from cycling of thermal power plant in electricity systems with high penetration of wind power: LCA for Ireland. Appl. Energy 131: 1–8. 49. Riesz, J. and Milligan, M. (2015). Designing electricity markets for a high penetration of variable renewables. Wiley Interdiscip. Rev. Energy Environ. 4: 279–289.

50. Lannoye, E., Flynn, D., and O’Malley, M. (2015). The impact of transmission networks on power system flexibility. IEEE Trans. Power Syst. 30: 57–66. 51. Meegahapola, L., Fox, B., Littler, T. et  al. (2013). Multiobjective reactive power support from wind farms for network performance enhancement. Eur. Trans. Electr. Power 23: 135–150. 52. Cuffe, P., Smith, P., and Keane, A. (2012). Transmission system impact of wind energy harvesting networks. IEEE Trans. Sustain. Energy 3: 643–651. 53. Energinet.dk. Cable action plan 132–150 kV. May 2011. 54. Hong, Y.Y. and Liao, C.M. (1995). Short‐term scheduling of reactive power controllers. IEEE Trans. Power Syst. 10: 860–868. 55. Qin N, Abildgaard H, Lund P et  al.Automatic voltage control (AVC) of Danish transmission system—concept design. In: Proceedings of the 2014 Grid of the Future Symposium, CIGRE USNC, Houston, TX, October, 2014. 56. Wallnerström, C.J., Huang, Y., and Söder, L. (2015). Impact from dynamic line rating on wind power integration. IEEE Trans. Smart Grid 6: 343–350. 57. Greenwood, D.M., Gentle, J.P., Myers, K.S. et  al. (2014). A comparison of real‐time thermal rating systems in the US and the UK. IEEE Trans. Power Del. 29: 1849–1858. 58. Oscar A. New trend in transmission power lines and related stringing equipment development. In: IEEE PES General Meeting, Vancouver, BC, Canada, 2013. 59. Tortora AC. Current and future role of energy storage in the grid. EU Sustain Energy Week June 2015, 2015. 60. Kundur, P. (1994). Power System Stability and Control, The EPRI Power System Engineering Series. New York, NY: MGraw‐Hill. 61. Matevosyan J. Wind power integration in power systems with transmission bottlenecks. In: IEEE PES General Meeting, Tampa, FL, 2007. 62. Haubrich H‐J, Zimmer C, von Sengbusch K, et al. Analysis of electricity network capacities and identification of congestion. Report commissioned by the European Commission Directorate‐General Energy and Transport, 2001 63. Vittal, E., O’Malley, M., and Keane, A. (2010). A steady‐state voltage stability analysis of power systems with high penetrations of wind. IEEE Trans. Power Syst. 25: 433–442. 64. Ellis A, Kazachkov Y, Muljadi E, et al. Description and technical specifications for generic WTG models—a status report. In: IEEE PES General Meeting, Phoenix, AZ, 2011 65. Sorensen P, Andresen B, Fortmann J, et  al. Overview, status and outline of the new IEC 61400–27—electrical simulation models for wind power generation. In: 10th Wind Integration Workshop, Aarhus, Denmark, 2011. 66. Pourbeik, P. (2014). Specification of the Second Generation Generic Models for Wind Turbine Generators. Knoxville, TN: Electric Power Research Institute. 67. IEC 61400‐27‐1. Electrical simulation models—wind turbines. Ed.1.0 2015–02, 2015.

Technical Impacts of High Penetration Levels of Wind Power on Power System Stability  63

68. Sorensen P, Fortmann J, Buendia FJ, et al. Final Draft International Standard IEC 61400–27–1. Electrical simulation models of wind turbines. In: 13th Wind Integration Workshop, Berlin, Germany, 2014. 69. Hansen, A.D., Altin, M., Margaris, I.D. et  al. (2014). Analysis of the short‐term overproduction capability of variable speed wind turbines. Renew. Energy J. 68: 326–336. 70. Ackermann, T., Ellis, A., Fortmann, J. et  al. (2013). Code shift—grid specifications and dynamic wind turbine models. IEEE Power Energy Mag. 11: 72–82. 71. Sørensen P, Andresen B, Bech J, Fortmann J, Pourbeik P. Progress in IEC 61400–27. In: 11th International Workshop on Large‐Scale Integration of Wind Power into Power Systems as well as on Transmission Networks for Offshore Wind Power Plants, Lisbon, Portugal, November, 2012. 72. Zhao, H., Wu, Q., Margaris, I. et  al. (2015). Implementation and validation of IEC generic type 1A wind turbine generator model. Int. Trans. Elect. Energy Syst. 25: 1804–1813. 73. Kristoffersen, J.R. and Christiansen, P. (2003). Horns Rev offshore windfarm: its main controller and remote control system. Wind Energ. 27: 351–360. 74. Hansen, A.D., Sorensen, P., Iov, F. et  al. (2006). Centralised power control of wind farm with doubly fed  induction generators. Renewable Energy 31: 935–951. 75. Miller N, Wachtel S, Longtin K, et  al. The GE hybrid wind turbine with integrated battery energy storage: concept, implementation and first field test results. In: 13th International Workshop on Large‐Scale Integration of Wind Power into Power Systems as well as on Transmission Networks for Offshore Wind Power Plants, Berlin, Germany, November, 2014. 76. Zeni L, Glasdam J, Hesselbæk B, et al. Coordinated system services from offshore wind power plants connected through HVDC networks. In: CIGRE Session 45, Paris, France; 2014. 77. Asmine, M., Brochu, J., Fortmann, J. et  al. (2011). Model validation for wind turbine generator models. IEEE Trans. Power Syst. 26: 1769–1782. 78. van der Meer AA, Hendriks RL, Kling WL. Combined stability and electro‐magnetic transients simulation of offshore wind power connected through multi‐terminal VSC‐HVDC. In: IEEE PES General Meeting, Minneapolis, MN, July, 2010. 79. IEEE Task Force on Load Representation for Dynamic Performance (1995). Standard load models for power flow and dynamic performance simulation. IEEE Trans. Power Syst. 10: 1302–1313. 80. WECC (2012). Composite Load Model for Dynamic Simulations. Salt Lake City, UT: Western Electricity Coordinating Council. 81. Welfonder, E., Weber, H., and Hall, B. (1989). Investigations of the frequency and voltage dependence of load part systems using a digital self‐acting measuring and identification system. IEEE Trans. Power Syst. 4: 19–25.

82. Chamorro HR, Ghandhari M, Eriksson R. Wind power impact on power system frequency response. In: North American Power Symposium, Manhattan, KS, 2013. 83. Singhvi V, Zhang Y, Pourbeik P, et  al. Impact of wind active power control strategies on frequency response of an interconnection. In: IEEE PES General Meeting, Vancouver, BC, Canada, 2013. 84. Zhang YC, Gevorgian V, Ela E, et al. Role of wind power in primary frequency response of an interconnection. NREL Report, 2013. 85. Sharma S. Frequency response in ERCOT. In: NREL/ EPRI Active Power Control from Wind Power Workshop II, May, Golden, Colorado; 2013. 86. Miller N, Shao M, Pajic S, et al. Western wind and solar integration study phase 3—frequency response and transient stability. NREL/SR‐5D00–62906, December, 2014. 87. Fox B, Littler T, Flynn D. Measurement‐based estimation of wind farm inertia. In: PowerTech 2005, St. Petersburg, Russia, June, 2005. 88. Gautam, D., Vittal, V., and Harbour, T. (2009). Impact of increased penetration of DFIG‐based wind turbine generators on transient and small signal stability of power systems. IEEE Trans. Power Syst. 24: 1426–1434. 89. Beddoes A, Thomas P, Gosden M. Loss of mains protection relay performances when subjected to network disturbances/events. In: CIRED, Turin, Italy, June, 2005. 90. Eto JH, Undrill J, Mackin P et  al., Use of frequency response metrics to assess the planning and operating requirements for reliable integration of variable renewable generation. Lawrence Berkley National Laboratory, LBNL‐4142E, December, 2010. 91. Commission for Energy Regulation (2014). Rate of Change of Frequency Modification to the Grid Code: Decision Paper. Dublin, Ireland: CER. 92. Meegahapola, L. and Flynn, D. (2015). Characterization of gas turbine lean‐blowout during frequency excursions in power networks. IEEE Trans. Power Syst. 30: 1877–1887. 93. NERC. Industry advisory: turbine combustor lean blowout, 2008. Available at: www.nerc.com. 94. Wang, Y., Silva, V., and Lopez‐Botet‐Zulueta, M. (2016). Impact of high penetration of variable renewable generation on frequency dynamics in the continental Europe interconnected system. IET Renew. Power Gener. 10: 10–16. 95. MacDowell, J., Dutta, S., Richwine, M. et  al. (2015). Serving the future. IEEE Power Energy Mag. 13: 22–30. 96. Morren, J., de Hann, S., Kling, W. et al. (2006). Wind turbines emulating inertia and supporting primary frequency control. IEEE Trans. Power Syst. 21: 433–434. 97. Miller, N. (2015). Keeping it together. IEEE Power Energy Mag. 13: 31–39. 98. Yingcheng, X. and Nengling, T. (2011). Review of contribution to frequency control through variable speed wind turbines. Renewable Energy 3: 1671–1677. 99. Miao, Z., Fan, L., Osborn, D. et al. (2010). Wind farms with HVDC delivery in inertial response and primary frequency control. IEEE Trans. Energy Conver. 25: 1171–1178.

64  Advances in Energy Systems

100. Ullah, N.R., Thiringer, T., and Karlsson, D. (2008). Temporary primary frequency control support by variable speed wind turbines—potential and applications. IEEE Trans. Power Syst. 23: 601–612. 101. Ruttledge, L. and Flynn, D. (2016). Emulated inertial response from wind turbines: gain scheduling and resource coordination. IEEE Trans. Power Syst. 31 (5): 3747–3755. https://doi.org/10.1109/tpwrs.2015. 2493058. 102. Brisebois J, Aubut N. Wind farm inertia emulation to fulfill Hydro‐Québec’s specific needs. In: IEEE PES General Meeting, San Diego, CA, July, 2011. 103. Rawn BG, Gibescu M, Kling WL. Kinetic energy from distributed wind farms: technical potential and implications. In: IEEE Innovative Smart Grid Technologies Europe, Gothenburg, Sweden, October, 2010. 104. Ruttledge L, Flynn D. System‐wide inertial response from fixed speed and variable speed wind turbines. In: IEEE PES General Meeting, San Diego, CA, July, 2011. 105. Ruttledge, L., Miller, N., O’Sullivan, J. et  al. (2012). Frequency response of power systems with variable speed wind turbines. IEEE Trans. Sustain. Energy 3: 683–691. 106. Wang, Y., Delille, G., Bayem, H. et  al. (2013). High wind power penetration in isolated power systems— assessment of wind inertial and primary frequency responses. IEEE Trans. Power Syst. 28: 2412–2420. 107. Fischer M, Engelken S, Mihov N, et  al. Operational experiences with inertial response provided by type 4 wind turbines. In: 13th International Workshop on Large‐ Scale Integration of Wind Power into Power Systems as well as on Transmission Networks for Offshore Wind Power Plants, Berlin, Germany, November, 2014. 108. Asmine M, Langlois C‐É. Field measurements for the assessment of inertial response for wind power plants based on Hydro‐Québec TransÉnergie requirements. In: 13th International Workshop on Large‐Scale Integration of Wind Power into Power Systems as well as on Transmission Networks for Offshore Wind Power Plants, Berlin, Germany, November, 2014. 109. de Martini P. DR 2.0—A Future of Customer Response. Newport Consulting, San Francisco: Association for Demand Response and Smart Grid; 2013. 110. Molina‐Garcia, A., Bouffard, F., and Kirschen, D. (2011). Decentralized demand‐side contribution to primary frequency control. IEEE Trans. Power Syst. 26: 411–419. 111. Qazi, H. and Flynn, D. (2016). Analysing the impact of large‐scale decentralised demand side response on frequency stability. Int. J. Electr. Power Energy Syst. 80: 1–9. https://doi.org/10.1016/J.IJEPES.2015.11.115. 112. Samsarakoon, K., Ekanayake, J., and Jenkins, N. (2012). Investigation of domestic load control to provide primary frequency response using smart meters. IEEE Trans. Smart Grid 3: 282–292. 113. Short, J., Infield, D., and Freris, L. (2007). Stabilization of grid frequency through dynamic demand control. IEEE Trans. Power Syst. 22: 1284–1293.

114. Xu, Z., Ostergaard, J., and Togeby, M. (2011). Demand as frequency controlled reserve. IEEE Trans. Power Syst. 26: 1062–1071. 115. Tao, W. (2013). Recharging China’s electric vehicle policy. In: Policy Outlook. Beijing: Carnegie‐Tsinghua Center for Global Policy. 116. Shengnan, S., Pipattanasomporn, M., and Rahman, S. (2011). Demand response as a load shaping tool in an intelligent grid with electric vehicles. IEEE Trans. Smart Grid 2: 624–631. 117. European Photovoltaic Industry Association (2012). Connecting the Sun: Solar Photovoltaics on the Road to Large‐Scale Grid Integration. Brussels, Belgium: EPIA. 118. Hsieh, E. and Johnson, R. (2012). Frequency Response from Autonomous Battery Energy Storage. Paris, France: CIGRE US National Committee. 119. U.S. Department of Energy (2013). Grid Energy Storage. Washington, DC: U.S. Department of Energy. 120. D’Arco S, Suul JA. Virtual synchronous machines‐ classification of implementations and analysis of equivalence to droop controllers for microgrids. In: IEEE PowerTech, Grenoble, France, June, 2013. 121. D’Arco S, Suul JA, Fosso OB. Control system tuning and stability analysis of virtual synchronous machines. In: IEEE Energy Conversion Congress and Exposition, Denver, CO, September, 2013. 122. Boemer JC, van der Meer AA, Rawn BG, et  al. Network fault response of wind power plants in distribution systems during reverse power flows. In: 10th International Workshop on Large‐Scale Integration of Wind Power into Power Systems, Arhus, Denmark, October, 2011. 123. Gómez‐Lázaro, E., Fuentes, J.A., Molina‐Garcia, A. et al. (2007). Field tests of wind turbines submitted to real voltage dips under the new Spanish grid code requirements. Wind Energ. 10: 483–495. 124. EirGrid. Grid Code v6.0 July 2015. Available at: http:// www.eirgridgroup.com/site‐files/library/EirGrid/ GridCodeVersion6.pdf. 125. Estanqueiro, A., Castro, R., Flores, P. et  al. (2008). How to prepare a power system for 15% wind energy penetration: the Portuguese case study. Wind Energ. 11: 75–84. 126. van der Meer AA, Hendriks RL, Kling WL. A survey of fast power reduction methods for VSC connected wind power plants consisting of different turbine types. In: 2nd EPE Wind Energy Chapter Seminar, Stockholm, Sweden, April, 2009. 127. Rather Z, Flynn D. Voltage dip induced frequency events in wind integrated power systems. In: 9th IFAC Control of Power Plants and Energy Systems Symposium, New Delhi, India, December, 2015. 128. McMullan L, Horan P, Gallery T, et al. Medium‐term dynamic studies for a large island power system with high levels of wind. In: 13th International Workshop on Large‐Scale Integration of Wind Power into Power Systems as well as on Transmission Networks for Offshore Wind Power Plants, Berlin, Germany, November, 2014.

Technical Impacts of High Penetration Levels of Wind Power on Power System Stability  65

129. ENTSO‐E (2011). Offshore Transmission Technology. Brussels, Belgium: ENTSO‐E. 130. Abildgaard H. Installation of synchronous condensers for reliable HVDC operation and bulk power transfer. In: IEEE PES General Meeting, Denver, CO, 2015. 131. You, R., Barahona, B., Chai, J. et al. (2013). A novel wind turbine concept based on an electromagnetic coupler and the study of its fault ride‐through capability. Energies 6: 6120–6136. 132. Knüppel T, Akhmatov V, Nielsen JN, et al. On small‐signal stability of wind power system with full‐load converter interfaced wind turbines. In: AWEA WINDPOWER Conference & Exhibition, Chicago, IL, May, 2009. 133. Aherne, R. (2015). DS3 Programme Status. Dublin, Ireland: EirGrid and SONI DS3 Industry Forum. 134. Rather Z, Chen Z, Thogersen P. Impact of renewable energy integration on reactive power reserve and its smart solution: a Danish power system case study. In: IEEE Conference on Power System Technology (POWERCON), Auckland, New Zealand, 2012. 135. Transpower (2007). Wind Generation Investigation Project 1B: Effect of Unpredictability of Wind Generation Output on Scheduling. Planning and Investigations. Wellington, New Zealand: Transpower. 136. Liu C, Rather ZH, Bak CL, et  al. Importance sampling based decision trees for security assessment and the corresponding preventive control schemes: a Danish case study. In: IEEE PowerTech, Grenoble, France, June, 2013. 137. Vijayan P, Sarkar S, Ajjarapu V. A novel voltage stability assessment tool to incorporate wind variability. In: IEEE PES General Meeting, Calgary, AB, Canada, 2009. 138. Paserba, J. (ed.) (1996). Analysis and Control of Power System Oscillations. Paris: CIGRE Technical Brochure. 139. Transpower (2008). Wind Generation Investigation Project 8: Effect of Wind Generation on Small‐signal Stability. Planning and Investigations. Wellington, New Zealand: Transpower. 140. Slootweg, J.G. and Kling, W.L. (2003). The impact of large scale wind power generation on power system oscillations. Electr. Power Syst. Res. 67: 9–20. 141. Fan L, Miao Z, Osborn D. Impact of doubly fed wind turbine generation on inter‐area oscillation damping. In: IEEE PES General Meeting, Pittsburgh, PA, July, 2008. 142. Nguyen MH, Saha TK, Eghbal M. Impact of high level of renewable energy penetration on inter‐area oscillation. In: 21st Universities Power Engineering Conference (AUPEC), Brisbane, Australia, September, 2011. 143. Qiao W, Harley RG. Effect of grid‐connected DFIG wind turbines on power system transient stability. In: IEEE PES General Meeting, Pittsburgh, PA, July, 2008. 144. Fayek HM, Elamvazuthi I, Perumal N, Venkatesh B. The impact of DFIG and FSIG wind farms on the small signal stability of a power system. In: 5th International Conference on Intelligent and Advanced Systems (ICIAS), Kuala Lumpur, Malaysia, 2014. 145. Tsourakis, G., Nomikos, B.M., and Vournas, C.D. (2009). Contribution of doubly fed wind generators to oscillation damping. IEEE Trans. Energy Conver. 24: 783–791.

146. Chaudhuri NR, Chaudhuri B. Impact of wind penetration and HVDC upgrades on dynamic performance of future grids. In: IEEE PES General Meeting, San Diego, CA, July, 2011. 147. Mehta, B., Bhatt, P., and Pandya, V. (2014). Small signal stability analysis of power systems with DFIG based wind power penetration. Electr. Power Energy Syst. 58: 64–74. 148. Tsourakis, G., Nomikos, B.M., and Vournas, C.D. (2009). Effect of wind parks with doubly fed asynchronous generators on small‐signal stability. Electr. Power Syst. Res. 79: 190–200. 149. Woong Shim J, Verbic G, Hur K, Hill DJ. Impact analysis of variable generation on small signal stability. In: Australasian Universities Power Engineering Conference (AUPEC), Perth, WA, 2014. 150. Knuppel, T., Nielsen, J.N., Jensen, K.H. et al. (2012). Small‐signal stability of wind power system with full‐ load converter interfaced wind turbines. IET Renew. Power Gener. 6: 79–91. 151. Mishra, Y., Mishra, S., Li, F. et al. (2009). Small‐signal stability analysis of a DFIG‐based wind power system under different modes of operation. IEEE Trans. Energy Conver. 24: 972–982. 152. Rueda J, Erlich I. Impacts of large scale integration of wind power on power system small‐signal stability. In: 4th International Conference on Electric Utility Deregulation and Restructuring and Power Technologies (DRPT), Weihai, China, July, 2011. 153. Shi L, Dai S, Ni Y, Yao L, Bazargan M. Transient stability of power systems with high penetration of DFIG based wind farms. In: IEEE PES General Meeting, Calgary, AB, Canada, July, 2009. 154. Ullah, N.R. and Thiringer, T. (2008). Effect of operational modes of a wind farm on the transient stability of nearby generators and on power oscillations: a Nordic grid study. Wind Energ. 11: 63–73. 155. Yu C, James G, Xue Y, Xue F. Impacts of large scale wind power on power system transient stability. In: 4th International Conference on Electric Utility Deregulation and Restructuring and Power Technologies (DRPT), Aswan, Egypt, July, 2011. 156. Jindal AK, Irwin GD, Woodford DA. Subsynchronous interactions with wind farms connected near series compensated AC lines. In: 9th International Workshop on Large‐Scale Integration of Wind Power into Power Systems as well as on Transmission Networks for Offshore Wind Power Plants, Quebec City, Canada, October, 2010. 157. Massimo, B., Andreas, P., and Evert, A. (2011). The Impact of Wind Farms on Sub‐Synchronous Resonance in Power Systems. Stockholm, Sweden: ELFORSK. 158. Varma, R. and Moharana, A. (2013). SSR in double‐ cage induction generator based wind farm connected to series‐compensated transmission line. IEEE Trans. Power Syst. 28: 2573–2583. 159. Leon, R. and Solsona, J.A. (2015). Sub‐synchronous interaction damping control for DFIG wind turbines. IEEE Trans. Power Syst. 30: 419–428.

4

Understanding Constraints to the Transformation Rate of Global Energy Infrastructure Joe L. Lane1, Simon Smart1, Diego Schmeda‐Lopez1, Ove Hoegh‐Guldberg2, Andrew Garnett3, Chris Greig4 and Eric McFarland1 Dow Centre for Sustainable Engineering Innovation, Department of Chemical Engineering, The University of Queensland, St. Lucia, Australia 2  Global Change Institute, The University of Queensland, St. Lucia, Australia 3  Center for Coal Seam Gas, Sustainable Mining Institute, The University of Queensland, St. Lucia, Australia 4  UQ Energy Initiative, The University of Queensland, St. Lucia, Australia

1 

A massive transformation of the global energy supply system is required if deep reductions in atmospheric carbon dioxide emissions are to be achieved. A top‐down review of historical data and energy forecasts provides a perspective on the magnitude of the challenge. Global engineering capability has expanded significantly over the last two decades, accommodating more than 100 GW yr−1 increase in electricity generation infrastructure. However, business‐as‐usual (BAU) demand forecasts to 2050 will require more than double the global rates of energy project execution. Transforming to a low‐carbon energy supply mix requires 30–70 GW yr−1 of additional infrastructure, due to the increased reliance on intermittent renewables, and the earlier‐than‐expected replacement of existing coal power plants. Although all power systems share many similar subsystems that will need to be delivered regardless of the technology type, meeting the extra demands for engineering design, construction and/or supply chains may not be possible. The discussion focuses only on physical limitations of electricity generation, specifically around the timing and scale of retiring and/or replacing coal‐fired power generation capacity to meet the International

Energy Agency’s (IEAs) two‐degree scenario. We ignore the economics and politics of the transition scenarios and the transformation of the transportation and industrial sectors. What is clear is that the longer the delay in starting a significant transformation, the greater the challenge will become. Decision makers must understand the constraints to technology transitions to deliver effective policy. A broad international consensus is not required; instead, reaching agreements and developing economically sustainable pathways to technology transitions in the United States, China, and India is more likely to be successful and the only means for significantly curbing global emissions. ­INTRODUCTION An enormous, and unprecedented, transformation of global energy supply systems is required to provide the energy services needed to sustain increasing prosperity worldwide; while at the same time reducing carbon dioxide emissions low enough to limit atmospheric temperature rise and potentially serious ­ environmental impacts. Current emission

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

68  Advances in Energy Systems

trajectories suggest that, without significant change in the trends for industrial capacity or carbon intensity of the economy, average global temperature is on track for a 3.7–4.8 °C rise above the Preindustrial Period – which could have devastating consequences for human beings[1, 2]. Even under conservative projections of the change required to avoid that outcome, the dominant role of fossil fuels would have to end rapidly, falling from over 80% of global energy supply to 40% or lower by mid‐century[2, 3]. The energy output from renewable sources and nuclear power would need to increase nearly fourfold to fill the supply gap[2, 4–6]. Ongoing population growth and energy intensification in the developing economies are expected to require an increase in global primary energy demand of between 37% and 50% by 2040 under the IEA’s “New Policies” and “Current Policies” scenarios, respectively[5]. The greatest growth will be in developing regions, many of which will see demand increase by two to three times, while demand in Organization for Economic Cooperation and Development (OECD) countries is expected to remain relatively constant, potentially even declining, over the same period. The combination of strong regional demand growth and the decarbonization of the energy supply requires an enormous effort in design, construction, and start‐ up of new energy supply infrastructure. The scale of the infrastructure rollout would vary substantially across different regions, as would the practical challenges involved. International efforts to limit atmospheric greenhouse gas (GHG) emissions have largely focused on obtaining agreements on targets and possible incentives to achieve them[2]. There have been studies on resource limitations[7–10] that might constrain the overall capacity of certain nonfossil energy technologies[11–13]. However, little attention has been given to the question of what are the practically achievable rates of transforming the energy infrastructure, even if there were a global consensus to do so, and, even less on understanding the engineering and logistic factors (e.g. design skills; component supply chains; construction delivery) that will likely limit that achievable rate of transition. Our imperfect understanding of CO2–climate system interactions, makes climate‐change policy an exercise in risk reduction[3]. It is likely that more detailed understanding of these processes will alter the required emissions reduction targets and rates needed to achieve them. Central to the analysis of risk reduction is cognizance of how rapidly mitigation measures can be deployed. The longer it takes to

implement a transformation, the sooner we may need to act[3, 14]. This article focuses on one element of global energy services–power generation. It explores three aspects of the challenges involved with rapidly transforming global power production: 1. What infrastructure deployment rates are possible? We consider historical evidence of infrastructure growth rates that have been achieved, and opportunities to change or overcome the rate‐limiting constraints. 2. What are the extra burdens imposed by transforming the power generation systems away from the fossil‐fuel‐dependent, BAU, forecasts? We look at both the similarities and differences across different supply technologies, and at more direct stimulus, to understand their effect on overall infrastructure requirements. 3. How challenging the burden of early asset retirement might become, in the pursuit of ambitious yet necessary GHG mitigation targets? The example of coal‐fired electricity is used to illustrate that this requirement for early retirement and replacement of existing plants might be ­substantial, particularly if the transformation is delayed. The complexities of shifting the transportation sector from fossil fuels to a mix of electric and biofuel‐based vehicles have not been studied for this analysis, nor have the details of transitioning industrial processing and heating. The electricity demand for transport will account for only 13% of the total energy demand in 2050 according to the model in the Energy Technology Perspectives (ETP) 2014[4], and biofuel refineries are assumed to require similar engineering and material assets as would be required under a BAU scenario. ­ HAT IS POSSIBLE? – HISTORICAL W (AND FUTURE) CONTEXT Rates of Infrastructure Rollout At a global level, approximately 105  GW  yr−1 of new fossil‐ and nuclear‐fueled electricity generation capacity has been commissioned over the last two decades, dominated by growth in coal (33%) powered plants (Figure  4.1). The indicative “rate” shown for the plants under construction illustrates that the scale of growth has not diminished in recent years. Furthermore, given the expected growth in energy demand and the need to replace aging infrastructure, the IEA’s

Understanding Constraints to the Transformation Rate of Global Energy Infrastructure  69

200

5

Oil

4

Gas 120

Coal

3

China

Under Construction

2005–2014

1995–2004

0 1985–1994

0 1975–1984

1

1965–1974

40

1955–1964

2

1945–1954

80

Total installed capacity (TW)

160

1935–1944

Capacity increase (GW/year)

Nuclear

Figure 4.1  Historical rates of installing new fossil‐ and nuclear‐fueled electricity generation plant over the last century. The bars and left‐hand vertical axis show the increase in (average) net installed capacity for each 10‐year period. The dotted line and right‐hand axis show the cumulative total installed capacity at each point in time. The rate for plants under construction was calculated using the capacity under construction on the database [15] and each category was divided by construction timespans using the upper bound of IEA/WEO14 [5] (10 years for nuclear plant, 5 years for coal, and 3 years for oil and gas). Source: Global Data Power database [15].

mid‐case emissions estimate,1 requires sustained r­ollout of new fossils‐, nuclear‐, and renewables‐ based generation plant at unprecedented levels. Their scenario requires that fossil‐ and nuclear‐fueled plant installations continue at rates of ∼139 GW yr−1 for at least 20 years – nearly three new, large‐scale (1 GW) plants every week – supplemented with an additional 149 GW yr−1 of installed capacity from renewable sources[5]. The maximum rate achieved historically has been approximately 120 GW yr−1[15]. Whether or not the higher development rates can be achieved and sustained remains to be seen. Successful delivery of that infrastructure will depend in part on the political, technical, and business capability of the developing regions driving future infrastructure growth. Critical to executing a transition are the engineering‐procurement‐construction (EPC) firms responsible for the delivery of large engineering “megaprojects” (>$1  billion USD, 2003)[16], such

1  The “New Policies” scenario of the IEA (approximately the same as the ETP 4DS) is based on policy settings that continue the trend toward renewable energy uptake but without concerted action to meet the 2 °C temperature target[3].

as power plant construction. Recent trends in EPC activity suggest this crucial sector is capable of the growth required to service the increased demands for energy infrastructure[17]. The construction of electricity generation and distribution infrastructure makes up ∼11% of total activity of the international EPC contracting firms[18]. If we assume that the rate of rolling out electricity supply infrastructure will need to increase by 85%2 that might increase total

2  This estimate for increase in generation plant construction assumes a current rollout rate of 155 GW/year and a future value of 288 GW/year. The current value is estimated by combining 105 GW/year of fossil‐ and nuclear‐fueled power plant (as described above); and 50 GW/year of renewables generation plant as approximated from the difference in the IEA estimates for installed capacity in 2012[4] and 1992[19]. The future value of 288 GW/year combines the estimates described above – 139 GW/year of plant capacity using nonrenewable sources, and 149 GW/year of plant using renewable sources. For the sake of simplicity, combining the nonrenewable and renewable capacity estimates into a single value overlooks any potential differences in the implications for an EPC of providing plants using renewables versus nonrenewable sources.

70  Advances in Energy Systems

EPC activity by ∼9% all else being equal. The IEA also forecasts that investment requirements (to 2040) in oil and gas infrastructure will be 1.5 times higher than for electricity services[5], hence this might add another 14% to demand for EPC services depending on whether or not the transportation fuels sector is also making a transition. However, these increases appear relatively modest, given that overall EPC revenue (now valued at ∼$1.4 trillion[18]) has increased by a factor of ∼3 over the last decade[17]. Much of that growth is associated with the rise of domestic firms in non‐OECD countries, indicating there is existing capability to operate in the developing markets that will dominate future energy infrastructure growth. However, the EPC sector also needs to cope with other demands of a growing global economy such as the anticipated $350 trillion (USD, 2000) investment required for urban infrastructure and operations over the next three decades[20, 21]. Construction in developing economies will represent ∼34% of this activity[22]. Of course, EPC business growth is neither necessarily an indicator of capability, nor is EPC business capability the only important consideration in terms of large energy project delivery. Various studies have demonstrated that mega‐projects are plagued by cost  overruns, schedule delays and reduced performance regardless of the industry or global EPC status[23, 24]. These issues are only compounded during infrastructure booms where additional and often unforeseen supply‐chain bottlenecks become apparent. The current US shale gas boom is a case in point, where key unit operations have doubled in both cost and expected delivery time due to extraordinary demand, yet in terms of global EPC capabilities, the increase has been small. This effect of cost escalation is independent on the geographical location of the project, since similar projects often compete for the same limited resources, resulting in strains within the  supply chain (already seen in US natural gas ­projects as well as global wind power installations[21, 25]), and consequent increases in costs and wages[22]. Likewise, the oil refinery and process plant build‐outs of the 1970s and 1980s[24, 26, 27] and rise of nuclear power in the United States[28] over the same time period saw substantial supply‐chain constraints. The availability of suitably skilled workers and engineers is another important consideration. With the rapid expansion of chemical and pipeline facilities associated with the US shale gas boom, there has been a severe shortage of skilled welders, which has caused delays and cost increases[29]. Despite impressive growth in the engineering labor pool over the last two decades, it is still an open question as to how

quickly that workforce might be retrained to support a massive expansion in the design and deployment of new low‐carbon power technologies. The rapid rise in nuclear power plant construction in the 1970–1980s provides an interesting historical example, as it required a much more specialized workforce than for conventional energy supply technologies. The rollout of nuclear power plants peaked at ∼25 GW yr−1 at a time when there were ∼200 000 qualified engineers in the world, albeit largely unskilled in the design and implementation of nuclear technology. Today, “elite universities”[30] worldwide are providing almost six times as many engineering graduates[31, 32], with China and India accounting for ∼19% of the annual graduation output[33]. The total number of engineering graduates is significantly larger if all universities are included. The fraction of total engineering students graduating from “nonelite” universities in India and China is ∼96% and ∼83%, respectively, and growing[30, 34]. It remains to be seen whether the burgeoning engineering workforce will have the necessary experience and skills needed for the industry to keep pace with the anticipated growing demand for new power plants on top of the other demands of a growing economy. Experience in several engineering sectors suggests that despite adequate numbers, mismatches between project demands and available engineering skills and experience lead to massive project delays and cost overruns. Different Models for Infrastructure Delivery Different industrialization pathways and production processes will have very different constraints. Starting from a low manufacturing and skills base in the early 1990s, China’s annual motor vehicle output has increased by 15‐fold, to the point it now produces more than 22 million vehicles/year  –  the equivalent of approximately 1100 GW yr−1 of thermal energy capacity in the vehicle engines (Figure 4.2). This rate is nearly two orders of magnitude higher (in energy terms) than the rollout of large‐scale electricity generation plants (40 GW yr−1) in China over the last decade. Standardization of designs, assembly‐line manufacturing, and modularization of major subsystems for power generation facilities might help overcome some of the rate‐limiting constraints to plant construction. Wind and solar photovoltaics (PV) technologies already benefit from mass‐production of key components[31, 33]. Recent innovations in nuclear reactor design have already raised the prospect of an increasing role for small modular nuclear reactors[35]. These modular subsystems might be sufficiently

Understanding Constraints to the Transformation Rate of Global Energy Infrastructure  71

12,000

80 Power capacity built (thermal) - Chinese automotive sector

Power capacity built (GW/year)

60 8,000

50

6,000

40 30

4,000

20 2,000

Cummulative capacity (TW)

70

10,000

10

0

0 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20

Figure 4.2  Historical evolution of the automobile industry in China[49] shown as a function of the thermal energy output of the engines. As a proxy for an “energy system,” each vehicle engine represents approximately 500 kW of thermal energy generation capacity (100 kW mechanical), hence the rapid growth since 2008 represents an annual output of 215 GW yr−1. The continuous line is the cumulative total capacity and is linked to the right hand side vertical axis.

small to benefit from assembly‐line/mass‐production manufacturing, quality assurance, and inspection ­ approaches. Large (∼50 MW) aircraft‐style jet engines are factory produced routinely at rates of approximately one per day, 18 GW yr−1 in a single facility. Smaller power units might be combined at a single facility for the equivalent output of a large‐scale power plant. A recent market report suggests that both the gas and steam turbine market was moving in the direction of small to medium turbines suited to flexible generation[29]. Another possible means for decreasing the needed rate of transformation to all new technologies is through conversion of existing power plants to utilize less greenhouse‐gas intensive fuel sources and/ or technologies. Selected older coal‐fired plants in the United States, otherwise slated for closure, are converting to the use of natural gas[36]. Elsewhere, the first conversions of conventional oil refineries into biofuel processing plants have been delivered or are in planning[37–41]. While promising, it is doubtful whether this current conversion activity will continue at sufficient scale and speed to significantly help with the goal of limiting atmospheric temperature increase. Our experience is that the repurposing of existing plants can end up being riskier, slower, and require greater capital spending than new construction, and that this introduces a substantial disincentive for utilities to attempt such a retrofit[42].

Reducing the “conversion risks” might be achievable if the initial plant design deliberately anticipates specific conversions at a later date[43]. Such a step might also open the door to more creative conversions in the future. For example, it might be feasible to substitute nuclear reactors, or solar‐thermal heat sources, for conventional boilers in a coal‐fired power plant, especially if that option were designed up front. While that will increase the up‐front costs for certain higher‐ grade components (e.g. turbines) and/or larger plant footprints, it could also conceivably reduce the risk for utilities that future changes in legislation might force early retirement of the capital asset.

­WHAT EXTRA BURDENS DOES AN ENERGY TRANSFORMATION INTRODUCE? Using the IEA forecasts as a guide (Figure 4.3) four important elements to the transformation story are described. Infrastructure Construction That Would Have Happened Anyway The first two elements of the transformation of the global power generation fleet will happen under all possible future scenarios. At first glance, GHG mitigation drivers are unlikely to significantly influence

72  Advances in Energy Systems

(a) Total energy demand (EJ/year)

700 Non - energy

650

Industry

600

Buildings/other

550 500

Transport Transport electrification

GHG scenario in 2050

Electricity

450

Total GHG

400

Total BAU

350 2011 2015

2020 2025

2030 2035 2040 2045

2050

Year of reference

(b) Change in demand (TWh/year) for fossil-fired electricity

35,000 30,000

Demand reduction Shift to renewables

25,000

Shift to biomass

20,000

Shift to nuclear

15,000

Install CCS

10,000 5,000 0 2011

2015

2020

2025

2030

2035

2040

2045

2050

(c)

Installed capacity (TW)

14 12 10

Renewables Biomass Nuclear Natural gas with CCS

8

Natural gas

6

Coal

4 2

Coal with CCS Oil Total BAU (high renewables) Total BAU Total GHG

0 2011 2015 2020 2025 2030 2035 2040 2045 2050 Figure 4.3  The changing role for electricity, comparing forecasts for business‐as‐usual (BAU) energy system growth under a scenario that achieves greenhouse gas (GHG) mitigation consistent with a 2 °C atmospheric temperature increase (ETP2014 study[4]). (a) Compares total energy demand, showing the portion of overall change attributed to the different sectors. The inset provides the demand profile for the GHG scenario in the year 2050. (b) Provides a pathway to achieve the GHG mitigation scenario by reducing demand and shifting the generation from fossil fuels to fossil‐free technologies. (c) Shows the evolution of the energy generation mix for the GHG mitigation scenario (colored areas). Also shown is the increase in capacity needed due to the removal of fossil fuels from the mix, e.g. the total installed capacity (in the year 2050) predicted for the GHG mitigation scenario is 0.7 TW higher than that for the “New Policies” scenario (green dotted line), and 1.9 TW higher than that for the BAU scenario (black dotted line). Those increases represent an additional 17 or 49 GW yr−1 (respectively) of infrastructure that must be constructed (on average) over the forecast period.

Understanding Constraints to the Transformation Rate of Global Energy Infrastructure  73

the overall infrastructure delivery scale for either of these elements. 1. Future electricity demand growth would be most strongly influenced by population growth and the intensification of energy use in developing economies (Figure  4.3a, b). GHG mitigation policy might have some influence if it encourages better demand management, but only to the extent that it slows the rate of overall growth. In the IEA/ ETP2014 scenario considered here (see, Figure  4.3a), these demand management measures offset and indeed outweigh the upward pressure caused by a partial (11%) electrification of the land‐based transport system. 2. Plant retirements will be largely driven by regulatory compliance and competitiveness in the electricity market. There may also be some shift in demand from fossil fuel based generators to low‐carbon sources, made possible by forces that have little to do with GHG mitigation policy. Drivers for this “natural retirement” will vary over time and across regions, but will invariably assist in meeting emission reduction targets, often at reduced economic impact (e.g. after the initial capital investment has been recouped). In some cases, aging plants will be decommissioned as their short‐run marginal costs rise above those of alternative energy sources. The current bout of coal plant closures in the United States is an example of this, instigated by a sharp drop in natural gas prices[36]. In other cases, concerns over local air quality might be the driver for closure of coal plants, such as those occurring in China. The drivers of demand growth and natural retirement both create the need for construction of new infrastructure, regardless of the choice of energy source. Any additional burden introduced by the transformation to a low‐carbon supply mix would therefore be the net difference between the rate of installing that low‐carbon supply infrastructure, and the rate of installing the fossil‐fuel‐based supplies that would have otherwise occurred. Quantifying these net differences would require a detailed, bottom‐up investigation into rate‐limiting factors for each technology option. Given that the available energy supply alternatives share many basic components and subsystems, it may be that the net differences involved are not that great. For example, there are many commonalities in the process equipment required for conventional and biomass‐based oil refineries, to the extent that conversions from the former to the latter are already happening

(see above). However, most projects will require some level of uniqueness especially around balance‐of‐plant systems. The unit operations comparison across electricity generation options is somewhat more complex; however, there are more similarities than differ­ ences between the conventional and nonconventional thermal technologies (Table  4.1). The nonthermal power supply technologies have a potentially simpler ­component set, however, far more units are required. While not considered here, the importance of differences in electrical distribution and storage infrastructure should not be overlooked, if the use of renewable supply technologies requires a more decentralized network. The Drivers of Additional Infrastructure Construction Two additional factors that will increase the overall scale of a low‐carbon energy infrastructure transition are: 1. Along with the “natural replacement” of aged assets, sufficient GHG mitigation would only be possible by incentivizing (or forcing) additional “early” retirement of fossil fuel generation plant. Since that must be replaced with new, low‐carbon supply sources, this increases the overall new infrastructure capacity required beyond that required for a BAU future. 2. The reduced capacity factor for the electricity supply system resulting from the replacement of fossil‐fuel power plants with intermittent sources such as PV and wind. To accommodate low‐ capacity factor sources requires an increased overall capacity of plant in service. This represents a net addition to the total infrastructure rollout that will be required in the future. For the IEA scenarios considered here, this penalty amounts for an extra 17–49 GW yr−1 of plants that must be installed over the 39‐year period, depending on the relative rates of future uptake of renewable technologies included in the BAU benchmark. Note also the significant additional infrastructure required for transmission, distribution, and storage associated with intermittent generation systems[4, 44]. ­HOW SIGNIFICANT IS THE EARLY REPLACEMENT CHALLENGE? Early retirement of existing fossil‐fuel powered generation plants will be necessary to deliver on GHG mitigation targets. Here a simplified, top‐down, approach

74  Advances in Energy Systems

Table 4.1 Unit operations required for different electricity supply technologies.

Coal Coal + CCS CCGT OCGT Nuclear Geo‐thermal Solar thermal Hydro Wind Solar PV

Feed

Waste

Heat generation

Heat exchange

Cooling

Turbine

Electricity generation Electrical conversion Specialized components

Rail Rail Pipeline Pipeline — — — — — —

Slag disposal Pipeline and wells — — Treat and store — — — — —

Boiler Boiler Combustion chamber Combustion chamber Reactor — Receiver — — —

Reheat Reheat HRSG — Steam generator Steam generator Steam generator — — —

Water/air Water Water/air Air Water/gas Water Water — — —

Steam Steam Gas/steam Gas Steam/gas Steam Steam Mech Mech —

AC AC AC AC AC AC AC AC AC PV panels

CCGT, combined‐cycle gas turbine; OCGT, open‐cycle gas turbine; HRSG, heat recovery and steam generation.

Transformer Transformer Transformer Transformer Transformer Transformer Transformer Transformer Transformer Inverter/transformer

— Compressor Compressor Compressor Containment Wells Thermal storage Dam/reservoir Towers —

Understanding Constraints to the Transformation Rate of Global Energy Infrastructure  75

is used to explore the significance of the early retirement of coal‐fired electricity generation plants. The example uses coal‐fired plants because of their importance in global electricity supply (Figure  4.1), and because it is the coal sector that is reduced most substantially in the IEA scenarios for GHG abatement (Figure 4.3c). The estimates developed use the following two scenarios, taken from the 2014 ETP modeling study3 of the IEA[4]. These scenarios were chosen because of the prominent role that the IEA models play in the international policy debate. In both cases, the ETP model simulations are delayed by five years, assuming the same forecast profiles and the same starting point in terms of demand and supply mix, but with a 2015 (rather than 2011) start date. • A BAU forecast, using the 6DS (6° average global temperature rise scenario of the ETP study[4]). This forecast is largely an extension of existing trends in evolution of the energy system, albeit strongly influenced by the IEA assumptions on future gross domestic product (GDP) growth. • A GHG forecast, taken from the 2DS (2° average global temperature rise scenario of the ETP study[4]). This provides an emissions trajectory considered to be consistent with the requirements to give at least a 50% chance of limiting average global temperature increase to 2°[3]. Our estimate for early replacement uses a seven‐ region representation of the world: the European Union (EU), the United States, the remainder of the OECD, remainder of non‐OECD, India, China, and the remainder of Southeast Asia (Figure  4.4). The intention is to capture key regional differences in the age of existing coal plant, and in the future role that coal‐fired power plays in the energy supply system. The global age distribution of coal‐fired generation plants is bimodal (Figure  4.4a). China has the youngest fleet (with a median age of 10 years), and virtually all plants in most regions of China are less than the nominal 50‐year “technical lifespan” of such infrastructure. The only fleets with substantial capacity older than that are in the United States and European Union, which have a collective median

3  The IEA also produces future modeling simulations under the World Energy Outlook (WEO) series of reports – e.g. the WEO 2014 study[5]. The ETP simulation results were used here because, of the two IEA approaches, that study provides a greater level of the detail (in the publicly available output) of use to our analysis.

age of 41 years. Under BAU forecasts, coal‐plant capacity in China, India, and Southeast Asia continues growing, to the point that it would comprise 75% of the global total by mid‐century, Figure 4.4b. Over that same period, substantial reductions are predicted for the coal‐plant fleet of the United States (by 40%) and EU (by 75%). The 2DS scenario assumes that mitigation targets can be reached if[1] the global stock of coal plants is reduced to ∼1 TW of installed capacity by 2050 (45% of installed capacity in 2011); and half of these remaining plants utilize carbon capture and storage (CCS; Figure 4.4c). That reduction includes the further decommissioning of 170 GW of coal plants in the United States, taking its fleet to 50 years, excluding any plant that has already been retired up to that point. The age distribution of the current stock of coal‐fired plants was used in the calculations (Figure 4.5). Estimating the Rate of Early Replacement For each of the seven geographical regions, the capacity of the plants that are retired early (CER) is defined as:

CER

CBAU

GHG

ACG

C NR

where ΔCBAU − GHG is the difference in capacity, between the BAU and GHG scenarios; ACG is the capacity growth that is avoided by the GHG scenario; CNR is the capacity of coal‐fired plants that is replaced with noncoal alternatives, for reasons other than GHG mitigation (natural replacement). The rate of early replacement (RER) is calculated as the cumulative sum of plant capacity that had to be retired over the transformation period, averaged across the length (in years) of that period. The need for early replacement varied substantially across the seven regions (Figure  4.6). For the United States and EU, the capacity of plants required for early replacement is simply the difference between the BAU and 2DS scenarios. For regions such as India, in the early stages of rapid growth in energy demand, the situation is very different. In India’s case, the GHG scenario involves a stabilization of growth in coal‐plant capacity, rather than an absolute decline. As a result, there is little need for the replacement of coal‐fired assets other than that required for the natural retirement of aged plants. China, however, does see a substantial need for early replacement of infrastructure toward the end of the forecast period, as it shifts from its growth phase into one of coal‐plant substitution.

International commitment to significant emissions reduction and a major transformation of global energy systems appears to remain some way off[45]. It is therefore prudent to consider how much the scale of the early replacement burden might increase, the longer it takes for the transformation to begin in earnest. By recalculating the early replacement rates using hypothetical alternatives for the 2DS scenario, assuming that global agreement to commence a serious, rapid transformation of the energy system would be delayed until 2020, 2025, and 2030. In each case, the coal‐ plant capacity forecasts were assigned by[1] assuming that the world’s power systems evolve as per the BAU forecast during the delay period;[2] post‐delay evolution of coal‐plant capacity follows a similar functional form to that of the default 2DS scenario; and[3] the year 2055 capacity endpoint remains constant. Importantly, these hypothetical forecasts would not achieve the same level of GHG mitigation as the default ETP scenario, as the delay in acting will ensure that existing coal plants stay online during the delay. Instead, we assume that any such GHG mitigation shortfall would be addressed in some other way. The real‐world effect of delays will vary, depending on the regional circumstances (Figure  4.6). The overall capacity of plants that are replaced “early” does not change for the United States and European Union; however, that replacement happens over a shorter period. For India (also Southeast Asia), the delay has little effect because of the scale of demand growth under the default GHG mitigation scenario. For China, the effect of delaying the transformation is to increase both the amount and rate of early retirement that is required, hence an increase to the amount and rate of construction required for replacement infrastructure. China will have the greatest challenge to delayed action. Aggregating the seven regional forecasts provides a measure of the overall effect (Figure 4.7). Depending on the length of delay before the transformation begins, the requirement for early replacement of coal‐ fired plant ranges from 8 to 18 GW yr−1 (Figure 4.8). While this may appear a relatively small addition to the overall BAU forecast of 288 GW yr−1, 18 GW yr−1 represents the construction of one to two additional power plants every month, for the entire transformation period, on top of what is already required.

­SENSITIVITY ANALYSIS Given the conceptual nature of this analysis, a number of the simplifying assumptions warrant closer scrutiny. Four such issues are considered here, three of

78  Advances in Energy Systems

10

0

0.4 0.3 0.2 0.1 2025

2035

2045

Coal-fired power plant (TW-installed)

0.7

10

0.6

5

0.5

0

0.4 0.3 0.2 0.1 0.0 2015

2025

2035

2045

0.4

2 1

0.3

0

0.2 0.1 0.0 2015

2055

2025

2035

2045

2055

0.5

4 3

Coal-fired power plant (TW-installed)

0.0 2015

4 3

Coal-fired power plant (TW-installed)

5

0.5

Retirement rate (GW/year)

Coal-fired power plant (TW-installed)

0.6

EU

0.5 Retirement rate (GW/year)

0.7

Early replacement

0.4

2 1

0.3

0

0.2 0.1 0.0 2015

2055

Retirement rate (GW/year)

Natural replacement

USA

2025

2035

2045

Retirement rate (GW/year)

Avoided growth

(a)

2055

(b)

6

0.3

5 4

0.2

3 2

0.1

1

1.4 Coal-fired power plant (TW-installed)

7

Retirement rate (GW/year)

Coal-fired power plant (TW-installed)

8 0.4

0.0 2015

2035

2045

50

1.2 1.0

40

0.8

30

0.6

20

0.4

10

0.2

0 2025

China

1.6

0.0 2015

2055

Retirement rate (GW/year)

India

0.5

0 2025

2035

2045

2055

6 5

0.20

4 3

0.10

2 1

0.00 2015

0 2025

2035

2045

2055

1.4 Coal-fired power plant (TW-installed)

7

0.30

Retirement rate (GW/year)

Coal-fired power plant (TW-installed)

8

1.2

50

1.0

40

0.8

30

0.6 20

0.4

10

0.2 0.0 2015

0 2025

2035

2045

2055

Figure 4.6  Early replacement forecasts for (a) United States and EU, and (b) India and China. The top row in each figure provides the default assessments. The bottom row, for each region, provides equivalent estimates, assuming the world makes no commitment to transformation until the year 2025. The vertical axes on the right are related to the bar graphs, which show the capacity retired in GW/year.

Retirement rate (GW/year)

1.6 0.40

Understanding Constraints to the Transformation Rate of Global Energy Infrastructure  79

those quantitatively – illustrating in each case that the early replacement rates are likely to be conservative on the low side.

(a) Coal-fired power plant (TW-installed)

3.0 2.5

What If Coal Power Stations Are Installed Faster than Forecast?

2.0 1.5 1.0 0.5 0.0 2015

2025

2035

2045

2055

2025

2035

2045

2055

(b) 3.0 Coal-fired power plant (TW-installed)

2.5 2.0 1.5 1.0 0.5 0.0 2015

Avoided growth Early replacement Natural replacement Figure 4.7  Early replacement forecasts for the world, ­obtained by aggregating the forecasts of each region: (a) provides the default assessments; and (b) provides equivalent estimates, assuming the world makes no commitment to transformation until the year 2025.

RER(GW/year)

40 30

If the current boom in coal power plant construction continues more rapidly than allowed for in the IEA forecasts, then the overall scale (and rate) of coal‐plant decommissioning in future years would necessarily be larger – if GHG mitigation targets are to be met. It is easy to envisage this being possible. The ETP 2DS assumes that coal capacity in India and Southeast Asia stabilize by the year 2020. Given the time lags involved in the infrastructure development cycle, this pathway might mean that no further coal‐fired plants are built beyond those already fully financed and under construction. All plants at other stages in the planning process would have to be canceled. It is hard to imagine that these regions, with their priority on increasing access to electricity, would accept such an interruption to their development plans while the world strives for international agreement on climate‐ change policy. Alternatively, the early replacement burden might increase if future demand for electricity has been understated. Again, it is easy to envisage this possibility if the population growth occurs faster than assumed in the IEA models  –  recent studies have highlighted that the future global population could peak much higher than the conventional wisdom[46]. The infrastructure retirement estimates would also be sensitive to an overestimation by the IEA of how much electricity demand can be suppressed over the longer term  –  either because their demand management targets are not achievable or because a greater‐than‐expected degree of transport fleet electrification occurs.

20

What If Coal‐Fired Power Capacity Must Be Further Reduced?

10

For the sake of exploring this quantitatively, we consider a possible example of why the reduction in coal‐ plant capacity might need to be greater than included in our default estimates. The IEA scenarios are based on ambitious levels of CCS, coupled to 530 GW of coal‐ fired generation plant by 2050. This ambitious target would require that CCS technology be successfully integrated with one large‐scale coal‐plant every month (on average), for the next 35 years. To put this in context, the CO2 storage rate in 2050 would be on the order of current global oil and gas (O&G) extraction rates.

0 2015

2020

2025

2030

Commencement year for the transformation Figure 4.8  Overall (averaged) rate of early replacement for the global aggregate, depending on the commencement year. The solid line represents the default assessment, and the dotted line represents all regions with natural ­retirement threshold of 70 years (this excludes United States and European Union).

80  Advances in Energy Systems

The results were tested using the modified assumption that only 25% of that CCS rollout is achieved, with the remaining 75% of associated coal‐fired plants being decommissioned so as to maintain a similar overall carbon budget. This approximately doubles the global aggregate estimates for early replacement rates, as shown in Figure 4.8.

retired earlier than their technical lifespan, with all the possible economic and investment penalties that such retirement could entail, will be in developing regions that might not have the capacity or desire to absorb such impacts.

How Old Is Too Old?

Many assume that stabilization of GHG emissions to limit the average global temperature to 2° above the Preindustrial Period is “safe.” There is, however, a growing consensus in the scientific community that 2° above the Preindustrial Period might be too high and that significant impacts will occur at atmospheric carbon dioxide levels associated with lower temperatures than previously thought[47]. If evidence mounts that the assumption of 2° as a safe upper limit is incorrect, then there might be a need for an even more rapid transition.

The historical evidence from developed economies shows it has been rare for coal plants to be decommissioned. Even if a plant is far older than its technical lifespan estimated as 50 years here, operators do not shut down operating plants. The calculations used here assuming a 50‐year plant lifespan will likely overestimate future rates of natural retirement and replacement, particularly in developing regions with strong upward pressure on demand for electricity. If that threshold is increased to 70 years, for all regions other than the United States and the European Union (which use scenario‐based natural replacement estimates), the global aggregate early replacement rates are more than doubled (Figure  4.8). Also noteworthy are the low ages of some plant being retired in the forecasts, particularly in China and Southeast Asia, which have plants being retired at ∼40 years of age using the default assumptions (Figure 4.9). This drops to ∼35 years under the two sensitivity tests that involve a greater overall retirement of coal plants. Crucially, this emphasizes that almost all plants

What If 2° Is Too Much?

­CONCLUSIONS Understanding achievable rates of energy system transitions is fundamental to developing meaningful energy and climate‐change policy. Limitations in engineering, labor, project management, material resources, and supply‐chain logistics have the potential to constrain the pace at which the economy is able to transition to a low‐carbon future. Focusing governmental efforts on reaching agreements and developing realistic pathways

Age of plant at the time of retirement (year)

100 90 80 70 60 50

European Union United States

40

Reminder OECD

30

Reminder Non

20 10 0 2015

2015 2020 2025 2055

S.E. Asia India China 2020

2025

2030

2035

2040

2045

2050

2055

Year of retirement Figure 4.9  Minimum age of coal‐fired power plant at the time of retirement obtained, from the default scenarios. The different shapes represent the different delay periods to the start date for the transformation. The different colors represent the different regions. China and Southeast Asia are the two regions with particularly young plants being retired − approximately 40 years old is the oldest plant.

Understanding Constraints to the Transformation Rate of Global Energy Infrastructure  81

to technology transitions in the United States, China, and India, rather than trying to obtain a global consensus, is both more relevant to achieving significant curbs on global emissions and more likely to be successful. Global dialog should continue, but consensus is not needed. The transition to a low‐carbon power generation fleet will likely increase the required rate of construction of new power plant infrastructure, over and above the already ambitious deployment rates necessary to meet projected demand growth (all else being equal). The evidence considered in this study is equivocal on whether we have the social, business, and engineering capability to sufficiently accommodate that increased infrastructure rollout challenge. Policy makers concerned about the need to implement rapid energy system transformation so as to avoid climate‐change impacts, should be equally concerned about the risk that the necessary transformation rates are actually achievable. Among other factors, the increased rate of infrastructure rollout will be driven by the need for early replacement of coal‐fired power plants and an increased reliance on intermittent renewables. Increasing deployment of intermittent power also has implications for transmission, distribution, and storage infrastructure, which have not been considered here. Estimates for infrastructure requirements would also increase dramatically if the modeling were to be extended to include other energy supply and utilization infrastructure for industrial processing, heating, and transportation to more fully capture the required transition away from a fossil‐fuel‐dependent world. Adopting low‐emissions thermal power generation technologies to meet electricity demand growth might have a relatively minor impact on these infrastructure delivery rates if we consider that there are a significant number of common unit operations and basic components. Not only does this mean that net construction rates might not be that different, it also suggests that there could be opportunities to lessen the transition burden through novel conversions of existing plants to use low‐carbon energy sources. Other opportunities to enhance the deployment rate might include an increasing use of modularization to harness the benefits of mass production lines. Recent history in the automotive sector suggests that mass‐­ production output can be ramped up more quickly, and to much greater levels, than traditional methods for large‐scale, field‐erected energy plants. This preliminary analysis shows that further scrutiny is required into the rate at which new infrastructure can be delivered to meet GHG mitigation targets. A more detailed bottom‐up analysis of rate‐limiting factors and

opportunities to overcome those, across all technologies and all regional/temporal contexts, is needed. Further work should also move beyond the focus on plant construction toward other aspects of climate‐change risk mitigation, e.g. the stabilization wedges of Pacala and Socolow[48]. In the interim, it is suggested that future modeling studies, including those by the IEA, should provide greater transparency in the assumptions and strategies for commissioning/decommissioning infrastructure, as these clearly have a major impact on the realistic rate for decarbonizing the economy.

­REFERENCES 1. World_Bank. Turn down the heat: why a 4°C warmer world must be avoided. 2012. Available at: http:// documents.worldbank.org/curated/en/2012/11/17097815/ turn‐down‐heat‐4%C2%B0c‐warmer‐world‐must‐ avoided. (Accessed December 2, 2014). 2. IPCC (2014). Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge; New York, NY: Cambridge University Press. 3. Meinshausen, M., Meinshausen, N., Hare, W. et  al. (2009). Greenhouse‐gas emission targets for limiting global warming to 2°C. Nature 458: 1158–1162. 4. International Energy Agency (IEA) (2014). Energy Technology Perspectives. Paris: OECD Publishing, IEA. 5. International Energy Agency (IEA) (2014). World Energy Outlook. Paris: OECD Publishing, IEA. 6. International Energy Agency (IEA) (2014). World Energy Investment Outlook. Paris: OECD Publishing, IEA. 7. Andersson, B.A. (2000). Materials availability for large‐ scale thin‐film photovoltaics. Prog. Photovoltaics 8: 61–76. 8. Feltrin, A. and Freundlich, A. (2008). Material considerations for terawatt level deployment of photovoltaics. Renew. Energy 33: 180–185. 9. UNFCCC. United Nations Framework Convention on Climate Change. Available at: http://unfccc.int/essential_ background/convention/background/items/1349.php. (Accessed January 2, 2015). 10. Tao, C., Jiang, J., and Tao, M. (2011). Natural resource limitations to terawatt solar cell deployment. ECS Trans. 33: 3–11. 11. Davidsson, S., Grandell, L., Wachtmeister, H. et  al. (2014). Growth curves and sustained commissioning modelling of renewable energy: investigating resource constraints for wind energy. Energy Policy 73: 767–776. 12. Erdmann, L. and Graedel, T.E. (2011). Criticality of non‐fuel minerals: a review of major approaches and analyses. Environ. Sci. Technol. 45: 7620–7630. 13. Kleijn, R. and van der Voet, E. (2010). Resource constraints in a hydrogen economy based on renewable energy sources: an exploration. Renewable Sustainable Energy Rev. 14: 2784–2795.

82  Advances in Energy Systems

14. IPCC (2007). IPCC Fourth Assessment Report: Climate Change 2007. Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, vol. 4. Cambridge; New York, NY: Cambridge University Press. 15. GlobalData. GlobalData Power: Power Plants. 2014. Available at: http://www.globaldata.com. (Accessed December 15–20, 2014). 16. Altshuler, A.A., Luberoff, D.E., and Lincoln Institute of Land Policy (2003). Mega‐Projects: The Changing Politics of Urban Public Investment. Washington, DC: Brookings Institution Press. 17. Reina, P. and Tulacz, G.J. (2005). The top 225 international contractors. Eng. News Rec. 255: 40–54. 18. Reina, P. and Tulacz, G.J. (2014). It’s a competitive world after all. Eng. News Rec. 273: 1–1. 19. International Energy Agency (IEA) (1995). World Energy Outlook. Paris: OECD Publishing, IEA. 20. Pennell, N., Ahmed, S., and Henningsson, S. (2010). SUSTAINABILITY reinventing the city to combat climate change. Strategy + Business 34. 21. Inbound Logistics (2013). The Natural Gas Factor: New Market for Providers, New Competition for Capacity. New York, USA: Inbound Logistics. 22. Kemp J. COLUMN‐Oil Industry Starts to Squeeze Costs, Wages: Kemp. Reuters; 2014. Available at: http:// www.reuters.com/article/2014/01/30/oil‐and‐gas‐ idUSL5N0L41LJ20140130. (Accessed December 20, 2015). 23. Flyvbjerg, B., Garbuio, M., and Lavallo, D. (2009). Delusion and deception in large infrastructure projects: two models for explaining and preventing executive disaster. Calif. Manage. Rev. 51: 170–193. 24. Merrow, E.W., McDonnell, L.M., and Arguden, R.Y. (1988). Understanding the Outcomes of Mega‐Projects. Santa Monica, CA: RAND Corporation. 25. Kidela_Capital_Group. Rare earth element shortages threaten global wind power development. AltEnergyStocks.com, 2010. Available at: http://www. altenergystocks.com/archives/2010/12/rare_earth_ element_shortages_threaten_global_wind_power_ developm ent.html. (Accessed December 31, 2014). 26. Merrow, E.W. (1978). Constraints on the Commercialization of Oil Shale. Santa Monica, CA: RAND Corporation. 27. Merrow, E.W., Phillips, K.E., and Myers, C.W. (1981). Understanding Cost Growth and Performance Shortfalls in Pioneer Process Plants. Santa Monica, CA: RAND Corporation. 28. Zimmerman, M.B. (1982). Learning effects and the commercialization of new energy technologies: the case of nuclear power. Bell. J. Econ. 13: 297–310. 29. Arnsdorf I, Murtaugh D, Kaskey J. Labor shortage threatens to bust the shale boom. Bloomberg, 2014. ­ Available at: http://www.bloomberg.com/news/2014‐ 04‐17/midnight‐welding‐picks‐up‐slack‐that‐imperils‐ shale‐boom.html. (Accessed December 31, 2014). 30. Loyalka, P., Carnoy, M., Froumin, I. et  al. (2014). Factors affecting the quality of engineering education in

the four largest emerging economies. High. Educ. 68: 977–1004. 31. UNESCO (2010). Engineering: Issues Challenges and Opportunities for Development. Paris: UNESCO Publishing. 32. OECD. Graduates by field of education. 2014. Available at: http://stats.oecd.org/Index.aspx?DataSetCode= RGRADSTY. (Accessed January 2, 2015). 33. Gereffi, G., Wadhwa, V., Rissing, B. et  al. (2008). Getting the numbers right: international engineering education in the United States, China, and India. J. Eng. Educ. 97: 13–25. 34. Loyalka P, Carnoy M, Froumin I, et al.Getting the quality right: engineering education in the BRIC countries. 2012. Available at: http://cepa.stanford.edu/content/ getting‐quality‐right‐engineering‐education‐bric‐ countries. (Accessed December 31, 2014). 35. Carelli, M.D., Garrone, P., Locatelli, G. et  al. (2010). Economic features of integral, modular, small‐to‐ medium size reactors. Prog. Nucl. Energy 52: 403–414. 36. Kanellos M. GE Confirms That Wind Turbine Supply Is Getting Worse. CNET; 2008. Available at: http://www. cnet.com/au/news/ge‐confirms‐that‐wind‐turbine‐ supply‐is‐getting‐worse. (Accessed January 15, 2015). 37. Eni S.p.A. Green refinery. Available at: http://www.eni. com/en_IT/innovation‐technology/technological‐focus/ green‐refinery/green‐refinery.shtml. (Accessed December 20, 2014). 38. Eni S.p.A. Venice refinery—porto marghera. Available  at: http://www.eni.com/en_IT/company/ operations‐strategies/refining‐marketing/refining‐rf/ refinery‐marghera.shtml or http://www.eni.com/en_IT/ attachments/azienda/attivita‐strategie/refining‐marketing/ eni_Venezia%20ENG_esecutivo.pdf. (Accessed December 20, 2014). 39. United Hub. United teams up with AltAir on aviation biofuels. Available at: https://hub.united.com/en‐us/ news/company‐operations/pages/united‐and‐altair‐partner‐ on‐biofuels.aspx. (Accessed December 31, 2014). 40. Messenger B (2013). 30m Gallon Facility to Produce Aviation Biofuels from Waste for United Airlines. Tulsa, OK: Waste Management World. 41. Solecki M. AltAir fuels to produce renewable jet fuel for United. Fueling Growth, 2014. Available at: http://www. fuelinggrowth.org/altair‐fuels‐to‐produce‐renewable‐ jet‐fuel‐for‐united. (Accessed December 31, 2014). 42. Smith, R. (2005). Chemical Process Design and Integration. Chichester, Hoboken, NJ: Wiley. 43. Mac Donald M. Definition of CCS ready. Available at: http://www.globalccsinstitute.com/insights/authors/ christophershort/2010/11/03/definition‐ccs‐ready. (Accessed December 15, 2014). 44. Australian Energy Market Operator, 100 per cent renewable study—modelling outcomes. 2013. 45. Pearce F. Lima talks set up climate deal for a ‘bad outcome’. New Scientist, 2014. Available at: http://www. newscientist.com/article/dn26700‐lima‐talks‐set‐up‐ climate‐deal‐for‐a‐bad‐outcome.html#.VKYqpE2_ yUk. (Accessed January 2, 2015).

Understanding Constraints to the Transformation Rate of Global Energy Infrastructure  83

46. Gerland, P., Raftery, A.E., Ševčíková, H. et al. (2014). World population stabilization unlikely this century. Science 346: 234–237. 47. Allen MR, Barros VR, Broome J, et  al.IPCC fifth assessment synthesis report‐climate change 2014: ­ synthesis report. IPCC Report, 2014. Available at: ­ http://www.ipcc.ch/report/ar5/wg3. (Accessed January 2, 2015).

48. Pacala, S. and Socolow, R. (2004). Stabilization wedges: solving the climate problem for the next 50 years with current technologies. Science 305: 968–972. 49. US‐DOT. World motor vehicle production, selected countries. 2014. Available at: http://www.rita.dot.gov/ bts/sites/rita.dot.gov.bts/files/publications/national_ transportation_statistics/html/table_01_23.html_mfd. (Accessed January 2, 2015).

5

Physical and Cybersecurity in a Smart Grid Environment Jing Xie1, Alexandru Stefanov1 and Chen‐Ching Liu1,2  School of Mechanical and Materials Engineering, University College Dublin, Dublin, Ireland  School of Electrical Engineering and Computer Science, Washington State University, Pullman, WA, USA 1

2

As a critical infrastructure, the electric power grid is vulnerable with respect to malicious physical and cyberattacks. Indeed, the importance of physical and cybersecurity for smart grids has been highlighted by recent intrusion incidents worldwide. Various types of sensors, instrumentation, information and communications technology (ICT), computational methodologies, and security controls have been developed to enhance the physical and cybersecurity of the smart grid. In this chapter, critical vulnerabilities of a smart grid that can be exploited for physical and cyber intrusions are discussed. A comprehensive survey is conducted on the state‐of‐the‐art research to enhance the physical and cybersecurity in a smart grid environment. Furthermore, the interdependency of physical and cybersecurity is illustrated with an intrusion scenario. The emphasis is on the physical and cybersecurity of power substations. Computer simulation results for a case study are presented.

­INTRODUCTION Critical infrastructures are complex systems that form the foundation of a modern society, e.g. large‐scale networks for water, transportation, communications, and energy. Their reliability and secure operation are essential for the economy and national security. A disruption of service may lead to damages or even a loss of lives. Critical infrastructures have become more and more interconnected in different ways, e.g. physical and cyber connectivity[1]. As a fundamental infrastructure of the society, the electric power grid is a critical infrastructure upon which other infrastructures depend. The safe operation of power systems is critical for energy security[2]. In a smart grid environment, power grids are increasingly dependent on information and communications technology (ICT) for the operation and control of physical facilities. It can be envisioned that on top of the physical power infrastructure reside ICT layers that are coupled with the power grids[3].

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

86  Advances in Energy Systems

Generation, transmission, and distribution are three main subsystems of the electric power infrastructure. The transmission subsystem consists of widely distributed substations and high‐voltage pylons. Compared with power plants and control centers, the transmission subsystem is easier to access and is considered to be the most vulnerable. Threats to power facilities arise from external and internal sources, e.g. power system component failures, human errors, natural calamities, and sabotage[4, 5]. Among these threats, physical and cyber intrusions deserve special attention due to the potential hostile nature. In the power grid, a large number of power substations are unmanned and located in isolated, remote, rural, or mountainous areas. They are vulnerable to malicious physical attacks as the security personnel cannot reach these facilities easily. Physical intrusion attempts and events are not uncommon. In a recent incident, the attacker caused a shutdown of a power substation by firing at the transformers[6]. Thus, an advanced system to monitor physical security of substations and other facilities is needed in a future smart grid environment. Clearly, the physical intrusion is not the only way to breach the substation security and cause damages to the power infrastructure. ICTs on the power grids have evolved from isolated structures into an open and networked environment. The power and cyber systems are becoming more and more interdependent[7]. Power system communication and its cybersecurity are essential for a smart grid infrastructure. The private communication networks for power systems can be vulnerable with respect to ­cyberattacks[8]. This chapter discusses critical issues related to physical and cybersecurity and their interdependency in a smart grid environment, particularly the security of power substations. Major intrusion incidents are presented to highlight the importance of physical ­security and cybersecurity for smart grids and the critical vulnerabilities that can be exploited. The current security technologies that protect power systems against physical and cyber intrusions are identified. A state‐of‐the‐art survey is conducted on research efforts to enhance the smart grid security. In addition, the interdependency of physical and ­cybersecurity is illustrated. An intrusion scenario is presented with computer simulation results as a case study. The physical security of a substation can be breached to allow cyberattackers to access the substation communication network. Attackers in remote locations gain unauthorized access to the private ICT infrastructure supporting the supervisory control and data acquisition (SCADA) system for the power grid. Cyber intruders have unauthorized control capabilities

over the substation assets, e.g. remote terminal units (RTUs), with a possible devastating impact on power system operation.

­MAJOR INTRUSION INCIDENTS Physical Intrusion Incidents Worldwide physical intrusion incidents serve as reminders of the importance of physical security for power systems. Incidents are categorized into three areas: power plants, transmission towers, and power substations. As the sources of a power grid, power plants are critically important. Unlike substations and transmission towers, power plants are usually manned. Thus, the difficulty of sabotage is high. However, if attackers are able to successfully sabotage a power plant, the damage can be very significant. Damages to power plants can be caused by vandals, protesters/activists, and terrorists. In June 2014, a makeshift bomb was placed next to a diesel tank at the UniSource Energy Services Valencia Plant. The explosion caused a small spill. Fortunately, the fuel did not ignite. Had there been a catastrophic explosion, the entire city of ­Nogales could have been devastated. Besides conventional power plants, there are more and more wind and solar farms as the level of penetration of variable renewable generation increases[9]. Therefore, the physical security of wind turbines and solar panels becomes a new concern. In January 2013, two wind turbines were brought down at East Ash Farm in the United Kingdom. Bolts were missing from the base, and the second tower that collapsed was not far away from the first one. Thus, sabotage was suspected. Among different types of power plants, nuclear facilities are the most attractive to terrorist attacks[10]. Security is the focus of debate over nuclear power[11]. In August 2014, an act of sabotage was reported at the Doel 4 nuclear power plant in Belgium. Someone tampered with the system for emptying lubricating oil from the main turbine. The turbine became overheated, leading to a shutdown of the power plant for four months. Power grids can stretch hundreds or thousands of miles and be geographically dispersed over a wide area. Damage to transmission towers can be caused by thefts, malicious sabotage, and terrorist attacks. Tower bolts and foundation steel provide essential support for tower structures. If the support legs are broken, they will be brought down. In October 2013, two power poles at Lonoke County, Arkansas, USA, were intentionally cut. A blackout occurred, affecting more than 10 000 people over several days[12].

Physical and Cybersecurity in a Smart Grid Environment  87

Although key substations are surrounded by solid fences, acts of sabotage, criminal activities, and malicious behaviors are observed regularly. As a common scenario, theft is the most difficult incident to prevent[13]. Copper theft is indicated as a dangerous and costly threat to the reliability of power systems by the Canadian Electricity Association[14]. In August 2014, intruders cut fencing at the Metcalf substation in San Jose, California, USA, of Pacific Gas and Electric Company (PG&E) and stole construction equipment for security upgrades[15]. Vandals damage substation facilities for various reasons. The recent incident (April 16, 2013) at PG&E’s Metcalf substation draws attention to the physical security of substations. Attackers cut the fiber‐optic telecommunications cables of American Telephone and Telegraph Company (AT&T) and shot transformers for 19 minutes[6]. In total, 17 transformers were knocked out. It took 27 days to recover the substation. Disgruntled employees may pose a threat, as they may have the key to the substation and be familiar with facilities and vulnerabilities. For instance, a disgruntled employee knocked out a substation in Salt Lake City, USA, during the 2002 Winter Olympics (February 2002), causing an area power outage. Finally, a group of attackers can attack multiple substations simultaneously with a carefully coordinated plan targeting the bulk power system. In 1981, two substations in Florida, USA, were heavily damaged by simultaneous dynamite explosions[16]. Cyber Intrusion Incidents Major intrusion incidents also occurred due to vulnerabilities of cybersecurity. A widely publicized malware attack is the Stuxnet worm. It was discovered in June 2010[17]. Stuxnet targets the SCADA system of industrial facilities. The objective is to search for a specific type of programmable logic controller and reprogram parts of its code. It waits for certain conditions and then takes control. Cyber intrusion incidents were reported by the industrial control systems cyber emergency response team (ICS‐CERT)[18]. In October 2012, a power plant was affected by a virus infection. The virus infected 10 computers in the turbine control system network. A third‐party technician used an infected USB flash drive for software updates during a scheduled outage for equipment upgrades. The technician was not aware that the USB flash drive was infected with a variant of the Mariposa virus. As a result, the power plant’s restart was delayed for three weeks. A malware infection was also reported at a different power plant. Common and sophisticated malware was discovered

in the industrial control system environment. The infections originated from an employee that routinely used an USB flash drive to back up the control system configurations. The malware affected two engineering workstations critical to the operation of the power plant. Their operations were significantly impaired as the workstations had no backup. In February 2013, cyberattacks were reported against control systems of critical infrastructures connected to the Internet. Attackers targeted a gas compressor station. Brute‐force attempts were made to access the process control network. An alert was issued to warn other critical infrastructure asset owners, including the natural gas industry, to guard against similar activities. The alert discovered related attempts to compromise the control networks of other gas compressor stations across the Midwest and Plains states in the United States. None of the brute force cyberattacks were successful. A report on the risk of Internet‐accessible control systems was issued for the period of January– April 2014. Two new cyber incidents were reported. Intruders exploited the weak network configurations and lack of perimeter security. In the first incident, the target was a public utility. The attackers breached the security controls and accessed the utility’s control system network. The remote access capability was protected by a simplistic authentication system. The standard brute‐force techniques were used to crack the password. In the second incident, attackers targeted the control system operating a mechanical device. Intruders accessed the control system server, connected to the Internet via a cellular modem, through a SCADA protocol. It was not protected by firewalls or authentication access controls. Fortunately, the control system was physically disconnected from the mechanical device for scheduled maintenance. A summary on the cybersecurity incident response for 2014 is presented in the ICS‐CERT report for the period September 2014–February 2015. The industrial control system’s cyber emergency response team responded to 245 cybersecurity incidents that were reported by industrial asset owners. The energy sector reported the most security incidents compared to other industrial sectors in 2014. More than half of the incidents involved persistent threats or sophisticated actors. Some of the actors were hacktivists, insider threats, and criminals. However, many threat actors were unknown. The overall objective of the reported incidents was to gain access to both business and control systems infrastructure. Attackers attempted to have unauthorized access and control over the SCADA devices. They tried to

88  Advances in Energy Systems

exploit zero‐day vulnerability −a loophole in the control system devices and software that is unknown to the vendor − before the vendor became aware and remediated it. Structured Query Language (SQL) injection attacks were performed by exploiting the web application vulnerabilities. SQL injection is a code injection technique where malicious SQL statements are inserted into an entry field to unveil database contents to the attacker. Targeted spear‐phishing campaigns and lateral movement between network zones were performed. Spear‐phishing is an e‐mail‐ spoofing fraud attempt that targets a specific organization for unauthorized access to confidential data. After the initial successful breach into the corporate network, lateral movement techniques allow the ­attacker’s movement in the internal network from one system to another until the goal is achieved. The precise intrusion methods were not identified from the forensic examinations because of a lack of detection and monitoring capabilities within the compromised networks. The reported intrusion incidents demonstrate the importance of cybersecurity in a smart grid environment.

­SMART GRID VULNERABILITIES The smart grid environment is vulnerable to malicious attacks. Growing threats to infrastructure security are considered in smart grid policies[19]. Security challenges in smart grid implementation are articulated in Ref.[20]. It covers network threats, cyber‐physical threats, and smart metering threats as well as privacy issues in the smart grid. Along with the threats, the book discusses the means to the improvement of smart grid security and the standards emerging in the field. The legal issues in smart grid implementation are discussed, particularly from a privacy point of view. Although physical and cyberattacks on power infrastructure are not new threats, they are continuously evolving to be more dangerous and more costly. As worldwide electricity sectors enter this transformative period toward a future smart grid environment, effective approaches to protecting power grids should be developed to meet the growing threats. With respect to the interconnections between the electric power grid and other critical infrastructures, there are three types of threats: attacks on, by, and through power systems[21]. In the first case, the electricity infrastructure itself is the primary target. In the second case, the ultimate target is the population. Attackers can use parts of the power infrastructure as a weapon, e.g. nuclear power plants. In the third case, the target is the civil infrastructure. For example, an

electromagnetic pulse could be coupled through the grid by terrorists to damage computers or the telecommunications infrastructure. Physical Vulnerabilities As industry restructuring introduces several new economic objectives into operation, power systems are being stressed closer to the stability limit. Damages to the physically distributed power facilities that form the fundamental layer of power grids have a direct impact on the smart grid operation. It may cause the system to lose balance and even result in catastrophic failures. In addition, the observability and controllability of power grids may be compromised due to disruptions of communication and measurement facilities[22], e.g. smart meters and Phasor Measurement Units (PMUs). Furthermore, intruders can steal or sabotage the protection devices, leading to the invalidation or malfunction of special protection schemes (SPSs)[23]. For example, if load‐shedding relays at substations are damaged, the load‐shedding scheme is compromised and may not be able to mitigate the disturbance on the power grid during contingencies. In the power grid, power substations are vulnerable to physical intrusions, as a large number of them are unmanned and located in isolated or rural areas. It is indicated that the sabotage of nine key substations is sufficient for a large blackout in the United States[24]. In fact, physical intrusion attempts and events are observed regularly at substations. After an alarm is triggered, it typically takes the police a minimum of 20 minutes to reach the substation[25]. Excluding the time spent on observing the scene and assessing the threat, the system operator may have approximately 10 minutes to perform the preventive and/or remedial actions before the police arrive. In the current industry practice, however, system operators simply inform the police if an intrusion event is detected at a substation. Without the support of an advanced monitoring system, the system operator located in the remote control center cannot investigate the intention of the intruder promptly. Some security systems on sites only record data for postmortem investigations. Consequently, the severity of consequent damage cannot be estimated. While thefts lead to significant economic losses, there is generally no immediate damage to the power system. On the other hand, an undetected intrusion inside substation buildings is dangerous. There are multiple options for an intruder in the substation control room: (i) opening circuit breakers via the mimic switching board; (ii) connecting a PC to the substation local area network (LAN) and, consequently, accessing the entire SCADA system; and (iii) stealing or destroying

Physical and Cybersecurity in a Smart Grid Environment  89

valuable apparatus. Cascading events may be triggered by the major disturbance due to the loss of important facilities. Therefore, preventive and remedial actions may be necessary to avoid a system collapse. Physical intrusions and coordinated attacks on multiple substations can cause considerable damage to the power grid. Usually, if one substation is damaged, power can be rerouted through other substations. However, the economic dispatch may not be achievable. In addition, power flows could be disrupted if a number of critical substations are disabled simultaneously due to sabotage[26]. The cost of repairing damaged facilities is also appreciable. While not all substations require special protection systems, the critical ones should be identified and protected. Therefore, risk assessment methods are needed to assist utilities in optimizing the investment in physical security enhancement under budgetary constraints. However, the actual use of risk assessment by security practitioners had not been researched in a scientific manner. In the risk assessment survey[27], it is indicated that about one‐third of respondents fail to conduct a cost−benefit analysis while evaluating security options to mitigate risks. In addition, less than half of respondents measure the effectiveness of security systems after installation. Thus, smart grid security should be enhanced with respect to the physical vulnerabilities. Cyber Vulnerabilities The physical power grid and communication infrastructure supporting the smart grid together form a large, complex, and interdependent cyber‐physical system (CPS)[28]. The general CPS model for power grids is illustrated in Figure  5.1. Disruption of services at the cyber layer can make a major impact on the smart grid operation and assets. Denial of service (DoS) and increased latency for measurements and control commands at the cyber system layer can affect the power grid’s observability and controllability[29]. Following a system disturbance, during the initial transients, the control mechanisms rely on local information to take local control actions. However, at later stages, a control action may be delivered from a remote controller, and communication delays in CPSs are crucial for the performance of the control systems, e.g. power system stabilizers[30]. In case of large disturbances, the grid can experience instability, and without proper control, it may lead to a sequence of cascading events and cause a blackout. Intrusions and targeted attacks against the cyber system are major concerns for smart grid security. As the ICT connectivity increases, so does the potential for cyber intrusions. The TCP/IP and

Cyber system Measurements

Controls

G

G G Power system

Figure 5.1  Cyber‐physical system.

ethernet technologies are deployed in modern ICT networks for high‐speed data exchange at reduced costs. However, they are susceptible to IP‐based attacks[28, 31]. Firewalls and electronic security perimeters do not guarantee cybersecurity. Misconfiguration of these widely adopted access control methods leads to a common vulnerability. Even with proper configuration, the vulnerabilities are not completely removed. For example, firewalls cannot detect insider attacks and connections from trusted sources[32]. Site engineers and vendor personnel have remote access to the substation and power plant LANs for maintenance purposes. RTUs and intelligent electronic devices (IEDs) are accessed from remote locations external to the grid. If they are not properly secured, such vulnerabilities can be exploited by attackers as unauthorized access points into the private communication networks of power grids. The protective relays are critical devices for system protection. Conventional relays can be accessed only by serial cable connections locally. However, the smart relays have a network interface and are accessed and configured remotely[33]. A malicious configuration change of protective relay settings can have a severe impact on the security of power grid operation in case of large disturbances, e.g. faults. Smart meters are vulnerable to cyberattacks, resulting in financial losses, e.g. altering of meter readings and sending false pricing information to customers[34]. The security of communication protocols provides assurance within the cyber layer only. However, the smart grid is a CPS. The interactions between the cyber and physical layers are often overlooked, resulting in common security vulnerabilities.

90  Advances in Energy Systems

This has serious security implications to smart grid applications, e.g. demand response, vehicle‐to‐grid[35]. The possible access points for cyber intrusions into the substation communication network were identified[36]. An example of a cyberattack that originates from inside the substation communication network is provided. The attacker can use an infected USB flash drive to install malware on the substation user interface. This can execute hacking tools or open predefined communication ports to provide unauthorized access into the communication network. Cyberattacks can originate from outside the substation communication network as well. The remote access points for maintenance and operation can be used to bypass the substation firewall and gain access to the communication network. A compromise of the substation user interfaces can lead to undesirable switching and modification of settings for the power devices. Network‐based intrusions can result in packet monitoring, modification, and replay. For example, altering of simple network time protocol (SNTP) messages can disrupt time synchronization, and GOOSE (Generic Object‐Oriented Substation Event) messages can open circuit breakers. ICT devices can be infected with viruses, worms, and Trojan horses that disrupt their normal operation. DoS attacks affect data processing and communication performance through resource exhaustion. ICT buffers can overflow, and communication bandwidth can be consumed by creating traffic avalanches over short periods of time, e.g. packet flooding attack. Cyberattacks affect the operation of the physical power grid. They may lead to instability, loss of load, equipment damage, and cause a partial or complete blackout[37].

­SECURITY CONTROLS FOR THE SMART GRID Various types of sensors, instrumentation, ICTs, computational algorithms, and security controls have been deployed to protect power grids against physical and cybersecurity threats. The current security technologies should be adapted to protect against the continuously evolving security threats. The security of power grids must be enhanced at all levels, i.e. generation, transmission, and distribution, to secure a smart grid environment. Physical Security Controls Physical Security Monitoring Systems A common definition of the Physical Security ­ onitoring Systems (PSMS) is reported in Ref.[38]. M

A  PSMS integrates people, procedures, and equipment for the protection of assets or facilities against theft, sabotage, or other malevolent human attacks. Various types of sensors can be applied to detect an entry into a substation, e.g. video, motion, sound, and seismic detection. Currently, physical security of substations is focused almost exclusively on fence line detection and security of substation control rooms using video cameras[39, 40]. From the conventional analog‐based closed circuit television (CCTV) system to the modern IP‐based system, the resolution of surveillance cameras has increased significantly to megapixels at a lower cost. A typical architecture of the surveillance system is shown in Figure 5.2. A comparison among commercial security monitoring systems is summarized in Table 5.1. A weakness of the fence line detection is that attackers cannot be detected before they break into the substation. In order to achieve the outside‐the‐ fence detection, there is a new trend in the application of radar sensors[51–53]. In the past, most users of radar sensors and related intrusion detection techniques were military and government agencies. It is reported that utilities are interested in radar sensors today[54]. Radar sensors transmit pulses of radio waves or microwaves that bounce off any object in their path and evaluate the reflected signal. The integration of radar sensors can expand the monitoring area significantly and reduce the number of video cameras. Thus, it becomes possible to respond to potential intrusions while intruders approach the substation fence. A weakness of visual and radar sensors is that weather, terrain, and foliage can influence the measurement accuracy. This gap can be bridged by adding buried‐line sensors to monitor the perimeter. The buried‐line sensor comprises detection probes that are cables buried in the ground, typically between fences that form the secure zone. In essence, buried‐line cables are coaxial cables in which one cable is used to transmit generated radiofrequency from a secured zone. The other cable is designated as a radiofrequency receiver. Under normal conditions, the receiving cable detects steady radiofrequency signals. If an approaching intruder tries to break into a secured zone, the characteristics of received signals are altered, and an alarm is triggered. Various types of motion detectors are widely used for indoor monitoring, e.g. passive infrared (PIR), ultrasonic, microwave, radiowave detectors. As a thermal sensor, PIR detectors passively measure incoming infrared radiation in its field‐of‐view (FOV). A human being radiates infrared energy with a wavelength between 9 and 10 μm. Thus, PIR is typically sensitive in the range of 8–12 μm. The other three types have

Physical and Cybersecurity in a Smart Grid Environment  91

Database Video cameras Storage of video records Visualization tool

Video streams

Humanmachine interface

On-site servers for intelligent surveillance Marked video stream

Detection module Locating module

2-D grid map data

Data fusion Tracking module Entire data grid

Patrolling module

Figure 5.2  Typical architecture of the surveillance system. Table 5.1  Comparison of the main properties of commercial and research PSMSs. Event Visualization of Site Crowded reconstruction multiple cameras map environment

System Commercial TNT systems IBM (S3‐R1)[42] Siemens SiteIQ[43] Honeywell Pro‐Watch[44] Forward Looking Infrared (FLIR) Thermal Fence[45] Research Video Surveillance and systems Monitoring (VSAM)[46] SPSM[47] Dynamic Object Tracking System (DOTS)[48] ADVISOR[49] Real‐time Baseline Video Surveillance System (RBVSS)[50] [41]

Online

Tracking

✗ ✓ ✓ ✗ ✗

✓ ✗ ✓ ✓ ✓

✗ ✓ ✗ ✓ ✓

✓ ✗ ✗ ✓ ✓

✓ ✗ ✓ ✓ ✓

✓ ✗ ✗ ✗ ✓

✗

✓

✗

✗

✓

✓

✓ ✗

✓ ✓

✓ ✓

✗ ✗

✓ ✓

✓ ✗

✗ ✗

✓ ✓

✗ ✗

✗ ✓

✓ ✓

✓ ✓

SPSM, substation physical security monitoring; TNT, tag and track.

the same mechanism but send waves of different frequencies. As transceivers, these types of detectors actively send out waves and sense disturbances through the received waves that travel through an area. It is useful to fill the visual gaps caused by obstacles. Advanced motion detectors that use dual technologies are more accurate and energy‐efficient. For example, after PIR detects proximity, the microwave will be turned on for certification.

Access Control Systems Power substations are visited by a few authorized staff members regularly. Therefore, an access control system is needed to distinguish between authorized and unauthorized access. As a subsystem of a PSMS, an access control system relies on a central database and verification technologies[55]. First, the database obtains the personal security information through the interaction between personnel and access control

92  Advances in Energy Systems

devices. By checking the level of authority, the access control system commands the access control devices to provide a proper response to the personnel. Various types of verification technologies are summarized in Table 5.2. Through access control systems, system operators in the control center can remotely monitor and control all access attempts to predefined areas. Therefore, the safety level of property, assets, and employees is improved. In addition, the movements of all authorized staff and visitors on the premises can be monitored in real time without triggering an alarm. Risk Assessment, Modeling, and Simulation Tools A modeling and simulation tool, Automated Vulnerability Evaluation for Risks of Terrorism (AVERT) [56, 57], has been developed by ARES security. It is ­designed to conduct security risk assessments on critical infrastructures in a four‐step process: (i) characterizing the facility and its security system; (ii) synthesizing the security system response to an intrusion event; (iii) analyzing results and

conducting the cost‐benefit analysis; and (iv) optimizing the budget for a system installation or update. For generating metrics on the system performance, the probability of detection, interruption, and neutralization of threat is evaluated against time and distance independently. As an interactive training simulator, BluTrain[58] is designed specifically for security forces. It helps with instruction on tactics, techniques, and procedures by providing an interactive three‐dimensional (3D) representation of the facility. Virtual training improves training effectiveness and reduces training costs. Security forces can familiarize themselves with possible attack scenarios and site layouts. Thus, the response time against various threat scenarios can be improved. Cybersecurity Controls The cybersecurity controls that protect the ICT networks supporting the smart grid are shown in Figure 5.3.

Table 5.2  Verification technologies for access control. Verification technologies Card reader systems

Characteristics Magnetic strip cards Wiegand type cards Bar code cards Smart cards (contact type) Smart cards (contactless type)

Key fobs Biometric devices Pin access keypad

Firewalls

Economical, practical but brittle Difficult to forge or counterfeit Cheap, but easy to forge or counterfeit Moderate throughput rates High throughput rates Portable High security level Convenient

Intrusion detection systems

Cyber system (ICT networks) Intrusion-tolerant replication

Encryption

Authentication

Figure 5.3  Cybersecurity controls for the smart grid.

Physical and Cybersecurity in a Smart Grid Environment  93

Firewalls Network firewalls are ICTs that hold the first line of defense against malicious cyberattacks, e.g. unauthorized access and control of smart grid assets. They protect the entry points of private communication networks. The incoming and outgoing communication traffic is inspected based on sets of rules. A rule‐based engine inspects each packet rule by rule to find a match. The legitimate data packets are allowed to pass while malicious packets are blocked and discarded[59]. The rule criteria for packet acceptance or rejection may include the IP address range, specific port services, and type of protocols. The packet audits are recorded for offline analyses of malicious behaviors. Firewall analyzers are used to automatically detect anomalies in the datasets. The firewall logs are used to assess the number of records rejected compared to the total number of traffic records. Thus, the probability of cyberattack occurrence is computed[32]. Firewalls are deployed to guard the smart grid communication networks against cyber intrusions. They enable electronic security perimeters that allow only the legitimate data traffic to access the gateways for smart grid communications[60].

Intrusion Detection Systems Anomaly detection systems are early warning mechanisms of cyber intrusions. Anomaly detection uses event correlation methods to identify unauthorized activities. A detailed analysis of data logs is required. Patterns that are not of expected behaviors are identified and correlated[61]. Anomaly detection systems are deployed to protect the communication networks of power substations. They rely on the fact that intruders’ behaviors generate logs across the substation network. The logs of intruders’ footprints are used to detect anomalous data items. Integrated host‐ and network‐based anomaly detection systems proved to be useful in detecting cyberattacks on multiple substations. Intrusion detection systems are effective security controls to discover intrusions and mitigate cyberattacks in a timely manner. They monitor user access, file access, and system event logs to  identify anomalies in the private communication networks. The network‐based intrusion detection systems collect packets from a communication network and analyze the traffic activity. Host‐based intrusion detection systems monitor the host system and raise alarms if malicious behaviors are identified[36]. Apart from anomaly‐based intrusion detection, there are two other broad categories of intrusion detection techniques, i.e. signature‐based and stateful protocol

analysis. Signatures are patterns that correspond to known threats. The intrusion detection technique based on signatures compares the available signatures with observed events to detect malicious activities. Predetermined indisputable activities of each protocol state are compared with the observed events to detect mismatches in the stateful protocol analysis[62]. Encryption The smart grid relies on an extensive data communication infrastructure. Different communication protocols are used for data exchange at all levels. In a smart grid environment, the transmission system operators, distribution system operators, and electricity suppliers monitor the energy generation and consumption in real time. End users know the current energy price. They can adapt their energy consumption to the current supply situation. The international standard communication protocols used in a smart grid environment, e.g. Distributed Network Protocol (DNP) 3.0 over TCP/IP, IEC 60870–5, IEC 60870–6, IEC 61850, are secured with different cryptographic schemes. Otherwise, data contents can be modified in the case of man‐in‐the‐middle attacks with important security implications[63]. Standard encryption algorithms are widely used to protect the integrity and confidentiality of smart grid information, e.g. data encryption standard (DES), triple DES, and advanced encryption standard (AES). Symmetric cryptography is used to efficiently handle large amounts of data, e.g. ZigBee for wireless communication uses 128‐bit AES. The encryption keys must be known by both the sender and receiver and must remain secret. The encryption key must be transmitted from the sender to receiver separately from the encrypted message. Thus, symmetric encryption has a short lifespan. The key can be discovered, or brute force can be used by attackers to decrypt the message. The alternative is asymmetric encryption. The sender uses the receiver’s public key to encrypt the message. The receiver uses both its public and private keys to decrypt. The communication security is enhanced, as only the receiver knows the private key. However, asymmetric encryption is more expensive in terms of costs and computational effort[64]. Authentication Authentication mechanisms are widely adopted security controls that confirm the identity of an ­ entity that tries to access the system resources. Human user and machine‐to‐machine authorizations are necessary to verify the permissions and rights

94  Advances in Energy Systems

to access the requested resources. Authentication and authorization mechanisms protect the privacy of data against unauthorized control of smart grid assets[35]. Logon authentication is the most common security control. This includes passwords, smart cards, and associated personal identification numbers (PINs), fingerprints, voice pattern samples, and retinal scans. Network access authentication is used to verify the identity of the entity when an attempt is made to access a network service. Message signatures and encryption by using IP Security (IPSec) technology ensure data confidentiality, integrity, and authenticity. IPSec transmission often uses public key certificates from a trusted certificate authority. Digital certificates are used for authentication and secure communications of data on unsecured networks. They relate a public key to an entity that has the corresponding private key[65].

as for security enhancement of the network electronic security perimeter. This technology allows only legitimate messages to pass even when some of the firewall replicas are compromised[67].

Intrusion‐Tolerant Replication

Physical Security Monitoring Systems

The intrusion‐tolerant replication method was used to design a SCADA system that can operate correctly and with minimal performance degradation when malicious attackers compromise parts of the system. The intrusion‐tolerant protocol assumes that some of the SCADA system components are compromised by attackers and do not operate properly. The protocol allows SCADA to operate correctly as long as no more than a threshold fraction of the components are compromised. Furthermore, the intrusion‐tolerant ­protocol uses proactive recovery techniques to restore the compromised system components[66]. The intrusion‐ tolerant techniques were applied to firewalls as well

As a real‐time application to power substations, the reliable and resilient design of PSMSs faces theoretical and practical challenges: (i) trade‐offs between technical capabilities and costs, (ii) online locating with good accuracy and low computational burden, (iii) robust real‐time tracking of intruders, and (iv) data fusion and visualization. A comparison between PSMSs reported in the literature is summarized in ­Table 5.1. Much effort has been devoted to improving the detecting, locating, and tracking functionalities of the PSMS. Motion detection in an image sequence serves as the fundamental step. There are three major classes

­ NHANCEMENT OF THE SMART GRID E SECURITY Research on Physical Security In recent years, research efforts have been focused on PSMSs, sensor placement, risk assessment methods, and computer simulation technologies. Novel sensors have been developed and incorporated to enhance system performance and reduce the false alarm rate. In addition, the number of research projects and programs that contribute to the physical security of power grids is increasing (see Table 5.3).

Table 5.3  Research projects and programs on physical security. Project title

Organizer

Related objectives

AFTER (A Framework for electrical power sysTems vulnerability identification, dEfense, and Restoration) EPCIP (European Programme on Critical Infrastructure Protection)

EU’s Seventh Framework Programme (FP7)

To develop a method for detection and early warning of physical intrusions into energy system facilities, especially substations

European Commission

IPSIG (Infrastructure Protection and Security Interest Group) Programme

Center for Energy Advancement through Technological Innovation (CEATI) North American Electric Reliability Corporation (NERC)

To develop procedures for the identification and assessment of European Critical Infrastructures, and approaches to protecting them To enhance the design and functioning of security operations for dams and power houses, substations, and transmission/ distribution grids To develop a standard that identifies and protects transmission substations and their associated primary control centers with respect to physical attacks

CIP‐014–2 (Physical Security)

Physical and Cybersecurity in a Smart Grid Environment  95

Colored images

Grayscale images Optical flow method

Binary images Foreground - white Background - black

Marked colored images

Connected-component labeling A red rectangle for each intruder

Figure 5.4  Typical flowchart of image processing for detection.

of motion detection methods: background subtraction, temporal differencing, and optical flow. A typical flowchart of image processing for detection is shown in Figure  5.4. Background subtraction methods [68, 69] maintain a reliable background model to achieve high accuracy. Therefore, they are generally used for stationary cameras. A foreground detection algorithm[68] is developed to be robust against illumination changes and noise using gauss mixture models. It provides a novel and practical choice for intelligent video surveillance systems using static cameras. However, this algorithm requires the environmental illumination to be only white and objects to have Lambertian surfaces. In practical substation scenes, these conditions are not always satisfied. An effective background subtraction method[69] is proposed to handle dynamic backgrounds and illumination variations. An array of dynamic texture models within the spatiotemporal representations is maintained, resulting in a high computational burden. In order to improve the algorithm efficiency, parallel computing algorithms may be needed to process each part of the scene in parallel. To achieve a good tracking performance, pantilt‐ zoom (PTZ) cameras need to detect intruders actively. The acronym PTZ indicates that the camera can be moved horizontally, vertically, and zoomed in and out. Considering the outdoor environment of a power substation, the optical flow method is more practical than the temporal differencing method for PTZ ­cameras. Although temporal differencing is adaptive to environmental changes, its threshold is sensitive to noises and variations in illumination[70]. In addition, it is sensitive to undesired camera motions. These usually result in incomplete detection of the shapes of moving objects. The optical flow method was initially proposed in Ref.[71]. It shows the projected motion on the image plane to handle complex backgrounds.

Most variations of the basic optical flow method are based on the well‐known brightness constancy assumption[72]. Therefore, they cannot work properly with large illumination changes. Future research is needed to develop more illumination‐robust optical flow methods with lower computational costs. The location of intruders is important for the determination of threats and proper responses. If the size of the object is unknown, the 3D location of an object cannot be obtained through the video stream from only one camera without assumptions. While the distance between two points can be accurately obtained through laser distance sensors[73], the cost of maintenance and calibration may be high. Binocular positioning methods are developed, leading to twice the number of video cameras and pictures[74]. Therefore, a method that uses triangulation is proposed to calculate the intruder’s location using a single camera[75]. It requires a realistic assumption that intruders are in contact with the ground. In real  practice, video cameras should be calibrated periodically to guarantee the accuracy of locating. Future research is needed to develop an adaptive self‐calibration algorithm for robust and precise camera calibration. A critical capability of the surveillance system is to track intruders in the time‐varying environment. Centralized schemes are used by most tracking methods[76]. For example, good performance can be achieved by the master−slave mode[77, 78] through collaborations between fixed and PTZ cameras. The master−slave mode utilizes staticwide FOV cameras as masters to monitor a large area and PTZ cameras as slaves to track an intruder detected by the master. In some applications, the centralized scheme may be prohibitively expensive. An alternative solution is to use distributed schemes that utilize distributed resources to perform local computations. Furthermore, the distributed system can self‐adapt to changes in the system topology. For example, distributed tracking algorithms[79, 80] in the camera network are developed using a multi‐agent system (MAS). Each camera is modeled as an agent that performs the computation based on its own sensed data and data from neighboring agents. As a computational paradigm, MAS is a good solution for distributed control schemes through the use of distributed artificial intelligence techniques. When the number of targets is larger than the number of cameras, the coordination of cameras for tracking multiple targets becomes a critical problem for both centralized and distributed schemes. Future research is needed to bridge this gap for enhancing the performance of an online practical ­surveillance system.

96  Advances in Energy Systems

Sensor Placement The problem of sensor placement is not emphasized among many multisensor monitoring systems[81]. However, the layout of sensors is highly related to the coverage of the monitoring area that affects the system performance significantly. For example, although PTZ cameras are very flexible, their visible areas are restricted by their positions. The optimal placement of multiple sensors can be modeled and solved as a multiobjective problem. Given the ground plan of a substation, specific task requirements, regions of interest, and budgetary constraints, the optimal allocation of resources can be estimated and evaluated. In general, an approach of multisensor placement includes five steps[82]: modeling sensors, modeling the field (i.e. space discretization), visibility analysis, setting objectives, and optimization calculation. As a critical step, visibility analysis is used to calculate the observable area of a sensor based on its characteristics and the geographical features including topography, layout, vegetation, and buildings[83]. A maximum effective distance should be set for each sensor. Those areas that are beyond a distance threshold from the sensor are regarded as invisible even if they are within the sensor’s FOV. The visibility analysis can be performed by geographic information systems (GISs). The GIS is a system designed to capture, store, manipulate, analyze, manage, and present all types of spatial and/or geographical data. In general, sensor positions are limited by potential locations for simplifying the optimization calculation. Particle swarm optimization (PSO) and potential field‐based coverage enhancing algorithm (PFCEA) are applied to optimize sensor placement problems[82]. Regarding the difficulties of optimization calculation, some commercial optimization software packages are adopted, e.g. IBM ILOG CPLEX[84]. Most of the methods consider model sensor placement as an optimization problem with a single objective, using the coverage rate as the final index. However, sensor placement should be modeled as a multiobjective optimization problem. For example, the rate of coverage and sufficient overlap between FOVs should be considered together. In Ref.[83], a simple weighted sum of multiple objectives is used as the cost function. Although multiple objectives are considered, the selection of weighting factors remains an issue.

Risk Assessment Methods Risk assessment methods are desired to assist utilities in optimizing investments and improving the security of critical power facilities. Various bilevel

and trilevel models have been proposed for identifying system vulnerabilities and obtaining an optimal defense plan. An interdiction model[85] is proposed for identification of critical sets of components in a power grid and solved by a decomposition‐based heuristic algorithm. Similarly, an interdict power flow (IPF) model[86] is developed and solved by a global Benders decomposition method. In addition, attacker‐defender (AD), defender‐attacker (DA), and defender‐attacker‐defender (DAD) models[87] are proposed for enhancing the resilience of critical infrastructure against terrorist attacks. All these models rely on an assumption that the attacker knows the defender’s reactions to attack scenarios and uses the defender’s reaction to maximize the damage. Although this assumption is not necessarily true, the target of these models is to protect power systems against worst‐case attacks. Similarly, a solution procedure is proposed to use the mixed‐integer bilevel programming model of the electric grid ­security under a disruptive threat[88]. This method requires the identification of the adversary’s motives before an attack happens, which is hard to achieve. While the worst‐case scenario leads to the maximum loss, its probability of occurrence is low. As a result, the expected value of investment in the worst‐case attacks may be low. Therefore, some methods [25, 89] are developed to protect critical components from high‐probability, low‐loss events. By ranking components with respect to their impact on the system performance, vulnerabilities are identified, and the reliability of the power system is improved. The severity risk index (SRI) and its refinement are proposed by North American Electric Reliability Corporation (NERC) for evaluating the impact of all risk‐significant events[89]. With respect to potential threats to the physical security of power substations, the threat severity index (TSI)[25] is developed based on the SRI in evaluating the impact of physical intrusion events. In addition, the substation critical index (SCI) is established to identify and rank critical substations. Power flow analysis is conducted to determine the amount of load shedding if a substation is de‐energized. Currently, the assessment of PSMSs and physical security of power facilities are performed separately. Physical attacks are simply modeled as a loss of components in the power flow model. The success rates of physical attacks are either assumed without considering the PSMSs or even set to be determined. There is no model that integrates security systems, the process of physical attacks, and power systems together. The cost−benefit analysis of PSMSs should be conducted for the physical security of power systems.

Physical and Cybersecurity in a Smart Grid Environment  97

Figure 5.5  Three‐dimensional substation scene.

Modeling, Simulation, and Visualization of Power Substations It is important to collect real data from a drill for validation of security monitoring methodologies. However, it is time‐consuming and costly to conduct an actual drill. An alternative solution is to develop a computer simulation environment that produces synthetic data for ­debugging, testing, and verification. A computer method[90] is proposed to generate the pseudo‐synthetic video. In view of physical intrusion scenarios, a critical functionality of such a computer simulation environment is to establish 3D substation scenes. Computer simulation technologies, especially virtual reality technologies, have been used to develop substation simulation systems for training system operators[91–93]. 3D virtual models of power substations[94] and methods for rendering 3D substation scenes[95] are reported. This substation scene building and rendering method includes three parts: objects modeling, scene graph building, and scene graph rendering. Different modeling process stages of creating 3D virtual models for power substations are described. A real‐case application of the development of a 3D virtual model for a transmission substation (230 kV and 60 MVA) is provided. In the authors’ related work[47, 75], computer simulation environments are established for this substation based on Webots[96] and the built‐in Virtual Reality Toolbox of MATLAB, respectively. Webots is a development environment for modeling, programming, and simulating mobile robots. 3D graphic objects in the substation scene are designed using Trimble SketchUp,

as shown in Figure 5.5. In addition, a data presenter is developed to present the processed data to the system operator for situation awareness. For lack of a specialized GIS, the widely available Google Maps is utilized to present a geographic display by the data presenter. Furthermore, intrusion events can be reconstructed and represented by the data presenter. Coordinated attacks on multiple substations need to be simulated and evaluated simultaneously. In the future, a platform that supports simultaneous simulation and visualization of multiple substations is desired. It is challenging to render a performance, especially when elaborate virtual models are adopted. In addition, the environmental changes and disturbances to sensors should be considered. Research on Cybersecurity Power system communication and its cybersecurity were identified as essential parts of a smart grid infrastructure  8. The main cybersecurity issues and access points into the substation communication networks were highlighted. The importance of cybersecurity for smart grids was recognized as a sensitive issue in the context of emerging cyber threats[97]. Cyberthreats Potential threats from a broad range of cyberattacks have become a serious concern. Research is emerging to enhance the cybersecurity of smart grids. The false data attacks in a smart grid environment were

98  Advances in Energy Systems

investigated[98]. The authors analyzed the vulnerabilities of the measurement system to false data attacks on the communicated measurements. A method was proposed for the efficient computation of a security index that shows the minimal number of corrupted measurements in order to provide incorrect consistent information without being detected by bad data detection algorithms. Based on such an index, the security resources can be better assigned. The proposed method improves the robustness of system observability, assuming a full set of measurements and considering malicious tampering of some measurements. Further research is needed to extend the proposed method to the general case, i.e. the system is not fully measured. The false data‐injection attacks against the power system state estimator were analyzed as well[99]. It is admitted that intruders can bypass the bad data detection schemes, which poses a threat to power grid operation. The authors researched ways to find the optimal data‐injection attack strategy that selects a set of meters to manipulate and cause maximum damage. Defense strategies against such attacks are proposed. A spatial‐ and temporal‐based detection mechanism was developed to identify the false data‐injection attacks. An algorithm is proposed to identify and protect the critical sensors. The proposed defense methods improve the resilience of power grids with respect to data‐injection attacks. The data‐integrity attacks against the smart grid were also investigated[100]. The authors addressed the issue of unobservable low‐sparsity cyberattacks by compromising a small number of meters. Countermeasures are proposed against a collection of data‐integrity cyberattacks. These are based on a minimum number of known secure phasor measurement units at chosen buses. Future research is needed to solve the stale data problem in state estimation. As meter readings arrive asynchronously, time delays may occur that increase the time for the algorithm to converge. In this case, state estimation may not be suitable for detecting data‐integrity cyberattacks. Another motivating type of cyberattack is the undetectable man‐in‐ the‐middle attack on the smart grid topology[101]. The intruder alters data to mislead system operators with an incorrect system topology. For a set of vulnerable meters, the authors set the conditions that make such cyberattacks undetectable. A countermeasure is proposed based on the implementation of strong authentication measures at a subset of meters. This research shows that such undetectable cyberattacks would not occur if a subset of meters that satisfy certain conditions was protected by security controls. However, further research is needed to identify other security mechanisms based on more sophisticated bad data

detection techniques and system dynamics. Taxonomy of cyberattacks for smart grids was proposed[102]. The framework systematically constructs the taxonomy that relies on the structure of space‐time and data flow direction, security features, and cyber‐physical causality. The cyber‐physical relationship was used to propose a decentralized cyberattack detection scheme. The detection model senses the effects of cyberattacks and provides the corresponding mitigation actions to defend the smart grid. Further investigations are needed for a realistic and comprehensive analysis of cyber threats that incorporates the power grid operating practice under normal and contingency operations. Experimental evaluations of cyberattacks to power grid control systems were reported[103–105]. The flooding‐based DoS and exploit‐based attacks were modeled and simulated. The experiments demonstrate that cyber intrusions into the communication infrastructure supporting smart grids are possible. Their success depends on the means of bypassing or circumventing the security controls protecting the ICTs, e.g. network access controls and user authentication mechanisms. Attacks may target the SCADA system to affect the power grid’s monitoring and control capabilities, e.g. opening circuit breakers and modifying power and voltage set points. As a result, there is a critical need for future research on the interactions between ICTs and power grids. Encryption‐Based Security Technologies The CPS security was analyzed[28]. The current CPS security concerns and research efforts to secure the smart grid are presented. The authors propose a risk assessment methodology that incorporates the cyber and physical characteristics for impact analysis of cyberattacks. A classification of power applications for smart grid control and their cybersecurity concerns are reported. Further research challenges on cybersecurity of CPS are highlighted. Security technologies for smart grids were proposed to guard against cyber threats[106]. The authors suggest the use of public key infrastructure technologies along with trusted computing elements, supported by firewalls, strong user and device authentication, and message privacy and integrity. Naturally, the time to encrypt/decrypt data may increase with the key size, which can affect the performance of control systems. Further investigations are needed for high‐performance encryption algorithms that effectively secure the control systems in a smart grid environment. A comprehensive survey on the security issues for critical infrastructures was reported[107]. The authors proposed a framework to secure SCADA systems and smart grids. It consists

Physical and Cybersecurity in a Smart Grid Environment  99

of four main components, i.e. real‐time monitoring, anomaly detection, impact analysis, and mitigation. Attack trees were used to evaluate system vulnerabilities by identifying the objectives of intruders. The algorithm for ICT vulnerability assessment involves port auditing and evaluation of the password entropy. A methodology based on attack trees is proposed for impact analysis of cyberattacks. However, the proposed methodology hypothetically evaluates system vulnerabilities in a simplified manner. Further research is needed for methods to efficiently deliver information from substations to help operators identify critical messages quickly and better assess system vulnerabilities. The ICT and power grid vulnerabilities with incomplete information were investigated using graph theory[108]. In this research, cyber and physical vulnerability models given incomplete information are integrated. The authors proposed the model for a cyberattack that resulted in the rapid opening and closing of a circuit breaker near a generator. This causes excessive torque and leads to physical damage. Research is needed to develop appropriate computational methods to mitigate such coordinated cyber‐physical attacks against the smart grid. Intrusion Detection A novel anomaly inference algorithm was reported for early detection of cyber intrusions in the substation communication networks based on a systematic extraction of the intrusion footprints[61]. The method considers temporal anomalies in datasets. The detection algorithm is used for single and multiple simultaneous intrusions in substation networks. An impact factor is computed to assess the impact of substation outages on the power grid. Further research is needed to extend the capabilities of the proposed anomaly detection algorithm for large‐scale CPS evaluations in a control center environment. The temporal anomaly detection method was generalized for host‐based anomaly detection[36]. In this research, the authors proposed an efficient algorithm for attack similarity. The proposed intrusion detection system searches for the same type of attacks on multiple substations. The generalized method uses a comprehensive set of data logs and messages from multiple substation communication networks. A method is also proposed for network‐based anomaly detection for multicast messages in the substation automation network. The IEC 61850 communication standard was used for the multicast messages. The integrated anomaly detection methods were validated by realistic intrusion scenarios on a SCADA test bed. Future research is needed to include other substation automation communication

protocols, e.g. DNP‐and IEC 60870‐5‐based anomalies. The techniques and challenges for integrating intrusion detection systems into a CPS were discussed[62]. The authors examine the unique properties and security requirements of a CPS. With the proposed framework, the intrusion detection systems are better integrated for an enhanced protection of the CPS layers. However, other emergent research problems must be addressed in terms of architecture, auditing of data source, detection method, response, and performance evaluation. Authentication Mechanisms An efficient authentication and authorization framework for smart objects was reported[65]. The proposed set of lightweight authentication and authorization mechanisms secure certain assets of a smart grid environment during their life cycle. Further research is needed for the design and development of security solutions of other smart grid components. A framework to secure the smart grid electric vehicle ecosystem was presented[35]. A two‐factor cyber–physical device authentication protocol was proposed to guard against the coordinated cyberattacks in a smart grid environment that target the integration of electric vehicles into the grid. This method can be extended to efficiently secure other devices in the smart grid. Multicast messages are critical to the control systems in a smart grid environment. They are used in many smart grid applications, e.g., wide area protection and demand response. Therefore, multicast authentication is crucial for large and distributed systems. The security requirements for multicast communication and authentication for smart grids were identified[109]. A one‐time signature scheme is proposed for multicast authentication. The method has a short authentication delay and low computational cost. Compared to other similar methods, this authentication scheme has a reduced storage cost and signature size. Thus, the proposed method is suitable to secure the smart grid applications. A different method was proposed for a rapid authentication that is appropriate for time‐ critical authentication of command and control messages[110]. However, the rapid authentication method depends on predefined structures. Research is needed to mitigate this issue. To protect against the message injection and replay attacks, the Merkle hash tree technique was employed for an efficient authentication scheme[34]. Furthermore, the proposed security solution has a reduced computational cost and can resist the message analysis and modification attacks. Further investigations are required to protect against other challenging cyber threats such as the DoS attack.

100  Advances in Energy Systems

A security protocol was proposed for the advanced metering infrastructure in a smart grid environment[111]. The protocol addresses different security vulnerabilities in the metering infrastructure for smart grids. The user privacy and behavior are secured. The protocol provides message authentication for meter readings and control messages. The proposed integrated authentication and confidentiality protocol has a good performance in terms of end‐to‐end delay and packet loss. Future research is needed to adapt the proposed security protocol to multicast and broadcast in the advanced metering infrastructures of a smart grid. Overall, the proposed solutions against malicious cyberattacks are encouraging, but further security issues need to be addressed, especially in the emerging context of continuously evolving cyber‐ physical threats. ­PHYSICAL AND CYBERSECURITY INTERDEPENDENCY Two intrusion scenarios were developed to emphasize the interdependency between physical and ­cybersecurity. They were derived from the physical

and cyber intrusion scenarios presented in Refs.[75, 112], respectively. The physical intrusion scenarios in Ref.[75] were designed with the industry support. The scenarios demonstrate the importance of both physical and cybersecurity for power substations. A physical breach of the substation security could open a gateway to allow cyberattackers to access the substation communication network and launch cyberattacks. Intrusion Scenario Without the PSMS Scenario 1 is concerned with an attempt by an individual to physically break into the premises of an unmanned substation. The target is the control room. The substation layout and intrusion path are shown in Figure 5.6. The objective is to break into the control room and plug in the preconfigured USB flash drive with malware into substation ICTs, e.g. routers/firewalls. The malware changes firewall rules and opens predefined communication ports to provide unauthorized remote access into the communication network for cyberattackers. In this scenario, it is assumed that the physical intruder does not have the knowledge or intention to operate the

Intrusion path C20 C19

C18

Substation fence

C17

C13

C14

C11 C10

C12

C15

C16

Conductors in storage Transformers

C3 C4

C2

C8

C5 C6

Transformers

Substation control room

C9

C7

C1

Figure 5.6  Substation layout and the intrusion path.

Physical and Cybersecurity in a Smart Grid Environment  101

Engineering workstation

Station HMIs

LAN

WEB HMI

Router firewall

RTU

Server

Server IED i

IED 1 Figure 5.7  Substation information and communications technology model.

power devices and cause damage to the power infrastructure. Furthermore, it would be dangerous for the physical intruder to manually disconnect switches without opening the circuit breakers first. The intruder may be part of an organization that planned the cyberattack to deliver the malware and open the gateway into the communication network. The ICT model for the substation communication network[112] is presented in Figure 5.7. It includes the bay and station levels. At the bay level, the local operating network (LON) connects the IEDs and RTU using a star topology. They directly interact with the power devices. The IEDs monitor and control the substation bays. The RTU facilitates the communications with both the substation control room and control centers. At the station level, the LAN is based on ethernet that networks the ICT devices, e.g. communication servers, human machine interfaces (HMIs), and workstations for remote access. Due to the lack of an efficient physical security monitoring system, the attacker successfully intrudes into the substation control room and achieves the goal. The firewall that protects the substation communication network is compromised. The cyber intruders have remote access to the substation network. The cyber intrusion path is represented in Figure 5.7. They further compromise the security controls that protect the communication server and RTU. Once the RTU is compromised, the cyberattackers obtain unauthorized control capabilities over the substations’ IEDs. The cyberattack model is presented in Ref.[112]. The considered attack is the unauthorized access and control of one of the main assets in the substation, i.e. RTU. A series of attack steps are needed for the attack path to reach the target. Each step is to compromise an ICT device along the path. Compromising an ICT device means that the attackers defeat the security controls, which guard the ICT device against cyberattacks. It is assumed that all ICTs are vulnerable to cyberattacks. An attack takes t hours to breach the security controls. The time delays to compromise

each ICT device are considered to be between tattack, ∈ [0.1, 100] hours. If it is 0.1, the ICT device is considered extremely vulnerable. The attack is successful. Otherwise, if the attack time delay is 100 hours, it is considered secured, and the attack fails. Usually, penetration tests are used to discover the ICT vulnerabilities in communication networks and penetration times[113]. The success rate to compromise the ICT min min device is attack ,i tattack /tattack ,i 100 [%], where tattack is the lower limit on the attack time delays interval. The rate of success considers the time needed by the attackers to bypass the security controls and access the  ICT. The result is the percentage to compromise the ICT device. It ranges from 0.1% to 100% success. The success rate for the entire jth attack path with nMC ICT machines to be compromised is path attack , j

nMC

path attack , i .

i 1

The cyberattack was performed on a SCADA test bed [37,112]. The penetration times and success rates to compromise the firewall, communication server, and RTU along the path and the success rate for the entire attack path are provided in Table 5.4. With help from inside the substation, the time needed by cyberattackers to compromise the firewall is low, i.e. 0.1001 hour. Thus, the success rate to compromise the firewall and access the substation network is 99.9%. Security controls are in place to protect the communication server and RTU. However, they are vulnerable to cyberattacks. The cyber intruders need only 0.145 and 0.110 hour to defeat them. The success rates to compromise these ICT devices are 68.9% and 90.9%, respectively. The overall success rate for the attack path is 62.5%. Assuming that 100 cyber intrusions follow the same attack path, around 62 would succeed in compromising all ICT devices. Thus, the cyberattack is considered successful. Unauthorized controls are sent to connected IEDs based on the RTU’s control capabilities, e.g.  triggering open circuit breakers, modify voltage and

102  Advances in Energy Systems

Table 5.4  Cyberattack penetration times and success rates. Without physical security protection

Firewall Server RTU

tattack [h]

ξattack [%]

0.1001 0.1450 0.1100

99.9 68.9 90.9

path attack

With successful physical security protection

%

62.5

tattack [h]

ξattack [%]

Firewall Server RTU

3.7 0.1450 0.1100

path attack

%

2.7 1.7 68.9 90.9

RTU, remote terminal unit.

60.20

Frequency [Hz]

60.15 60.10 60.05 60.00 59.95

0

10

20

30

40

50

60

70

80

90

100

Time [s] G2: Electrical frequency in Hz Figure 5.8  Frequency variations.

active power set points or tap positions. The most destructive scenario is considered. IEEE 39‐bus system was used for testing. It is assumed that the power system was at a steady state condition before the ­cyberattack. The targeted substation includes buses 29 and 38. The IEDs supervise two buses, two transmission lines, one power transformer, a generator, and a load. The cyber intruders disconnect transmission lines 26–29 and load 29, i.e. 283.5 MW at 20.3 and 27.4 seconds, respectively. The tap position of the main transformer is modified at 33.3 seconds. As the ­cyberattack is ongoing at the cyber layer, the power grid is experiencing N–2 contingencies that disturb the operating condition. The power system moves from a secure state to an insecure or emergency state. The ­cyberattack has a significant impact on all electrical parameters. ­Figure  5.8 provides the grid’s frequency behavior; Figure  5.9 shows the loading levels of lines 26–28 and 28–29; Figure 5.10 indicates the voltage level at bus 29; and Figure 5.11 shows the active power consumption of load 28. The disconnection of transmission lines 26–29 increased the power flows on lines 26–28 and 28–29. The voltage at bus

29 was severally affected. The bus voltage drops from 1.02 to 0.81 p.u. The load varies with the bus voltage. At bus 28, about 74 MW load cannot be supplied. The system finds a new operation point. The security constraints are violated. After some time, the overload and under voltage may lead to undesired events that could trigger cascading events or even a blackout. Intrusion Scenario with the PSMS In Scenario 2, the intrusion path and objective are the same as those of Scenario 1. However, a substation physical security monitoring (SPSM) system[47] is used as an example of the PSMS to enhance the physical security of power substations. The layout of sensors and system architecture are shown in Figures  5.6 and 5.12, respectively. The SPSM system includes sensors and subsystems installed locally at critical substations and the event analysis system with HMI in the control center. The intelligent video surveillance system is a core part that manages the video cameras in a substation and performs detection, locating, tracking, and patrolling functionalities. When an intrusion is detected, the

1.8 1.6

Current [kA]

1.4 1.2 1.0 0.8 0.6 0.4 0.2 0

Line 26-28: Current magnitude in kA Line 28-29: Current magnitude in kA

0

10

20

30

40

50

60

70

80

90

100

60

70

80

90

100

Time [s] Figure 5.9  Current variations on lines 26–28 and 28–29.

1.10 1.05

Voltage [p.u.]

1.00 0.95 0.90 0.85 0.80 0.75 0.70

0

10

20

30

40

50 Time [s]

Bus 29: Voltage magnitude in p.u. Figure 5.10  Voltage variations at bus 29. 220

Active power [MM]

200 180 160 140 120 100

Load 28: Active power in MW

0

10

20

30

40

50 Time [s]

Figure 5.11  Load variations at bus 28.

60

70

80

90

100

104  Advances in Energy Systems

Control center

Communication network

Event analysis system On-line alarm processing module

Substation

Network interface

Data visualization module

Central database

Substation

Substation 1 Video cameras Video stream

Human-machine interface

Substation

Sensors for indoor and perimeter monitoring

System operator

Alarms

Intelligent video surveillance system

On-site database

Alarms

Remote backup Network interface

Figure 5.12  Architecture of the substation physical security monitoring system.

SPSM system processes sensor measurements locally. Only the processed data is transmitted to the control center. It assists the system operator in situation awareness. Otherwise, much time may be spent on exhaustive search of the events and implications. In addition, it provides real‐time event analysis that estimates the intention of the intruder through evaluating the intrusion trajectory. Thus, the system operator is able to evaluate the impact of threats on the power grids and take proper preventive and remedial actions. The system response to a physical intrusion is shown in Figure  5.13. The SPSM system provides continuous online monitoring. An alarm is triggered once an intrusion is detected. With the support of the SPSM system, the remote security operator checks the situation quickly. If an intrusion is confirmed, the police are informed immediately. The intention of the intruder is investigated and judged by an intrusion classification method. The intrusion target is estimated, and proper actions can be applied. In this scenario, the intruder is detected while climbing over the substation fence, and the police are informed. Note that early detection is critical to gain the response time. The intrusion target is judged to be the substation control room by analyzing the ongoing intrusion trajectory. Sirens, strobe lights, and loud speakers are activated to deter and slow down the intruder. A remote backup is launched to protect the local data. In addition, the system operator remotely blocks all access to computers, firewalls, and the substation LAN in the substation control room through operating the access control system. Thus, the intruder has to break the door even with an authorized key or the

password to the substation control room. This results in a delay in sabotage. As a result, the intruder fails to use the infected USB flash drive before the police arrive. Finally, the intruder is arrested while escaping. The physical attack was performed using a test platform with synthetic datasets[47]. The false positive fp rate of tracking tra , false negative rates of detection fn fn and tracking tra, and the success rate for the det security monitoring of five intrusion paths are presented in Table  5.5. The success rate of the SPSM path fn fp fn is SPSM 1 det 1 tra 1 tra . The false positive rate of detection is not considered, as it only affects the rate of false alarms when there is no intrusion. Table  5.5 shows that the success rate of the SPSM in an intrusion scenario is highly relevant to the selection of intrusion path. This is true in practice. The intrusion path in Scenario 2 is selected as the fifth one. Assuming that 100 intrusion events use this path in a year, around six events may not be detectable with respect to the 5.71% false negative rate. Considering the failure of the tracking functionality, the success rate of SPSM is calculated to be 77.3%. This is a good ratio as around 77 events will be handled successfully. As the physical intruder fails to accomplish the objective, the cyberattackers do not have access to the substation communication network. They try to bypass the firewall by exploiting the existing vulnerabilities. The penetration times and attack success rates are presented in Table  5.4. The success rate to compromise the firewall using hacking tools is 2.7%. The success rates for the communication server and RTU remain the same. The overall success rate for the attack path is 1.7%. This is significantly lower than

Physical and Cybersecurity in a Smart Grid Environment  105

Online monitoring

Decision support in preventive and remedial actions

Recognize an intrusion?

No

Threat assessment

Yes Trigger an alarm

Intrusion classification Support security operators in situation awareness

Mute the alarm

Yes

Evidences of an intrusion?

No

If needed Turn on/off the substation illumination system

Inform police and power grid operators

If needed

If needed

Activate sirens, strobe lights, and loud speakers

Block access to substation control room

Figure 5.13  Procedure of the system response. Table 5.5  Success rates of the SPSM system in five intrusion paths. Tracking Intrusion path path 1 2 3 4 5

Detection false fn negative det [%]

False positive fp tra [%]

False negative fp tra [%]

Success rate of path SPSM SPSM [%]

1.57 1.64 4.61 4.70 5.71

1.00 1.99 3.97 4.47 7.33

3.50 6.88 7.36 7.84 11.53

94.04 89.77 84.86 83.90 77.30

SPSM, substation physical security monitoring.

that of the first scenario, i.e. 62.5%. Therefore, the cyber intruders cannot access the communication network, and the cyberattack fails.

­CONCLUSION An interdisciplinary review of physical and cybersecurity in a smart grid environment is conducted. Worldwide physical and cyber intrusion incidents are reported. The importance of the smart grid security is highlighted. Smart grid vulnerabilities and current security controls are identified. A review

that summarizes the research efforts to enhance the smart grid security is presented. In addition, the interdependency of physical and cybersecurity is illustrated using an intrusion scenario with computer simulations. Obviously, to establish comprehensive physical protection to all power infrastructures would be too costly. However, it is feasible and desirable to strengthen some key facilities such as critical substations. Under budgetary constraints, substations should be selected for security upgrade to maximize the system reliability. It is envisioned that physical and cybersecurity systems will be reorganized and extended with respect

106  Advances in Energy Systems

to novel sensors and technologies. For instance, the deployment of radar sensors at power substations can increase the monitoring area significantly and reduce the number of video cameras. In addition, outside‐the‐ fence areas can be monitored in an economical way with a lower false alarm rate. Further enhancements can be achieved by developing systematic methodologies that evaluate the performance of security systems. Thus, multiple intrusion scenarios can be tested by drills or computer simulations over a short‐term or long‐term period. Other promising research directions include (i) distributed detection of cyber intrusions upon the physical security system; (ii) development of a model that integrates security systems, the process of physical attacks, and power systems together; (iii) the cost‐benefit analysis of security systems for optimizing investment in system update or installation; and (iv) optimal placement of sensors concerning multiple objectives. ­ACKNOWLEDGMENTS The research on physical security was supported by EU Framework Programme FP7 at University College Dublin (UCD) through the AFTER project, “A Framework for electrical power sysTems vulnerability identification, dEfense and Restoration.” The research on cybersecurity was started by the Science Foundation Ireland (SFI) at UCD through a Principal Investigator Award, and the follow‐on work is sponsored by the US National Science Foundation Grant 1202229, “Collaborative Research: Resiliency against Coordinated Cyberattacks on Power Grid.”

­REFERENCES 1. Hokstad, P., Utne, I.B., and Vatn, J. (2013). Risk and Interdependencies in Critical Infrastructures. New York: Springer. 2. Sovacool, B.K. (2012). Energy security: challenges and needs. Wiley Interdiscip. Rev. Energy Environ. 1: 51–59. 3. Bose, A. (2010). Smart transmission grid applications and their supporting infrastructure. IEEE Trans. Smart Grid 1: 11–19. 4. Liu, C.‐C., Jung, J., Heydt, G.T. et al. (2000). The strategic power infrastructure defense (SPID) system. A conceptual design. IEEE Control Syst. Mag. 20: 40–52. 5. Li, H., Rosenwald, G.W., Jung, J. et al. (2005). Strategic power infrastructure defense. Proc. IEEE 93: 918–933. 6. Smith, R. (2014). Assault on California power station raises alarm on potential for terrorism. Wall St. J. (February 5). 7. Chenine, M., Ullberg, J., Nordstrom, L. et al. (2014). A framework for wide‐area monitoring and control systems interoperability and cybersecurity analysis. IEEE Trans. Power Delivery 29: 633–641.

8. Ericsson, G.N. (2010). Cyber security and power system communication – essential parts of a smart grid infrastructure. IEEE Trans. Power Delivery 25: 1501–1507. 9. Ruttledge, L. and Flynn, D. (2014). Short‐term fre quency response of power systems with high non‐synchronous penetration levels. Wiley Interdiscip. Rev. Energy Environ. 4: 452–470. 10. Laqueur, W. (1999). The New Terrorism: Fanaticism and the Arms of Mass Destruction, 302. Oxford: Oxford University Press. 11. Ofuji, K. (2014). Recent improvements and challenges in Fukushima: prefecture’s economy and living environment. Wiley Interdiscip. Rev. Energy Environ. 4: 307–315. 12. Barclay, R.A. (2014). Regulatory economics: cybersecurity − who cares? Threat and apathy worldwide, outlook uncertain. Nat. Gas Electricity 30: 30–32. 13. McDonald, J.D. (May 2012). Electric Power Substations Engineering, 3e, 536. New York, NY: CRC Press. 14. Canadian Electricity Association (CEA). Copper theft from Canada’s electricity infrastructure: dangerous, expensive and a threat to reliability, 2015. Available at: http://www.electricity.ca/media/CopperTheft/Copper Theft.pdf (Accessed April 13, 2015). 15. Smith, R. (2014). PG&E Silicon Valley substation is breached again. Wall St. J.. 16. Bompard, E., Huang, T., Wu, Y., and Cremenescu, M. (2013). Classification and trend analysis of threats origins to the security of power systems. Int. J. Electr. Power Energy Syst. 50, 50: 64. 17. Chen, T.M. and Abu‐Nimeh, S. (2011). Lessons from Stuxnet. IEEE Comput. Soc. 44: 91–93. 18. Industrial Control Systems Cyber Emergency Response Team. Incident response activity. ICS‐CERT Monitor Newsletters. Available at: https://icscert.us‐cert.gov/ monitors (Accessed June 1, 2015). 19. Brown, M.A. and Zhou, S. (2013). Smart‐grid policies: an international review. Wiley Interdiscip. Rev. Energy Environ. 2: 121–139. 20. Goel, S., Hong, Y., Papakonstantinou, V. et  al. (2015). Smart Grid Security. London: Springer. 21. Amin, M. (2002). Security challenges for the electricity infrastructure. Computer 35: 8–10. 22. Abiri, E., Rashidi, F., Niknam, T. et al. (2014). Optimal PMU placement method for complete topological observability of power system under various contingencies. Int. J. Electr. Power Energy Syst. 61: 585–593. 23. Panteli, M., Crossley, P.A., and Fitch, J. (2014). Quantifying the reliability level of system integrity protection schemes. IET Gener. Transm. Distrib. 8: 753–764. 24. Smith, R. (2014). US risks national blackout from small‐ scale attack. Wall St. J.. 25. Xie J, Liu C‐C, Sforna M, et al. Threat assessment and response for physical security of power substations. In: Proceedings of the Innovative Smart Grid Technologies (ISGT Europe 2014), Istanbul, Turkey, October, 2014. 26. Salmeron, J. and Wood, R.K. (2014). The value of recovery transformers in protecting an electric transmission grid against attack. IEEE Trans. Power Syst. 30: 2396–2403.

Physical and Cybersecurity in a Smart Grid Environment  107

27. Ohlhausen P, Bates S, Yelton L. ASIS/SIA risk assessment survey—results and analysis. ASIS Foundations, Security Industry Association, and Today’s Facility Manager Magazine, 2007. Available at: https:// foundation.asisonline.org/FoundationResearch/ Publications/Documents/asissiaRickAssessment.pdf (Accessed February 7, 2015). 28. Sridhar, S., Hahn, A., and Govindarasu, M. (2012). Cyber‐physical system security for the electric power grid. Proc. IEEE 100: 210–224. 29. Wang, S., Meng, X., and Chen, T. (2012). Wide‐area control of power systems through delayed network communication. IEEE Trans. Control Syst. Tech. 20: 495–503. 30. Stahlhut, J.W., Browne, T., Heydt, G.T. et  al. (2008). Latency viewed as a stochastic process and its impact on wide area power system control signals. IEEE Trans. Power Syst. 23: 84–91. 31. Gungor, V.C., Sahin, D., Kocak, T. et al. (2011). Smart grid technologies: communication technologies and standards. IEEE Trans. Industr. Inform. 7: 529–539. 32. Ten, C.‐W., Liu, C.‐C., and Govindarasu, M. (2008). Vulnerability assessment of cybersecurity for SCADA systems. IEEE Trans. Power Syst. 23: 1836–1846. 33. Liu, C.‐C., Stefanov, A., Hong, J. et al. (2012). Intruders in the grid. IEEE Power Energy Mag. 10: 58–66. 34. Li, H., Lu, R., Zhou, L. et  al. (2014). An efficient Merkle‐tree‐based authentication scheme for smart grid. IEEE Syst. J. 8: 655–663. 35. Chan, A.C.‐F. and Zhou, J. (2014). Cyber‐physical device authentication for the smart grid electric v­ ehicle ecosystem. IEEE J. Sel. Areas Commun. 32: 1509–1517. 36. Hong, J., Liu, C.‐C., and Govindarasu, M. (2014). Integrated anomaly detection for cyber security of the substations. IEEE Trans. Smart Grid 5: 1643–1653. 37. Stefanov A, Liu C‐C. Cyber‐physical system security and impact analysis. In: Proceedings of the International Federation Automatic Control (IFAC), Cape Town, August, 2014. 38. Garcia, M.L. (2007). Design and Evaluation of Physical Protection Systems. Oxford, MS: Butterworth‐ Heinemann. 39. Kong Y, Jing M. An identification method of abnormal patterns for video surveillance in unmanned substation. In: 2011 Asia‐Pacific Power and Energy Engineering Conference (APPEEC), March, 2011. 40. Ge, J., Tong, L., Chen, Q. et  al. (2002). Unmanned ­substations employ multimedia network RTUs. IEEE Comput. Appl. Power 15: 36–40. 41. Tag and track (TNT). Ipsotek, 2011. Available at: http:// www.ipsotek.com/?q=en/news/48 (Accessed May 7, 2015). 42. Hampapur A, Brown L, Connell J, et  al. S3‐R1: The IBM smart surveillance system‐release 1. In: Proceedings of the 2004 ACM SIGMM Workshop Effect of Telepresence, New York, NY, USA, October 15, 2004. 43. Surveillance SiteIQ. Siemens, 2012. Available at: http:// www.buildingtechnologies.siemens.com/bt/global/en/ Pages/home.aspx (Accessed May 7, 2015).

44. Pro‐Watch. Honeywell, 2015. Available at: http://www. honeywellintegrated.com/products/integrated‐security/ sms/index.html (Accessed May 7, 2015). 45. FLIR Thermal Fence. FLIR, 2015. Available at: http:// www.flir.com/cvs/americas/en/security/view?id=44287 (Accessed May 7, 2015). 46. Collins, R.T., Lipton, A.J., Fujiyoshi, H. et  al. (2001). Algorithms for cooperative multisensor surveillance. Proc. IEEE 89: 1456–1477. 47. Xie, J., Liu, C.‐C., Sforna, M. et al. (2015). Online physical security monitoring of power substations. Int. Trans. Electr. Energy Syst. 26 (6): 1148–1170. 48. Girgensohn A, Kimber D, Vaughan J, et al. DOTS: support for effective video surveillance. In: Proceedings of the 15th International Conference on Multimedia, New York, NY, USA, 2007. 49. Siebel NT, Maybank SJ. The ADVISOR visual surveillance system. In: Proceedings of the ECCV Workshop on Applications of Computer Vision (ACV ‘04), Prague, Czech Republic, May, 2004. 50. Lee, S.C. and Nevatia, R. (2014). Hierarchical abnormal event detection by real time and semi‐real time multitasking video surveillance system. Mach. Vis. Appl. 25: 133–143. 51. Monti Guarnieri, A., Broquetas, A., Recchia, A. et  al. (2015). Advanced radar geosynchronous observation system: ARGOS. IEEE Geosci. Remote Sens. Lett. 12: 1406–1410. 52. Pelletier M, Sivagnanam S, Blasch EP. A track scoring MOP for perimeter surveillance radar evaluation. In: Proceedings of 15th International Conference on Information Fusion, Singapore, July, 2012. 53. Sobhani, B., Paolini, E., Giorgetti, A. et  al. (2014). Target tracking for UWB multistatic radar sensor networks. IEEE J. Sel. Top. Sign. Proces. 8: 125–136. 54. Smith, R. (2014). Breaking news: how the big business stories of 2014 turned out (power‐grid security raises alarm). Wall St. J.. 55. Baker, P.R. and Benny, D.J. (2013). The Complete Guide to Physical Security. New York, NY: CRC Press. 56. Ritter, L. (2013). Behind the curtain‐decision support tools. Maritime Profess. 3 (4): 49–51. 57. Bird N. A new approach for assessing operational nuclear security performance—an overview. National Nuclear Laboratory (NNL), 2013. Available at: http:// www.worldsecurity‐index.com/details.php?id=5794 (Accessed April 25, 2015). 58. BluTrain—the training solution for security professionals. ARES Corporation, 2009. Available at: http://www. aressecuritycorp.com/download/592 (Accessed April 25, 2015). 59. Salah, K., Elbadawi, K., and Boutaba, R. (2012). Performance modeling and analysis of network firewalls. IEEE Trans. Netw. Serv. Manage. 9: 12–21. 60. Ling‐Fang H. The firewall technology study of network perimeter security. In: Proceedings of IEEE Asia‐Pacific Services Computing Conference (APSCC), Guilin, December, 2012.

108  Advances in Energy Systems

61. Ten, C.‐W., Hong, J., and Liu, C.‐C. (2011). Anomaly detection for cybersecurity of the substations. IEEE Trans. Smart Grid 2: 865–873. 62. Han, S., Xie, M., Chen, H.‐H. et  al. (2014). Intrusion detection in cyber‐physical systems: techniques and challenges. IEEE Syst. J. 8: 1049–1059. 63. Fan, C.‐I., Huang, S.‐Y., and Lai, Y.‐L. (2014). Privacy‐ enhanced data aggregation scheme against internal attackers in smart grid. IEEE Trans. Industr. Inform. 10: 666–675. 64. Liu, T., Liu, Y., Mao, Y. et al. (2014). A dynamic secret‐ based encryption scheme for smart grid wireless communication. IEEE Trans. Smart Grid 5: 1175–1182. 65. Hernandez‐Ramos, J.L., Pawlowski, M.P., Jara, A.J. et  al. (2015). Toward a lightweight authentication and authorization framework for smart objects. IEEE J. Sel. Areas Commun. 33: 690–702. 66. Kirsch, J., Goose, S., Amir, Y. et al. (2014). Survivable SCADA via intrusion‐tolerant replication. IEEE Trans. Smart Grid 5: 60–70. 67. Bessani, A.N., Sousa, P., Correia, M. et al. (2008). The crutial way of critical infrastructure protection. IEEE Secur. Privacy 6: 44–51. 68. Li, D., Xu, L., and Goodman, E.D. (2013). Illumination‐ robust foreground detection in a video surveillance system. IEEE Trans. Circuits Syst. Video Technol. 23: 1637–1650. 69. Lin, L., Xu, Y., Liang, X. et  al. (2014). Complex background subtraction by pursuing dynamic spatio‐temporal models. IEEE Trans. Image Process. 23: 3191–3202. 70. Barbu, T. (2014). Pedestrian detection and tracking using temporal differencing and HOG features. Comput. Electr. Eng. 40: 1072–1079. 71. Horn, B.K. and Schunck, B.G. (1981). Determining optical flow. Artif. Intell. 17: 185–203. 72. Mohamed, M.A., Rashwan, H.A., Mertsching, B. et al. (2014). Illumination‐robust optical flow using a local directional pattern. IEEE Trans. Circuits Syst. Video Technol. 24: 1499–1508. 73. Paris, C. and Bruzzone, L. (2015). A three‐dimensional model‐based approach to the estimation of the tree top height by fusing low‐density LiDAR data and very high resolution optical images. IEEE Trans. Geosci. Remote Sens. 53: 467–480. 74. Shih, S.‐E. and Tsai, W.‐H. (2013). Optimal design and placement of omni‐cameras in binocular vision systems for accurate 3‐D data measurement. IEEE Trans. Circuits Syst. Video Technol. 23: 1911–1926. 75. Xie J, Liu C‐C. Monitoring of physical security at substations in a power grid. In: Proceedings of 11th Intelligent System Applications to Power Systems (ISAP ‘13), Tokyo, Japan, July, 2013. 76. Smeulders, A.W.M., Chu, D.M., Cucchiara, R. et  al. (2014). Visual tracking: an experimental survey. IEEE Trans. Pattern Anal. Mach. Intell. 36: 1442–1468. 77. Cui, Z., Li, A., Feng, G., and Jiang, K. (2015). Cooperative object tracking using dual‐pan‐tilt‐zoom cameras based on planar ground assumption. IET Comput. Vision 9: 149–161.

78. Huang, C.‐M. and Fu, L.‐C. (2011). Multitarget visual tracking based effective surveillance with cooperation of multiple active cameras. IEEE Trans. Syst. Man Cybern. Part B Cybern. 41: 234–247. 79. Song, B., Kamal, A.T., Soto, C. et al. (2010). Tracking and activity recognition through consensus in distributed camera networks. IEEE Trans. Image Process. 19: 2564–2579. 80. Liu, W., Chan, A.B., Lau, R.W.H. et  al. (2015). Leveraging long‐term predictions and online learning in agent‐based multiple person tracking. IEEE Trans. Circuits Syst. Video Technol. 25: 399–410. 81. Sun, Q., Xu, D., Wu, Z. et al. (2014). Coverage performance analysis of multi‐camera networks based on observing reliability model. Int. J. Light Electron Opt. 125: 2220–2224. 82. Morsly, Y., Aouf, N., Djouadi, M.S. et al. (2012). Particle swarm optimization inspired probability algorithm for optimal camera network placement. IEEE Sens. J. 12: 1402–1412. 83. Chen, C., Yao, Y., Hsu, W. et  al. (2015). Continuous camera placement using multiple objective optimisation process. IET Comput. Vision 9: 340–353. 84. Murray, A.T., Kim, K., Davis, J.W. et  al. (2007). Coverage optimization to support security monitoring. Comput. Environ. Urban Syst. 31: 133–147. 85. Salmeron, J., Wood, K., and Baldick, R. (2004). Analysis of electric grid security under terrorist threat. IEEE Trans. Power Syst. 19: 905–912. 86. Salmeron, J., Wood, K., and Baldick, R. (2009). Worst‐ case interdiction analysis of large‐scale electric power grids. IEEE Power Syst. 24: 96–104. 87. Brown, G., Carlyle, M., Salmeron, J. et  al. (2006). Defending critical infrastructure. Interfaces 36: 530–544. 88. Motto, A.L., Arroyo, J.M., and Galiana, F.D. (2005). A mixed‐integer LP procedure for the analysis of electric grid security under disruptive threat. IEEE Trans. Power Syst. 20: 1357–1365. 89. Integrated risk assessment approach—refinement to severity risk index. NERC, 2011. Available at: http:// www.nerc.com/docs/pc/rmwg/Integrated_Bulk_‐ Power_System_Risk_Assessment_Concepts_Final.pdf (Accessed May 18, 2015). 90. Black J, Ellis T, Rosin P. A novel method for video tracking performance evaluation. In: IEEE International Workshop on Visual Surveillance and Performance Evaluation of Tracking and Surveillance (VS‐PETS), Nice, France, October 11–12, 2003. 91. Okapuu‐von Veh, A., Marceau, R.J., Malowany, A. et al. (1996). Design and operation of a virtual reality operator‐ training system. IEEE Trans. Power Syst. 11: 1585–1591. 92. Liu Q, Tai N, Shang J, et al. Research and implementation of 3D training system for substation simulation. In: 22nd International Conference and Exhibition on Electricity Distribution (CIRED 2013), Stockholm, June, 2013. 93. Arroyo E, Arcos JLL. SRV: a virtual reality application to electrical substations operation training. In: Proceedings of the IEEE International Conference on Multimedia Computing and Systems, Florence, June, 1999.

Physical and Cybersecurity in a Smart Grid Environment  109

94. Quintana J, Mendoza E. 3D virtual models applied in power substation projects. In: Proceedings of the 15th International Conference on intelligent System Applications to Power Systems (ISAP ‘09), Curitiba, Brazil, November 8–12, 2009. 95. Guangwei Y, Zhitao G. Scene graph organization and rendering in 3D substation simulation system. In: Asia‐ Pacific Power and Energy Engineering Conference (APPEEC), Wuhan, China, March, 2009. 96. Eaton, M. (2015). Evolutionary Humanoid Robotics. New York, NY: Springer. 97. McBride, A.J. and McGee, A.R. (2012). Assessing smart grid security. Bell Labs Tech. J. 17: 87–103. 98. Hendrickx, J., Johansson, K.H., Jungers, R.M. et  al. (2014). Efficient computations of a security index for false data attacks in power networks. IEEE Trans. Autom. Control 59: 3194–3208. 99. Yang, Q., Yang, J., Yu, W. et al. (2014). On false data‐ injection attacks against power system state estimation: modeling and countermeasures. IEEE Trans. Parallel Distrib. Syst. 25: 717–729. 100. Giani, A., Bitar, E., Garcia, M. et al. (2013). Smart grid data integrity attacks. IEEE Trans. Smart Grid 4: 1244–1253. 101. Kim, J. and Tong, L. (2013). On topology attack of a smart grid: undetectable attacks and countermeasures. IEEE J. Sel. Areas Commun. 31: 1294–1305. 102. Hu, J., Pota, H.R., and Guo, S. (2014). Taxonomy of attacks for agent‐based smart grids. IEEE Trans. Parallel Distrib. Syst. 25: 1886–1895. 103. Dondossola G, Garrone F, Szanto J. Performance evaluation of standard power grid communications experiencing cyber anomalies. In: Proceedings of the FITCE Congress, Palermo, September, 2011.

104. Dondossola G, Garrone F, Szanto J. Supporting cyber risk assessment of power control systems with experimental data. In: Proceedings of the IEEE PES Power Systems Conference Exposition (PSC), Seattle, March, 2009. 105. Dondossola G, Garrone F, Szanto J. Cyber risk assessment of power control systems—a metrics weighted by attack experiments. In: Proceedings of the IEEE PES General Meeting, San Diego, July, 2011. 106. Metke, A.R. and Ekl, R.L. (2010). Security technology for smart grid networks. IEEE Trans. Smart Grid 1: 99–107. 107. Ten, C.W., Govindarasu, M., and Liu, C.‐C. (2010). Cybersecurity for critical infrastructures: attack and defense modeling. IEEE Trans. Syst. Man Cybern. Part A Syst. Humans 40: 853–865. 108. Srivastava, A., Morris, T., Ernster, T. et  al. (2013). Modeling cyber‐physical vulnerability of the smart grid with incomplete information. IEEE Trans. Smart Grid 4: 235–244. 109. Li, Q. and Cao, G. (2011). Multicast authentication in the smart grid with one‐time signature. IEEE Trans. Smart Grid 2: 686–696. 110. Yavuz, A.A. (2014). An efficient real‐time broadcast authentication scheme for command and control messages. IEEE Trans. Inf. Forensic Secur. 9: 1733–1742. 111. Yan, Y., Hu, R.Q., Das, S.K. et al. (2013). An efficient security protocol for advanced metering infrastructure in smart grid. IEEE Network 27: 64–71. 112. Stefanov, A., Liu, C.‐C., Govindarasu, M. et al. (2015). SCADA modeling for performance and vulnerability assessment of integrated cyber‐physical systems. Int. Trans. Electr. Energy Syst. 25: 498–519. 113. Arkin, B., Stender, S., and McGraw, G. (2005). Software penetration testing. IEEE Secur. Privacy 3: 84–87.

6

Energy Security: Challenges and Needs Benjamin K. Sovacool Institute for Energy and the Environment, Vermont Law School, South Royalton, VT, USA

This chapter describes four broad categories of emerging global energy security threats. After defining energy security, it discusses threats related to the availability of energy resources and fuels, the affordability of the services they provide, the efficiency of their use, and the effective management of their negative environmental and social impacts, notably climate change. It concludes by noting how the energy security of many industrialized and developing countries has eroded over the past few decades and proposes potential areas of future research. ­INTRODUCTION Energy security is an urgent priority for policy makers and scholars worldwide. As the global economy continues its climb out of recession, satisfying the demand for available, accessible, affordable, and environmentally acceptable mobility, cooking, heating, and cooling will remain a precondition for sustaining economic development[1]. The collective energy security challenges facing the world are more than just an inconvenience and present a sizeable barrier to worldwide prosperity. They result

in higher transportation and electricity costs that at the micro level can drive households into poverty[2]. Energy insecurity can also lead to bankruptcy of firms unable to pass on extra costs to their customers[3]. At the macro level, defending energy supplies, and the transfer of wealth from energy importers to exporters, can shape global geopolitics and involve trillions of dollars of resources[4–6]. After decades of discussion, the energy security debate now rightly focuses on a critical global dilemma: Can the world have secure, reliable, and affordable supplies of energy while also transitioning to a low‐­ carbon energy system? The evolution of this debate highlights the multifaceted nature of the energy security dilemma, although the key dimensions of the problem are still being disputed[7–9]. In turn, the global political economy exhibits a diverse array of energy security strategies and policies. This reflects a lack of consensus about the nature of the energy security problem as well as the need for different approaches to address diverse resource endowment, political levers, capital availability, risk aversion, and other particularities. Energy security has long centered on questions of reliable energy supplies, the regional concentration of

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

112  Advances in Energy Systems

energy resources, and the implications of the strategic withholding of energy fuels, mostly oil[10]. This view recognizes that energy is essential for any form of economic activities; increasing energy consumption has characterized industrialization and economic development over the past century[11]. Recent assessments of energy security have expanded to embrace electricity reliability as well as natural gas and petroleum security, and the entire energy supply chain including energy delivery infrastructure[10]. Many events have underscored the entire energy supply chain’s vulnerability to many different types of disruption, including political instability in the Middle East, natural disasters such as the Japanese tsunami and with the Fukushima nuclear reactor meltdown in 2011, gas disputes with Russia in 2008 and 2009 that have wreaked havoc in European electricity markets, and equipment breakdowns in the northeast United States blackout of 2003. Hurricanes Katrina and Rita in 2005 illustrated a major diversified energy disaster, with the simultaneous disruptions of oil, natural gas, and electric power.

­DEFINING ENERGY SECURITY Although security of supply, mostly procuring oil, gas, coal, and uranium, remains a core concern for national and global policymakers, much recent scholarship has recognized that priorities with respect to global warming, air pollution, economic growth, and energy affordability will define how the transition to a secure energy future is to be achieved[12–18]. This chapter defines energy security as “equitably providing available, affordable, reliable, efficient, environmentally benign, proactively governed, and socially acceptable energy services to end users.” This conception of energy security comes from both a literature review of peer‐reviewed studies on energy security offered in[8] as well as research interviews with energy experts[9] and surveys of energy end users[19]. Essentially, this definition breaks energy security down into four interconnected criteria or dimensions: availability, affordability, efficiency, and stewardship. Availability refers to diversifying the fuels used to provide energy services as well as the location of facilities using those fuels, promoting energy systems that can recover quickly from attack or disruption, and minimizing dependence on foreign suppliers. Affordability refers to providing energy services that are affordable for consumers and minimize price volatility. Efficiency involves improving the performance of energy equipment and altering consumer behavior to reduce energy price exposure and mitigate energy

import dependency. Stewardship consists of protecting the natural environment, communities, and future generations. Unfortunately, each of these energy security dimensions is currently under threat.

­THREATS TO AVAILABILITY The world’s four primary energy fuels are concentrated in a dramatically low number of countries, creating significant patterns of import dependence, and these reserves are running out. The world’s known 1.2 trillion barrels of oil reserves are concentrated in volatile regions of the world, as are the world’s largest petroleum companies. The three biggest of these – Saudi Arabian Oil Company, National Iranian Oil Company, and Qatar Petroleum – own more crude oil than the next 40 largest oil companies combined. The 12 largest oil companies control roughly 80% of petroleum reserves and are all state owned. The distribution of other conventional energy resources, such as coal, natural gas, and uranium, is equally consolidated. Figure 6.1 provides a graphic illustration of energy reserves by country. Strikingly, 80% of the world’s oil can be found in nine countries that have only 5% of the world population and 5% of gross domestic product (GDP), 80% of the world’s natural gas is in 13 countries with 12% of the population and 26% of GDP, and 80% of the world’s coal is in six countries (though these countries have 45% of the population and 46% of GDP). Many of the same countries are among the six that control more than 80% of global uranium resources. Although these reserves may seem vast, a growing worldwide demand for electricity and mobility threatens to exhaust them relatively soon. The world is transitioning from a position of abundant fossil energy supplies to a largely resource‐constrained supply future. World energy demand is expected to expand by 45% between now and 2030, and by more than 300% by the end of the century[20]. If levels of production were to remain constant worldwide, ­Table  6.1 illustrates that known coal reserves will become ­exhausted within 137 years, and petroleum and natural gas reserves would become exhausted in the next half century. If rates of production increase to keep up with growing demand, particularly in the rapidly developing BRIC countries (Brazil, Russia, India, and China), known fossil‐fuel reserves would be depleted much more rapidly. Additionally, if the world were to maintain its generation of nuclear power at 2004 levels, identified uranium resources would run out in 85 years.

Energy Security: Challenges and Needs  113

1200 Uranium Coal Natural gas Oil

Billion barrels of oil equivalent

1000

800

600

400

200

ab

zu

el

a

t ai Ve

ne

Ku

w

Ira

q

s ira

ai

Em

kr U

te

ne

ar at

a

Q

ut

h

Af

st

ric

an

U

ni

te

d

Ar

a

kh

a

di In

za Ka

So

U

ni

Sa

ud

iA

ra

bi

n Ira

lia ra

a

na

st Au

hi C

si us R

te

d

St

at

es

0

Figure 6.1  Major global energy reserves for the top 15 countries, 2008.

Table 6.1  Life expectancy of proven fossil fuel and uranium resources. Life expectancy (years) Proven reserves Coal

930 400 million short tons Natural gas 6189 trillion cubic feet Petroleum 1317 billion barrels Identified Uranium 4 743 000 tons (at $130/kg U)

Current production

0% Annual production growth rate

2.5% Production 5% Production growth rate growth rate

6807 million short tons

137

60

42

104.0 trillion cubic feet 30.560 billion barrels

60 43

37 29

28 23

40 260 tons

85

66

48

Source: Reprinted with permission from Ref 20. Copyright 2011, MIT press.

For oil in particular, a growing imbalance of production and consumption exacerbates the risk of fuel shortages and interruptions in supply, which will take a fairly rapid turn for the worst for many nations if alternative fuels are not widely deployed. The likely geographic pattern of oil production and consumption over the next two decades suggests that oil dependence in Europe, China, India, and other Asian countries could grow rapidly, each relying on imports to meet more than 75% of oil demand by 2030[21].

The issue of oil dependence is just as stark for many developing countries. Rises in the cost of crude oil and gasoline mean that the foreign exchange required for oil imports create a heavy burden on the balance of trade for many developing countries. While developed countries spend just 1% or 2% of their GDP on imported oil, those in the developing world spend an average of 4.5–9% of their GDP on crude oil imports. Higher prices for oil also hit developing countries twice: once for costlier barrels of oil and again for

114  Advances in Energy Systems

inflated transportation costs that reflect the increase in fuel prices[22]. Nor are the dangers of dependence limited to oil. The electricity sectors in the United States and European Union, among others, have become dependent on  foreign sources of natural gas. As suppliers in ­countries such as Algeria, Indonesia, Qatar, Russia, and Venezuela come to account for a greater proportion of global natural gas supply, the patterns (and subsequent costs) associated with import dependence could mimic trends in the oil sector. The costs of such dependence can be staggering. Researchers at Oak Ridge National Laboratory estimated that from 1970 to 2004 American dependence on foreign supplies of oil has cost the country $5.6– 14.6 trillion[5] or more than the costs of all wars fought by the country going back to the Revolutionary War (including both invasions of Iraq). To those who may express dismay at such a high estimate, consider that researchers from the University of California‐­Davis and University of Alaska‐Anchorage calculated that US defense expenditures to exclusively protect oil in the Persian Gulf amount to about US$27 billion to US$73 billion per year[6]. Importantly, the use of cleaner forms of energy – such as wind turbines, solar panels, hydroelectric dams, geothermal facilities, biomass power plants, and some biofuels  –  can reduce dependency and stabilize costs. Wind and solar generators, for example, lower demand for natural gas and thereby decrease natural gas rates for both electricity users and natural gas consumers. These cleaner technologies diversify the energy base, thereby providing more stable energy prices and insulating the industry from price spikes, interruptions, shortages, accidents, delays, and international conflicts. Unlike generators relying on oil, natural gas, uranium, and coal, renewable generators are generally not subject to the rise and fall of fuel costs. They thus provide a hedge against future environmental regulations (such as a carbon tax) that could make the price of conventional power unexpectedly rise, and diversify the energy portfolio. They can also respond more rapidly to supply and demand fluctuations, improving the efficiency of the market because of their modularity[23].

­THREATS TO AFFORDABILITY Challenges to the affordability of energy services are reflected not only in rising prices for oil, other fuels, and electricity, but also lack of access to modern energy carriers, largely because people in the developing world cannot afford it.

While electricity rates have not experienced the meteoric rise of oil prices, which have hovered above $100 per barrel for most of 2011, they have been steadily increasing. Over the past two years, residential electricity rates in the United States have increased 6% above the rate of inflation. In Europe, iron and steel manufacturers are experiencing increasing electricity and raw materials prices, which are damaging their already weakened industries. Other factors are contributing to escalating prices, including: • China, India, and even Europe are consuming more coal for power production and steel mills, pushing coal prices up. • The threat of tighter environmental regulations such as carbon caps has caused producers to invest in costly pollution abatement equipment. • Transportation bottlenecks are inhibiting the ability of global fuel suppliers to meet demand. The official electricity price forecasts in the United States predict a 25% rise in retail electricity prices over the next 20 years[24]. As another metric of value, consider the lives of the 1.4 billion people in the world who have little to no access to electricity. As a result of “electricity deprivation” or “energy poverty,” millions of women and children spend significant amounts of time searching for firewood, and then burning either it or charcoal indoors to heat their homes and prepare their meals, emitting localized pollution into the living space. As Table  6.2 reveals, worldwide, nearly 2.7 billion people use traditional biomass fuels for cooking and heating[25]. An additional billion people have access only to unreliable or intermittent electricity networks.

Table 6.2  Number of people without access to electricity and dependent on traditional fuels, 2009.

Africa Sub‐Saharan Africa Asia China India Other Asia South America World

Number of people lacking access to electricity (millions)

Number of people relying on the traditional use of biomass for cooking (millions)

587 585

657 653

799 8 404 387 31

1937 423 855 659 85

1417

2679

Energy Security: Challenges and Needs  115

Put another way, the poorest three quarters of the global population still only use 10% of global energy. Even accounting for significant increases in development assistance and rural electrification, by 2030 about 1.4 billion will still be at risk of having to live without modern energy services. The indoor air pollution resulting from cookstoves shortens the life of 2.8 million people every year, almost even with the number dying annually from HIV/AIDS[26]. Close to one million of these deaths – 910 000 – are children under the age of five years who must suffer their final months of life dealing with debilitating respiratory infections, chronic obstructive pulmonary disease, and cancer.

­THREATS TO EFFICIENCY Energy systems remain incredibly inefficient. Most commercial thermoelectric power plants convert the potential chemical energy of coal and other fuels into thermal energy at 100% efficiency, but typical usable energy is just 33% of that, with the remaining two‐ thirds simply wasted. One study suggested that the poor thermal efficiency of most power plants in the United States translates into enough potential power lost every year to meet the needs of 5–10 cities the size of Manhattan[23]. In the end, when you add the layers of inefficiency from a coal plant across the distribution network to the consumer, only 2–3% of the energy embodied within the coal used to produce that electricity is converted into lumens inside an incandescent bulb[20]. Similarly, a typical internal combustion engine in an automobile only utilizes 15−25% of the energy of the gasoline to move the car, with the rest of the energy, again, being lost along the way[20]. One of the problems with making the system more efficient is that energy savings come in small pieces rather than concentrated in the large chunks that attract ribbon cutters to energy‐supply facilities. Also, about 20 huge power plants operate to energize US appliances and equipment that are turned off, just to keep them in standby mode[27]. There is thus an immense amount of potential to improve the efficiency of how we produce and consume energy. Energy efficiency and demand‐side management – doing more with less, lowering levels of energy consumption by substituting fuels and technologies and altering consumer behavior – are often the cheapest and quickest ways to address energy ­security threats. Renewable power generators, in contrast to their nuclear‐ and fossil‐fueled counterparts, utilize sunlight, wind, falling water, biomass, waste, and geothermal heat to produce electricity from fuels

that are mostly free for the taking. These “fuels” happen to be in great abundance in every country in the world, and thus offer a way to make electricity sectors less susceptible to supply chain interruptions and shortages. Yet the potential of energy efficiency and renewable energy remains hindered by a collection of pernicious barriers. For instance, a litany of recent US and European studies has documented barriers to innovation for cleaner and more efficient energy systems at every stage of the commercialization and deployment process. The Interlaboratory Working Group identified scores of issues relating to misplaced incentives, inconsistent regulations, and information and market failures[28, 29]. Another study tabled the barriers to renewable energy penetration, highlighting in particular the problem of missing market infrastructure that may increase costs[30]. Yet another assessment found that subsidies for conventional forms of energy, high initial capital costs, imperfect capital markets, lack of skills or information, poor market acceptance, technology prejudice, financing risks and uncertainties, high transaction costs, and a variety of regulatory and institutional factors impede investments in clean energy[31]. One of my own studies involved research interviews with more than 180 experts working for utilities, government agencies, and the national laboratories and identified 38 nontechnical barriers to the deployment of distributed generation, renewable energy, and energy‐efficiency technologies[32]. Another investigation of the barriers to energy‐efficient technologies and energy‐efficiency programs found 80 separate types of barriers[27].

­THREATS TO STEWARDSHIP Every kilowatt‐watt hour of conventional electricity generated, barrel of oil produced, ton of uranium mined, or cubic foot of natural gas manufactured produces a laundry list of environmental damages that may include radioactive waste and abandoned uranium mines and mills, acid rain and its damage to fisheries and crops, water degradation and excessive consumption, particle pollution, and cumulative environmental damage to ecosystems and biodiversity through species loss and habitat destruction. If put into monetary terms, the social and environmental damage from just one type of energy – worldwide electricity generation – amounted to roughly $2.6 trillion in 2010[20]. Perhaps the most severe social and environmental consequence of current patterns of energy production and use is climate change. The destabilization

116  Advances in Energy Systems

of the world’s climatic zones, driven by relentless greenhouse gas emissions, has the potential to exacerbate food and water shortages, advance the spread of infectious disease, induce mass migration, damage trillions of dollars of property, and precipitate extreme weather events – all of which could lead to increased conflict worldwide[33]. The impacts of global warming on industrialized economies are likely to be of historic proportion, challenging previous global threats such as AIDS, the Great Depression, and global terrorism. The Pew Center on Global Climate Change estimates that, in the Southeast and Southern Great Plains of the United States, the financial costs of climate change could reach as high as $138 billion by 2100. Pew researchers warn that “waiting until the future” to address global climate change might bankrupt the US economy[34]. The Stern Report projected that the overall costs and risks of climate change will be equivalent to losing at least 5% of the world’s GDP, or $3.2 trillion, every year, now, and forever, and that these damages could exceed 20% of GDP if more severe scenarios unfold[35]. Many of these negative impacts will occur in the developing world. According to one study published in the Lancet, as many as 40 least developed countries with a total population of three billion people could lose 20% of their cereal production by 2080. Alterations to the ranges of agricultural pests and diseases with warmer winters could cause massive infestations of locusts, whiteflies, and aphids that could create extensive losses of crop yields. Over the past three decades precipitation across the Sahel in Africa has declined by 25%, contributing to hunger and malnutrition in the Niger Delta, Somalia, and Sudan. Most experts anticipate severe climate‐induced shortages of food in Angola, Burkina Faso, Chad, Ethiopia, Mali, Mozambique, Senegal, Sierra Leone, and Zimbabwe that will starve 87 million people[36]. Another study warned that as many as 75–250 million people in Africa could be exposed to increased water stress by 2020 as yields from rain‐fed farms fall by 50%[37]. Yet another assessment explored the climate change‐related risks for a sample of eight ­countries – China, Guyana, India, Mali, Samoa, Tanzania, United Kingdom, United States  –  covering a range of geographic locations, economies, and lifestyles[38]. It found that climate risks could cost some countries as much as 19% of their GDP by 2030, with the biggest impacts falling on developing countries. Some states, such as Maharashtra, India, will be prone to drought that will likely wipe out 30% of food production, inducing $7 billion in damages among 15 million small and marginal farmers. The Asian Development Bank also warned in a separate study that because of

their unique geography, countries such as Indonesia, Philippines, Thailand, and Vietnam will suffer more from climate change than the global average. These four countries alone are expected to lose 6.7% of combined GDP, or $86 billion, by 2100, more than twice the rate of global average losses if business as usual continues[39]. Similarly, Tuvalu, the Marshall Islands, and low‐ lying parts of the Caribbean, Papua New Guinea, and Bangladesh could be submerged within 60 years if sea levels continue to rise. The Republic of Kiribati, a small island country in the Pacific, has already had to relocate 94 000 people living in shoreline communities and coral atolls to higher ground. The Republic of the Maldives could lose 80% of its land because of rises in sea level and has already started purchasing land in Sri Lanka for its “climate refugees.” Melting glaciers will flood river valleys in Kashmir and Nepal, and reduced rainfall will aggravate water and food security so that 182 million people could die of disease epidemics and starvation attributable to climate change[20].

­CONCLUSION The world faces an interconnected, and daunting, set of energy security challenges that seem to be getting worse as time progresses, perhaps because the global economy continues to consume more energy. Indeed, one comparison of key energy security trends among 22 countries in the Organization for Economic Cooperation and Development (OECD) from 1970 to 2007 found that every country saw a deterioration of energy security, with even the best performers seeing indicators degrade across a variety of sectors and most countries seeing a net degradation of energy ­security[8]. A follow‐up study using a more fine‐tuned methodology evaluated energy security trends among Asia’s four large energy users – China, India, Japan, and South Korea – as well as the developing countries of Southeast Asia and reached the same conclusion[40]. The lesson appears to be that although the energy crises of the 1970s catalyzed global efforts to improve energy security, the opposite is occurring. Despite the establishment of the International Energy Agency, the creation of strategic petroleum reserves among its members, and the diversification of the fuel base for electricity as most countries moved away from their use of oil to produce electricity, energy security continues to worsen[41]. The international community has seen advances in low‐income energy services, efficiency, and demand reduction programs, renewable resources initiatives, and market restructuring of

Energy Security: Challenges and Needs  117

the various energy industries. Many countries have implemented aggressive renewable portfolio standards and systems benefit funds, started emissions trading schemes, passed feed‐in tariffs, and invested heavily in alternative fuels such as hydrogen, ethanol, and biodiesel. Yet these efforts have not meaningfully improved energy security. If this troubling conclusion is true, what should policy makers do about it, and what further research is needed? One area of research could be analyzing which factors cause energy security to rapidly degrade in particular countries, or, conversely, if any best practices exist that are empirically successful at improving energy security. Perhaps for some countries it would be the introduction of a new policy or technology, for others it might be changes in pricing or shifts in consumer attitudes, for still others major historical events. Determining these factors, and exploring both quantitatively and qualitative how they might contribute to “improved” or “degraded” energy security, could reveal previously unseen relationships between “external” events beyond a country, “internal” events within a country, and its overall energy security. Other research could look more closely at the tradeoffs involved with different aspects of energy security. As the evaluations given above[8, 40] demonstrate, as some energy security indicators, such as per capita carbon dioxide emissions or access to water, improve, others, such as price stability or subsidies, often worsen. Improvement in other dimensions, such as sustainability or governance, may also see availability and affordability worsen. Discovering the scale, scope, and structure of such trade‐offs would be an  instrumental part of designing synergistic energy security strategies that improve all elements of energy security simultaneously. A final area of inquiry could examine not the least or most energy secure country, or most or least improved, but the speed of change, both positive and negative, that a country’s energy security undergoes. This could help analysts help see which elements of energy security can be fixed or tweaked relatively quickly compared with those that may necessitate more effort[40].

REFERENCES 1. Modi, V., McDade, S., Lallement, D. et al. (2005). Energy Services for the Millennium Development Goals. Washington and New York: The International Bank for Reconstruction and Development/The World Bank and the United Nations Development Programme. 2. Pachauri, S. (2011). The energy poverty dimension of energy security. In: The Routledge Handbook of Energy

Security (ed. B.K. Sovacool), 191–204. London: Routledge. 3. Joscow, P.L. (2001). California’s electricity crisis. Oxford Rev. Econ. Policy 17: 365–388. 4. Klare, M.T. (2007). The futile pursuit of energy security by military force. Brown J. World Affairs 13: 144. 5. Greene D, Ahmad S. Costs of US oil dependence: 2005 update (January 2005). Report to the U.S. DOE, ORNL/ TM‐2005/45; 2005. 6. Delucchi, M.A. and Murphy, J.J. (2008). US military expenditures to protect the use of Persian gulf oil for motor vehicles. Energy Policy 36: 2253–2264. 7. Sovacool, B.K. (ed.) (2011). The Routledge Handbook of Energy Security. London: Routledge. 8. Sovacool, B.K. and Brown, M.A. (2010). Competing dimensions of energy security: an international review. Annu. Rev. Environ. Resour. 35: 77–108. 9. Sovacool, B.K. (2011). Evaluating energy security in the Asia Pacific: towards a more comprehensive approach. Energy Policy 39: 7472–7479. 10. Yergin, D. (2006). Ensuring energy security. Foreign Affairs 85: 69–82. 11. Warr, B.S. and Ayres, R.U. (2010). Evidence of causality between the quantity and quality of energy consumption and economic growth. Energy 35: 1688–1693. 12. Kruyt, B., Vuuren, D.P.V., de Vries, H.J.M. et al. (2009). Indicators for energy security. Energy Policy 37: 2166–2181. 13. Jacobson, M.Z. (2009). Review of solutions to global warming, air pollution, and energy security. Energy Environ. Sci. 2: 148–173. 14. Vivoda, V. (2010). Evaluating energy security in the Asia‐Pacific region: a novel methodological approach. Energy Policy 38: 5258–5263. 15. Jansen, J.C. and Seebregts, A.J. (2010). Long‐term energy services security: what is it and how can it be measured and valued? Energy Policy 38: 1654–1664. 16. Pode, R. (2010). Addressing India’s energy security and options for decreasing energy dependency. Renewable Sustainable Energy Rev. 14: 3014–3022. 17. Badea, A.C., CM Rocco, S., Tarantola, S. et al. (2011). Composite indicators for security of energy supply using ordered weighted averaging. Reliab. Eng. Syst. Saf. 96: 651–662. 18. Bollen, J., Hers, S., and van der Zwaan, B. (2010). An integrated assessment of climate change, air pollution, and energy security policy. Energy Policy 38: 4021–4030. 19. Sovacool, B.K., Valentine, S.V., Bambawale, M.J. et al. (2012). Exploring propositions about perceptions of energy systems and energy security: an international survey. Environ. Sci. Policy 16: 44–64. 20. Brown, M.A. and Sovacool, B.K. (2011). Climate Change and Global Energy Security: Technology and Policy Options. Cambridge: MIT Press. 21. International Energy Agency (IEA) (2008;; Figure 3.10). World Energy Outlook 2008, 105. Paris, France: IEA. 22. Sovacool, B.K. (2009). Sound climate, energy, and transport policy for a carbon constrained world. Policy Soc. 27: 273–283.

118  Advances in Energy Systems

23. Sovacool, B.K. (2008). The Dirty Energy Dilemma: What’s Blocking Clean Power in the United States. Westport: Praegar. 24. Energy Information Administration (EIA) (2009). Annual Energy Outlook 2009. Paris, France: IEA Table A8. Available at: http://www.eia.doe.gov/oiaf/archive/ aeo09/index.html. 25. International Energy Agency, United Nations Development Program, and United Nations Industrial Development Organization (2010). Energy Poverty: How to Make Modern Energy Access Universal? Paris: OECD. 26. Holdren, J.P. and Smith, K.R. (2000). Energy, the environment, and health. In: World Energy Assessment: Energy and the Challenge of Sustainability (ed. T. Kjellstrom, D. Streets and X. Wang), 61–110. New York: United Nations Development Programme. 27. Lovins, A.B. (2007). Energy myth nine—energy efficiency improvements have already maximized their potential. In: Energy and American Society (ed. B.K. Sovacool and M.A. Brown). New York: Springer Publishing Company. 28. Interlaboratory Working Group (2000). Scenarios for a Clean Energy Future, ORNL/CON‐476 and LBNL‐44029. Oak Ridge, TN/Berkeley, CA: Oak Ridge National Laboratory/Lawrence Berkeley National Laboratory. 29. Brown, M.A., Levine, M.D., Short, W. et  al. (2001). Scenarios for a clean energy future. Energy Policy 29: 1179–1196. 30. Painuly, J.P. (2001). Barriers to renewable energy penetration, a framework for analysis. Renew. Energy 24: 73–89. 31. Beck, F. and Martinot, E. (2004). Renewable energy policies and barriers. In: Encyclopedia of Energy (ed. C. Cleveland), 4. London: Academic Press/Elsevier Science. 32. Sovacool, B.K. (2009). Rejecting renewables: the sociotechnical impediments to renewable electricity in the United States. Energy Policy 37: 4500–4513. 33. Fingar T. ‘National Intelligence Assessment of the National Security Implications of Global Climate Change to 2030’, Testimony before the House Permanent Select Committee on Intelligence and House Select Committee on Energy Independence and Global Warming. June 25, 2008. 34. Claussen, E. and Peace, J. (2007). Energy myth twelve— climate policy will bankrupt the U.S. economy. In: Energy and American Society (ed. B.K. Sovacool and M.A. Brown), 311–340. New York: Springer Publishing Company.

35. Stern, N. (2006). The Economics of Climate Change: The Stern Review. Cambridge, UK: Cambridge University Press. 36. Haines, A., Smith, K.R., Anderson, D. et  al. (2007). Policies for accelerating access to clean energy, improving health, advancing development, and mitigating climate change. Lancet 370: 1264–1281. 37. Prouty, A.E. (2009). The clean development mechanisms and its implications for climate justice. Columbia J. Environ. Law 34: 513–540. 38. Economics of Climate Adaptation Working Group (2009). Shaping Climate‐Resilient Development: A Framework for Decision‐Making. New York: Climate Works Foundation. 39. Asian Development Bank (2009). The Economics of Climate Change in Southeast Asia: A Regional Review Manila. Manila, Philippines: ADB. 40. Sovacool, B.K., Mukherjee, I., Drupady, I.M. et  al. (2011). Evaluating energy security performance from 1990 to 2010 for eighteen countries. Energy 36: 5846–5853. 41. Brown, M.A. and Sovacool, B.K. (2007). Developing an ‘Energy Sustainability Index’ to evaluate energy policy. Interdiscip. Sci. Rev. 32: 335–349.

FURTHER READING Barton, B., Redgwell, C., Ronne, A., and Zillman, D.N. (eds.) (2004). Energy Security: Managing Risk in a Dynamic Legal and Regulatory Environment. New York, NY: Oxford University Press. Brown, M.A. and Sovacool, B.K. (2011). Climate Change and Global Energy Security: Technology and Policy Options. Cambridge, MA: MIT Press. Energy Security and Justice webpage for the Institute for Energy and the Environment: http://www.vermontlaw.edu/ Academics/Environmental_Law_Center/Institutes_and_ Initiatives/Institute_for_Energy_and_the_Environment/ Ongoing_Research_Projects.htm International Energy Agency’s ‘Energy Security’ webpage: http:// www.iea.org/subjectqueries/keyresult.aspPKEYWORD_ ID=4103. Sovacool, B.K. (ed.) (2011). The Routledge Handbook of Energy Security. London: Routledge. The International Institute of Applied Systems Analysis’s Global Energy Assessment: www.iiasa.ac.at/Research/ ENE/GEA. U.S. Energy Information Administration’s Country Profile Database: http://www.eia.gov/countries/index.cfm.

7

Nuclear and Renewables: Compatible or Contradicting? Lutz Mez Freie Universität Berlin, Environmental Policy Research Centre, Berlin, Germany

Since the report of the Club of Rome, “The Limits to Growth,” the exhaustion of fossil energy sources has been under discussion. Coal, oil, and gas – and energy technologies relying on them  –  will be ­ exhausted in the future. High expectations for nuclear energy have given way to the insight that nuclear fission is at best a bridging technology. The risks of this technology became apparent by nuclear catastrophes such as Chernobyl and Fukushima, and uranium resources are limited. Furthermore, fast breeder technology has failed and fourth‐generation reactors are still pure fiction. Before the Industrial Revolution, nearly all energy demands were supplied by renewable energy worldwide. From the beginning of the nineteenth century, fossil fuel was used in a steadily increasing amount for heating, lighting, transport, and other energy purposes. Electricity as a scientific form of energy played an important role in the transformation of agricultural to industrial societies. Electricity systems are the basic infrastructure of modern societies, influencing the industrial organization, the degree of automation, the communication system, and the industrial future. Today the electrical power industry is at a crossroads, last but not least for ­environmental

reasons. Beside the large cathedrals of electricity g­eneration, other types of generation technologies, each of which brings totally different social relations between energy producers and consumers, have a ­historical chance to emerge. The solar age will replace the fossil and the nuclear age sooner or later. This transformation is not only a technical and economic but also a social and political problem. ­INTRODUCTION Since the first report of the Club of Rome on “The Limits to Growth,”[1, 2] the exhaustion of fossil energy sources, in particular, oil, coal, and gas, has been discussed. The purpose of the report was not to make specific predictions but to explore how exponential growth interacts with finite resources. Since this publication, public consciousness over the finiteness of fossil resources was never lost. When Graham[3] examined the past 30 years of reality on the basis of the predictions made in 1972, he found that changes in industrial production, food production, and pollution are all in line with the predictions of an economic and societal collapse in the twenty‐first century. In the

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

120  Advances in Energy Systems

course of rising energy prices, the debate about the question of scarceness of energy resources and how energy supply and security can be ensured in the future resurfaced. In light of concerns over climate change, however, the ecological limits must be defined differently today than nearly 40 years ago. The increasing water shortage, the expansion of deserts, and the loss of biodiversity do not represent classical environmental problems but interacts closely with a number of social problems: the worsening of the hunger crisis and pauperization, the increase of environmental migration, and the conflicts about scarce resources, such as soil and water. Worldwide dramatic conflicts revolve around agricultural farmland and fresh water supply. The socioecological and socioeconomic crises resulting from it must be urgently counteracted. Not only the finiteness of resources, which will be reached in a few decades, is the main problem of mankind but also the effects of fossil energy consumption and the comprehensive and irreversible destruction of the ecological bases of life. Due to the social, ecological, and economic subsequent effects of climate change, the promotion of fossil energies has to be stopped – long before the last drop of oil from the Earth’s crust is pressed[4]. For social and ecological reasons, the energy industry thus already stands at a crossroads. Instead of large central power stations  –  “the cathedrals of electricity”  –  distributed generation technologies, e.g. small‐scale combined heat and power (CHP) plants, with completely different social relations between energy producers and consumers have a historical chance[5]. The fossil and the nuclear age can be replaced by the solar age. However, the political and cultural frame conditions still have to be created (for details on the cultural obstacles connected to the change of the energy system, see Ref.[6]). Politicoeconomically, however, the concentration of power in the energy sector is the biggest problem. Over the course of the twentieth century, the electricity industry has become a major player in the energy policy process in almost every country of the world. It often acts as a state in the state with a power base that could not be diminished but instead broadened over time. In its history, neither changing institutional conditions nor free market ideologies such as free trade or the political system change after the end of the East–West confrontation were able to change this industry. In its origins, it revolves strongly around technologically driven developments, which favor the emergence of powerful structures within the grid‐bound energy industry. Germany has taken key decisions concerning the structure of the power supply. Federal government and Parliament decided after Fukushima to phase

out stepwise all remaining 17 nuclear power plants (NPPs) until 2022, starting with eight reactors in 2011. Because many large power plants are close to retirement age in the next decade, including the NPPs, much of the generation capacity needed to be replaced anyhow. Actual electricity generation accounts for roughly 40% of total German carbon emissions, and electricity supply is a touchstone of German energy and climate policies. To reach the reduction target of 80–95% less carbon emissions in 2050, the German power plants need to be emission free as soon as possible because CO2 emissions cannot be reduced quickly and sufficiently in other sectors such as agriculture, transport, and heating and cooling. The only way to reduce Germany’s overall carbon emissions over the long term is to completely decarbonize the country’s electricity supply system. The low‐carbon technologies needed to do this are already available, or will be in the foreseeable future; these technologies include renewable energy such as wind, solar, biomass and geothermal energy, nuclear power, and fossil‐fuel power generation using carbon capture and storage technology[7]. The German Advisory Council on the Environment (SRU) has in several statements and reports argued that there exists a system conflict between the traditional power supply system and the emerging distributed generation by decentralized and renewable power technologies[7–9]. Although the utilities and the federal government use the term “bridging technology”1 mainly for the large nuclear power stations, the SRU refer to the environmental problems posed by coal‐fired power plants and NPPs “which are mainly of technological nature and the same, regardless of location.” Although greenhouse gas emissions from NPPs are far lower than for coal‐fired power plants, the use of nuclear power entails the risk of accidents  –  an eventuality that cannot be completely ruled out and that could have consequences for large areas and for extended periods of time; plus no viable solution has been found for long‐term storage of nuclear waste. In our view, this is a high price to pay; what’s more, nuclear power may not be a sustainable solution in view of the limited supply of uranium. Hence, in our view, neither coal‐fired power plants nor NPPs can be qualified as sustainable energy resources[7].

1  After the Fukushima accident, Federal Chancellor Angela Merkel, defined “bridging technology” as follows: “When we talk about nuclear power as ‛bridging technology’ this means nothing else that we want to phase‐out the use of nuclear power” (Süddeutsche Zeitung, March 14, 2011).

Nuclear and Renewables: Compatible or Contradicting?  121

350 300 250 200

in GWe, from 1954 to April 1, 2011 @ MYCLE SCHNEIDER CONSULTING

400

Number of reactors 500

Nuclear Reactors and Net Operating Capacity in the World

GWe

371 GWe 437 reactors

444 reactors 322 GWe 424 reactors

450 400 350

Reactors in operation

300

Operable capacity

250 200

150

150 100 100 50

50

2010 2011

2002

2000

1989 1990

1980

1970

1960

0 1954

0

Figure 7.1  Nuclear reactors and net operating capacity in the world since 1954.

The least cost‐intensive bridging technology available today for renewable electricity is not the expansion of operating time for NPPs but scaling back electricity demand by improving energy efficiency. Therefore, the SRU disbelieve in the compatibility of nuclear power and renewables, but to achieve the 100% target of renewable electricity, the ecological problems associated with renewables have to be managed and minimized through policy and planning measures.

­STATUS AND PERSPECTIVES OF NUCLEAR POWER 1. For three decades, politicians and the nuclear industry have predicted a renaissance of civilian nuclear power. On October 9, 1981, the New York Times featured an article titled: “President offers plans for revival of nuclear power”[10]. Under President Ronald Reagan, according to the article, the US government had taken concrete steps to revive commercial nuclear power. Periodically, various interest groups promise that the resurgence of nuclear energy is just around the corner. But the facts tell a different story – the nuclear industry is in terminal decline. According to the World Nuclear Industry Status Report, the number of reactors worldwide rose from 423 to 437 in the period from 1989 to 2011, an increase of less than

one reactor a year[11]. Moreover, in May 2011, there were seven fewer reactors in operation compared to 2002, when the total reached a historic high of 444 units (see Figure 7.1).2 Nuclear reactors are to be found in 31 countries, but around three quarters of global nuclear electricity is generated in only six countries – the three nuclear weapon states, the United States, France, and Russia, as well as Japan, Germany, and South Korea. Sixty‐ four reactor blocks are officially under construction and a further five are in long‐term shutdown. Closer examination of the building projects reveals that 12 reactors have been listed in the statistics as “under construction” for more than 20 years. The previous record holder for the longest construction time  –  the Bushehr‐1 NPP in Iran, for which the first foundations were poured on May 1, 1975 – has been overtaken by the American Watts Bar 2 reactor. Started more than 35 years ago in 1972, construction was halted on the project in 1985 and abandoned completely in December 1994. In 2007, the Tennessee Valley Authority announced plans to complete the reactor at an estimated cost of $2.5 billion[12]. 2  For the current information on the number of reactors, see; http://www.world‐nuclear.org/information‐library/current‐and‐ future‐generation/plans‐for‐new‐reactors‐worldwide.aspx.

122  Advances in Energy Systems

Table 7.1  Share of nuclear electricity of total final energy consumption (TFEC) of six largest producers (2008).

Total Primary Energy Supply (TPES; Mtoe) Electricity generation (TWh) Nuclear electricity generation (TWh) Share (%) TFEC (Mtoe) Share of electricity of TFEC (%) Share of nuclear electricity of TFEC (%)

France

Korea

Japan

Germany

The United States

Russia

266.50 570.3 439.5 77.1 165.55 22.5 17.3

226.95 443.9 151.0 34.0 147.54 23.7 8.1

495.84 1075.0 258.1 24.0 318.81 26.0 6.2

335.28 631.2 148.5 23.5 235.67 19.1 4.5

2283.72 4343.8 837.8 19.3 1542.25 21.3 4.1

686.76 1038.4 163.1 15.7 435.51 14.0 2.2

Source: International Energy Agency, 2010 [13].

Such long construction times incur enormous costs that few banks anywhere in the world are prepared to finance unless the risk is underwritten by the state. The decommissioning of 120 reactors with an average operating life of 22 years also militates against the ­renaissance theory. In 2006 alone, eight reactors were shut down, all of them in Europe, while only two were brought online, and construction began on six more. Assuming an operational life of 40 years, a total of 93 reactors will be taken off‐line by the year 2015, and a further 192 by 2025. At the time of writing, the International Atomic Energy Agency (IAEA) lists only two reactor blocks currently under construction in Western Europe – one in Finland and, since December 2007, and the other in France. The Finnish Olkiluoto‐3 reactor is the pilot and flagship power plant of the European Pressurized Water Reactor (EPR) project. Construction on the 1600 MW reactor began in 2005. It was sold in 2003 to a consortium under the leadership of Finnish operator Teollisuuden Voima (TVO) for a fixed sum of €3.2 billion, and was scheduled to begin commercial operation in 2009. However, the supplier AREVA NP has announced delays on different occasions since August 2005, and the earliest that operations can now begin is 2013. This means that the project is four years behind schedule and at least 90% over budget, reaching a total cost estimate of €5.7 billion ($8.3 billion) or close to €3500 ($5000) per kilowatt[11]. With AREVA, majority owned by the French state, and European competition rules a factor, it will be interesting to see who assumes the extra costs of up to €3 billion. The French state utility Electricité de France decided in October 2004 to build an EPR at the Flamanville nuclear facility in Normandy. The official construction started on December 3, 2007. This reactor was scheduled for completion in 54 months, but also at this site problems have occurred. As a result, the ambitious time schedule could not be met and commissioning is now scheduled for 2015.

The three major emerging nations – India, China, and Brazil – launched nuclear programs decades ago, but have only partially realized them. Their nuclear power output as a percentage of overall electricity and energy production is minimal[11]. This situation is unlikely to change much. A global building boom in new nuclear power stations can be ruled out in the short to medium term because of a lack of production capacity and the dwindling supply of skilled personnel. In addition, new production facilities and power plants need new staff to run them, but the nuclear industry and power plant operators are already struggling to fill jobs vacated by retiring older workers. A whole generation of engineers, atomic physicists, and radiation protection experts is missing. At the same time, closed plants have to be torn down and solutions finally found for nuclear waste. In 2008, the share of nuclear power of World Total Primary Energy Supply (TPES) was 5.8%, not even half the share of renewable energies. In 2008, electricity generation accounted for 17.2% of global final energy consumption and only 13.5% of those were produced by nuclear power. Thus, the share of nuclear electricity is less than 2.3% of global final energy consumption. France has the highest share of nuclear electricity worldwide but France’s nuclear program does not come close to ensuring its energy independence. While nuclear energy in 2008 provided 77.1% of France’s electricity, this corresponds to only 17.3% of the total final energy that consumers use (cf. Table  7.1). Oil meets almost half, and fossil fuels over 70%, of France’s final energy needs, as is the case in many other countries. Moreover, all of France’s uranium is imported. Even if the French model would be reproduced in other parts of the world, the contribution of nuclear power would be limited. And its contribution to slowing down global warming can be questioned. For several reasons, the share of nuclear power cannot increase significantly in the mid‐ or long‐term perspective. About six times as many NPPs would be needed to be installed around the world than exist today,

Nuclear and Renewables: Compatible or Contradicting?  123

which would be an order of 2100 large NPPs with a capacity of 1000 MW. In a study, the International Energy Agency (IEA) assumed investments of 45 trillion US dollars necessary to stop climate change. Besides the construction of 1400 NPPs, a massive expansion in wind turbines has to be financed[14]. The German energy expert Klaus Traube[15] gives cause to serious concern that new NPPs would have to be built in all regions of the world. The danger of terrorist attacks on NPPs would rather increase than decrease. Finally, the world’s uranium reserves are limited and currently about 70 000 t of uranium per year are required to operate the existing NPPs. A massive expansion of nuclear power would need an additional 260 000 t of uranium per year. As a consequence, the currently known reserves of uranium would be spent not after 70 years, as is currently assumed, but rather after only 18 years. The carbon footprint of NPPs also disfavors the nuclear expansion. Greenhouse gas emissions cannot be calculated solely on the basis of actual operation but have to take the entire production cycle into account. Then it quickly becomes clear that NPPs are not CO2‐free production systems, as presented in the advertisements of the operating companies. Already, nuclear power emits up to one third of CO2 as traditional gas‐fired power plants. Production‐related CO2 emissions of nuclear energy – depending on where the raw uranium and the fuel rods are produced – add up to 126 g of CO2 per kilowatt‐hour (all data see Ref.[16]). Fritsche[17] has calculated the carbon footprint for a typical NPP in Germany with 32 g of CO2 equivalents. By comparison, a natural gas cogeneration plant emits 49 g, an import‐coal CHP plant 622 g and a traditional lignite power plant is responsible for 1153 g kWh−1 CO2 equivalents. Much better is the carbon footprint of biogas total energy units with minus 409 g kWh−1 and also the carbon footprint of wind power with only 24 g kWh−1. The high expectations for nuclear power have ­therefore  – from a few NPP operators and nuclear power advocates, however, apart  –  given way to the insight that nuclear power is a phase‐out model, because there are always new risks shown (see Ref.[18]). ­Unresolved problems remain: the final disposal of radioactive waste, the limited uranium reserves, and the fast breeder technology (FBRs should produce plutonium as new nuclear fuel and expand the range of uranium resources by factor 60) has hitherto failed. It can be argued that swiftly shutting down NPPs is necessary to put more pressure on operators and the utility industry to innovate, to develop sustainable and more socially acceptable energy technologies, and, above all, to deploy them.

­RENEWABLE ENERGIES The era of large power plant units and condensing power plants will be finished by small, economical, and distributed systems as well as through renewable energy technologies[5]. The development of a sustainable energy system could restrengthen the link between the electric industry and utilities, and create innovations and new electrical engineering. This development can cause a potential of growth for renewable energies and could provide a faster access. The target is to replace fossil energies (coal, oil, and gas) and nuclear power by an environmentally friendly renewable energy system on a mid‐ and long‐ term basis, in order to actually and clearly decrease greenhouse gas emissions. Further technology and transfer of know‐how to newly industrializing and developing countries has to be intensified, so that the possibilities of the development of renewable energies can be used globally. Furthermore, renewable energies also have the potential to develop the self‐sufficiency of developing countries and to reduce their dependence on imported energy. Renewable energy sources are natural resources such as sunlight, wind, rain, tides, and geothermal heat, which may be naturally replenished. Renewable energy technologies range from solar power, wind power, hydroelectricity, micro hydro, biomass, and biofuels for transportation. Traditional biomass, such as woodburning, is still the most important source. Hydropower is the next largest renewable source, followed by hot‐water heating. But the share of modern solar, wind, and tidal energy to the renewable energies is still marginal. The technical potential for the use of renewables is very large, exceeding all other readily available sources. The use of renewable energy resources is on the rise worldwide. In 2010, all renewables excluding large hydro received $151 billion of global private investment and surpassed nuclear power’s total global installed capacity. Within a few years, new renewables will exceed the global nuclear power generation. “Just new solar power that is buildable sooner than one new reactor would outproduce and outcompete all 64 reactors that are currently under construction”[18]. In 2009, governments stepped up efforts to steer their countries out of recession by transforming industries and creating jobs. This gave a boost to the renewable‐energy sector. By early 2010, more than 100 countries had some type of policy target and/or promotion policy related to renewable energy; this compares with 55 countries in early 2005. Wind power and solar photovoltaics (PV) additions reached a record high during 2009, and in both Europe and the United

124  Advances in Energy Systems

States, renewables accounted for over half of newly installed power capacity in 2009. More than $150 billion was invested in new renewable‐energy capacity and manufacturing plants – up from just $30 billion in 2004. For the second year in a row, more money was invested in new renewable energy capacity than in new fossil fuel capacity[19]. An important factor in the pressure for change is the appearance of new actors on the energy industry stage. This tends to change the previously dominant role played by the electric power industry in the field of electricity policy. This has once again occurred most effectively in the United States, where consumer, environmental, and new supplier interests have occupied strong positions in energy supervisory institutions and advisory councils. The small, relatively homogeneous group of decision makers, especially in the field of electricity policy, has been expanded to include not least of all representatives of local authority interests and those of new energy interests. In the domain of energy research, too, the capacity and professionalism of new approaches has risen steadily. The claim of traditional electrical power suppliers to speak on behalf of research as a whole has been invalidated. In 2011, the World Wide Fund for Nature (WWF) published an energy report that explores powering the world entirely by renewable energy by the middle of this century. The report is a science‐based examination of a renewable and clean energy future on a global scale. It covers all energy needs and the challenge of providing reliable and safe energy to all. The report is based on conservative assumptions such as price of fossil fuel increases no more than 2% annually, deployment of already available technologies, and continuous renewable energy expansion[20].

the EU remains critical toward nuclear power and has a strong preference for other energy forms. The nuclear industry has failed to deliver in the past because of large budget overruns, construction delays, and excessive overall lead times. Much of this had to be covered either by the electricity consumers or by government credit guarantees, i.e. taxpayers. Problems with recent newly build projects indicate that there is no change to be expected. Nuclear energy will hinder rather than favor reliable, sustainable energy policies. Another important factor is the trend toward a transition from subsidizing the energy industry to taxing it. There is a growing international trend toward dismantling the considerable financial privileges granted to the traditional energy industry. The substantial support, in particular, for coal and for nuclear energy in many countries should be called to mind. Today, an opposing trend toward taxing energy has emerged. The long‐term sustainable energy solution can only be based on renewable energies, which guarantee protection of natural resources, reduction of health risks, and reduction of dependence on energy imports. Renewables stand for the promotion of economic development, a low level of automation. They create installation jobs and services locally, and offer environmental benefits – reduced CO2 emissions and primary energy savings. Renewables give access to modern, local, and reliable energy, and added value in rural areas. But renewables only provide solutions if implemented together with energy demand reduction. A sustainable and intelligent energy future will only be possible if, finally, energy efficiency and demand side management will go from decade‐long rhetoric to radical implementation.

REFERENCES ­CONCLUSION The coming reformation of the energy supply systems means that the domination of the 1000 MW fossil or nuclear steam turbine plant is over and that the age of small advanced gas turbines, decentralized production (cogeneration), and much more efficient electricity‐ using lights, motors, and appliances has dawned. Future electricity systems will probably be dominated by very flexible, decentralized technologies such as the fuel cell. Nuclear power plays, in the global energy supply, a very limited role, and it is highly likely that it will further decline. The nuclear industry has a long‐term workforce problem and will struggle to maintain competence levels for existing facilities. The nuclear industry is not trusted by the public. Public ­opinion in

1. Meadows, D., Meadows, D.L., Randers, J. et al. (1972). The Limits to Growth; A Report for the Club of Rome’s Project on the Predicament of Mankind. New York: Universe Books. 2. Meadows, D.H., Meadows, D.L., and Randers, J. (1992). Beyond the Limits: Confronting Global Collapse, Envisioning a Sustainable Future. White River Junction, VT: Chelsea Green Publishing. 3. Turner G; 2008. A comparison of the Limits to Growth with thirty years of reality. CSIRO Working Paper Series. Canberra, Australia. Available at: http://www.csiro.au/ files/files/plje.pdf. (Accessed April 7, 2011). 4. Mez, L. and Brunnengraber, A. (2011). On the way to the future—renewable energies. In: After Cancún: Climate Governance or Climate Conflicts (ed. E. Altvater and A. Brunnengraber), 173–189. Wiesbaden: VS Verlag. 5. Lönnroth, M. (1989). The coming reformation of the electric utility industry. In: Electricity: Efficient End‐Use

Nuclear and Renewables: Compatible or Contradicting?  125

and New Generation Technologies, and Their Planning Implications (ed. T.B. Johansson, B. Bodlund and R.H. Williams), 765–786. Lund: Lund University Press. 6. Amery, C. and Klimawechsel, S.H. (2001). Von der fossilen zur solaren Kultur. Munich: Kunstmann. 7. SRU German Advisory Council for the Environment; 2011. Special Report. In: Pathways towards a 100% Renewable Electricity System. Berlin: Erich‐Schmidt‐ Verlag. Ch. 10. Available at: http://www.umweltrat.de/ SharedDocs/Downloads/EN/02_Special_Reports/ 2011_01Pathways_Chapter10_ProvisionalTranslation. pdf. (Accessed May 5, 2011). 8. SRU Sachverstandigenrat für Umweltfragen Weichenstellung fUr eine nachhaltige Stromversorgung. Berlin: Thesenpapier; 2009. Available at: http://www. umweltrat.de/SharedDocs/Downloads/DE/06_Hinter‐ grundinformationen/2009_Thesen_Weichenstellungen_ Stromversorgung_Hohmeyer.pdf?blob=publicationFile. (Accessed May 16, 2012). 9. SRU Sachverständigenrat für Umweltfragen 100% erneuerbarer Stromversorgung bis 2050: Klimaverträglich, sicher, bezahlbar. Berlin; 2010. Available at: http://www. umweltrat.de/SharedDocs/Downloads/DE/04_ Stellungnahmen/2010_05_Stellung_15_erneuerbare Stromversorgung.pdf?__blob=publicationFile. (Accessed May 16, 2012). 10. New York Times. President offers plans for revival of nuclear power, October 9, 1981. Available at: http://www. nytimes.com/1981/10/09/us/president‐offers‐plans‐for‐ revival‐of‐nuclear‐power.html. (Accessed May 16, 2012). 11. Schneider M, Froggatt A, Thomas S. The World Nuclear Industry Status Report 2010/11. Paris/Berlin/ Washington. Available at: http://www.greens‐efa.eu/ fileadmin/dam/Documents/world_nuclear_industry_ [email protected]. (Accessed May 5, 2011). 12. TVA Tennessee Valley Authority. Watts bar nuclear plant. Available at: http://www.tva.gov/power/nuclear/ wattsbar.htm. (Accessed November 6, 2011). 13. International Energy Agency (2010). 2010 Key World Energy Statistics. Paris: International Energy Agency.

14. International Energy Agency (2008). Energy Technology Perspectives 2008—Scenarios and Strategies to 2050. Paris: International Energy Agency. 15. Traube, K. (2005). Atomenergie—unverantwortliche Bedrohung, marginale Potenziale. In: SPD, Atomausstieg—innovativ, nachhaltig, sicher, sozial, zukunftsweisend, 12–22. Berlin: SPD‐Bundestagsfraktion. 16. Fritsche, U. (2007). Treibhausgasemissionen und Vermeidungskosten der nuklearen, fossilen und erneuerbaren Strombereitstellung. Darmstadt: Oko‐Institut Arbeitspapier. 17. Schneider M., Kastchiev G, Kromp W, et  al. 2007. Residual Risk—An Account of Events in Nuclear Power Plants Since the Chernobyl Accident in 1986. Brussels. Available at http://www.sortirdunucleaire.org/actualites/ dossiers/risque/[email protected] (Accessed 16 May 2012). 18. Lovins, A. (2011). Foreword. In: The World Nuclear Industry Status Report 2010/11. Paris/Berlin/ Washington (ed. M. Schneider, A. Froggatt and S. Thomas), 4–6. 19. REN21. Renewables 2010 Global Status Report. Available at: http://www.ren21.net/Portals/97/documents/ GSR/REN21_GSR_2010_full_revised%20Sept2010. pdf. (Accessed April 7, 2011). 20. WWF. The energy report. In: 100% Renewable Energy by 2050. Available at: http://www.wwf.at/de/energy‐ report. (Accessed May 5, 2011).

FURTHER READING Mez, L. (ed.) (2007). Green Power Markets: Support Schemes, Case Studies and Perspectives. Brentwood: Multi‐Science Publishing. Mez, L., Schneider, M., and Thomas, S. (eds.) (2009). International Perspectives on Energy Policy and the Role of Nuclear Power. Brentwood, UK: Multi‐Science Publishing.

PART II

PERSPECTIVES ON GRIDS

8

Smart‐Grid Policies: An International Review Marilyn A. Brown and Shan Zhou School of Public Policy, Georgia Institute of Technology, Atlanta, GA, USA

The electric power systems of many industrialized nations are challenged by the need to accommodate distributed renewable generation, increasing demands of a digital society, growing threats to infrastructure security, and concerns over global climate disruption. The “smart grid”—with a two‐way flow of electricity and information between utilities and consumers  –  can help address these challenges, but various financial, regulatory, and technical obstacles hinder its rapid deployment. An overview of experiences with smart‐grid policies in pioneering countries shows that many governments have designed interventions to overcome these barriers and to facilitate grid modernization. Smart‐grid policies include a new generation of regulations and finance models such as regulatory targets, requirements for data security and privacy, renewable energy credits, and various interconnection tariffs and utility subsidies. ­INTRODUCTION The electric grid in most industrialized countries was designed to deliver electricity from large power plants via a high‐voltage network to local electric

distribution systems that serve individual consumers. Both electricity and information flow predominantly in one direction, from generation and transmission to distribution systems and consumers. One of the original rationales for this system design was the assumption that electricity production and supply is a natural monopoly, where a single firm can produce the total market output at a lower cost than a collection of competing firms. With the advancement of technology, the increasing demands of a digital society, growing threats to infrastructure security, and concern over global climate disruption, the current electricity infrastructure has become a constraint to progress. The result is a growing awareness of the need for electricity infrastructure modernization and the virtues of a “smart grid.” Smart‐grid architectures can integrate a diverse set of electricity resources, including large power plants as well as distributed renewable resources, electric energy storage, demand response, and electric vehicles. Figure  8.1 portrays a complex smart‐grid system with both central and regional controllers managing the two‐way flow of electricity and information ­between utilities and consumers. The

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

130  Advances in Energy Systems

Industrial plant with combined heat and power Commercial buildings

Isolated microgrid

Wind farm Energy storage Transmission substation

Demand response

Regional controller

Central controller Regional controller

Conventional power plant Distributed generation

Distribution substation

Residential buildings Solar powered charging station for EVs

Solar panels Electric vehicles

Figure 8.1  Smart grid – a vision for the future.

actual mix of controls and technologies will depend upon a region’s transmission and distribution system, its electricity governance and business model, the nature of the customers being served, and other demand‐side issues. By implementing a smart grid, electric systems can operate at higher levels of power quality and system security[1]. The efficiency of power delivery can be promoted by dynamic pricing and smart meters that enable consumers to play an active role in managing their demand for electricity. Payment systems can be made more efficient with digital communications and can reduce nontechnical losses that undermine grid economics in many developing countries. Without the development of the smart grid, the full value of individual technologies such as distributed solar photovoltaics (PVs), electric cars, demand‐side management, and large central station renewables such as wind and solar farms will not be fully realized. Despite their numerous benefits, various obstacles hinder smart grids from gaining rapid and widespread market share. A wide array of policies have emerged

worldwide to overcome these obstacles and protect the public’s interest in affordable, dependable, and clean electric power by promoting the deployment of smart grids. This chapter begins by providing an overview of barriers that hinder smart‐grid deployment and the drivers that motivate it. We then review experiences with smart‐grid policies in the United States, at both the federal and state levels. In particular, activities of four states (California, Georgia, New York, and Texas) are examined in detail. This chapter also provides insights into European Union (EU) smart‐ grid policies, with a special focus on Great Britain and Italy. To illustrate the smart‐grid policies used in other hemispheres, we also describe policy initiatives in China, South Korea, and Japan; and we discuss the unique value proposition for the smart grid in nations with substantial electricity poverty. Acknowledging that the transition to a smart grid is only beginning, this chapter ends with a brief discussion of lessons learned and recommended future directions.

Smart‐Grid Policies: An International Review  131

­BARRIERS AND DRIVERS IMPACTING THE DEPLOYMENT OF SMART GRIDS Although many smart‐grid technologies are available today, their widespread deployment is limited. To be effective, policies must address the key barriers that hinder deployment. Policies should also be designed to leverage the drivers that promote investments in smart‐grid technologies.

Smart‐Grid Barriers Access to Capital Large upfront cost and lack of access to capital is one of the greatest challenges to the deployment of smart grids[2]. Like many other green technologies, d­ eployment requires significant initial investment, yet the resulting benefits may not be fully realized for many years[3]. For example, the Electric Power Research Institute estimates that smart‐grid investments needed in the United States would cost the average residential customer $1000 to $1500; amortized over a 10‐year period, this would cause the average residential electricity bill to increase by 8–12% [4]. On the one hand, the benefits could be three to six times larger than these costs, such as lower meter reading costs, improved billing processes, reduction of nontechnical losses, enhanced reliability, improved power quality, increased national productivity, and enhanced electricity service. On the other hand, without guaranteed cost‐recovery timelines or sound business mechanisms to reduce risks for smart‐grid investment, utilities and policy makers tend to be reluctant to move toward a smart grid[2].

recovers utilities’ capital investments is still dominant in modern economies. When their profits are linked with sales, utilities have a financial incentive to maximize the throughput of electricity across their wires; hence, they are often reluctant to adopt technologies that improve the efficiency of power supply[5]. Moreover, rate‐of‐return regulation requires that utility rates are set to provide a “reasonable” return on invested capital, and any added investments must be demonstrated to be cost‐effective. As many societal benefits associated with smart grids are not fully ­rewarded by regulators, utilities that bear all the cost of smart‐grid investments have little incentive to invest in these technologies. From a consumer’s perspective, electricity rates do not typically reflect the marginal cost of electricity production or the conditions of the wholesale electricity market[6]. Without dynamic pricing that reflects the time‐dependent cost of electricity generation, customers who only receive an end‐of‐the‐month bill tend not to be interested in smart‐grid technologies or end‐use efficiency[7]. Under current policy schemes, smart‐grid technologies face disadvantages when competing with conventionally regulated power systems. To ensure system reliability, utilities and regulators often impose strict and discriminating rules on interconnection and DG. Incumbent electricity providers and distribution (and transmission) companies have incentives to discourage the deployment of smart grids in light of its potential to increase competition in the electricity market. A lack of consistency among policies at d­ ifferent levels of governments, together with outdated codes and standards has also prevented effective collaboration and integration across regions[6].

Technical Risks Network operators tend to be conservative and risk averse. Widespread and prolonged blackouts are costly and can threaten political stability in some nation states. The high‐level penetration of distributed generation (DG) on existing infrastructure can threaten system stability[2]. Developing complex integrated systems also places demanding requirements on a wide range of technologies, especially advanced metering infrastructure (AMI) and cost‐effective energy storage systems[2]. At the same time, AMI and storage systems are evolving rapidly, which introduce performance risks.

Incomplete and Imperfect Information Many consumers and investors do not see the benefits of a smart grid, nor do they understand the social and economic costs associated with today’s outdated power system[6]. The fear of carcinogenic effects from radio frequency waves has created negative public opinion about the safety of AMI, demand–response end‐use systems, electric vehicle charging stations, and other smart‐grid technologies[8]. Utilities and policy makers could play important roles in defining and communicating any health and safety effects and the benefits of smart‐grid systems to various stakeholders[3].

Regulation and Monopoly Structure Although electricity market reforms have been pursued in many countries, the utility business model based upon a negotiated rate‐of‐return that adequately

Privacy and Security Concerns Many technologies that enable the deployment of smart grids, such as smart meters and sensors, can

132  Advances in Energy Systems

increase the vulnerability of the grid to cyberattacks[3]. As the number of participants and distributed generators in the electric system increases, so does the complexity of security issues[6]. The tension between protection of consumer privacy and development of smart grid also imposes challenges on privacy protection rules. It is essential for both customers and smart‐grid service providers to have access to energy consumption data to optimize the use of smart‐grid technologies. This can be difficult when incumbent utilities that are currently controlling the meters and data on electricity consumption create barriers to market entry for new smart‐grid players[7].

that are largely based on technologies developed in the early twentieth century and that operate following a business‐as‐usual (BAU) approach will likely become increasingly vulnerable and fail to meet challenges associated with the new demands in the future. The huge economic and social losses caused by supply failures have stimulated efforts to enhance the reliability of electricity supply. Smart‐ grid technologies such as phasor measurement units allow utilities to monitor the grid system based on real‐time information, and prevent widespread electric service interruption by shedding loads and redispatching power.

Smart‐Grid Drivers

Climate Change and Clean Air Concerns

Over the past few decades, electricity markets and technologies have experienced rapid growth and development, with increasing focus on reliability. The desire for cleaner air through renewable resources and for oil independence through electric vehicles also motivates interest in smart grids.

Energy‐related human activities are a major emission source of greenhouse gasses and air pollutants. As in most industrialized countries, the electric power and transportation sectors in the United States are the largest carbon emission sources, accounting for 40% and 34%, respectively, of US total emissions in 2010[13]. Many countries have set targets for low‐carbon and renewable electricity generation to combat climate change, which require extensive changes to the current power systems. Smart grids could help to more fully exploit the potential of carbon emissions reduction and air quality improvement in energy sectors, as it enables low‐carbon distributed power generation and transport systems.

Increasing Electricity Demand Global electricity demand is expected to increase by over 150% between 2007 and 2050 under the International Energy Agency (IEA)’s 2010 baseline scenario[9]. Due to the rapid development of home appliances and a lack of real‐time pricing (RTP) signals, peak demand is expected to increase steadily over time. Since 1982, growth in peak demand for electricity in the United States has exceeded the growth of transmission system infrastructure by almost 25% every year[10], with an expected average annual growth rate of 1.7% between 2009 and 2019[11]. Rising peak demand stresses the electricity system and requires higher reserve margins for unforeseeable outages. Smart‐ grid technologies can help reduce demand by enabling demand‐side management programs, and can improve the efficiency of electricity supply through better integration of renewable DG. Energy Price Escalation and Electricity Reliability Concerns Under EIA’s (US Energy Information Administration) reference and high oil price scenarios, world oil prices are forecast to increase from $59 per barrel in 2009 to $135 and $210 per barrel, respectively, in 2035[12]. Rising petroleum prices have underscored the uncertainties associated with the long‐ term electricity market. Meanwhile, power systems

Deployment of Renewable Power and Electric Vehicles Efforts to combat climate change have encouraged the rapid development of environmentally friendly power generation and transportation technologies. As of 2008, 19% of world electricity was generated by renewable energy, and forecasts suggest a rise to 23% by 2035, with an annual growth rate of 3%[14]. Most of the new renewables is expected to come from variable and difficult‐to‐predict wind generation[15]. The transport sector is also undergoing an electrification revolution, which is expected to consume 10% of total electricity by 2050[9]. As electric vehicles gain market share, it may become difficult for conventional grid infrastructures to provide reliable and stable electricity services[9]. In particular, the intermittency of renewable energy and electric vehicle charging have to be managed intelligently to avoid supply failures, which provide an excellent opportunity for the deployment of smart grids.

Smart‐Grid Policies: An International Review  133

Economic Development and Business Opportunities Along with the development and diffusion of smart‐ grid technologies comes the growth of key industries, such as electric vehicle, smart appliance, and smart meter manufacturers. The savings businesses obtained through the adoption of smart‐grid technologies could also be redirected to other business investments, and hence improve their competitiveness in both domestic and international markets. It is clear that countries pioneering in smart‐grid deployment are building a competitive advantage for their future economy.

­ MART‐GRID POLICIES OF THE UNITED S STATES The United States aspires to a low‐carbon economy, but its current energy system is carbon intensive. The United States is second only to China in total energy‐related CO2 emissions  –  at 5610 million metric tons (Mt) of CO2 in 2010[16]. On a per capita basis, the United States is also highly carbon intensive  –  averaging 18.1 metric tons per person in 2010. Its CO2 emissions are down from a peak of 6016 metric tons in 2007 and from 19.9 metric tons per capita in the same year, just preceding the 2008 economic downturn[17]. In his 2011 State of the Union address, President Obama proposed a goal of generating 80% of the nation’s electricity from clean energy sources by 2035; however, only 11% of its electricity currently comes from renewable sources, compared with 27% in Italy and 19% in China. (See Table 8.7 for a list of data sources and a comparison with the other pioneering smart countries examined here.) Given the president’s clean energy imperative, the government recognizes that a smarter, modernized, and expanded electric system is essential to America’s world leadership in a clean‐energy future[18]. Development of policies has occurred at both federal and state levels to facilitate the evolution toward a twenty‐first‐century grid. Smart‐Grid Legislation and Policy Context The Energy Policy Act of 2005 was the first federal law that specifically promotes the development of smart meters. It directed utility regulators to consider time‐based pricing and other forms of demand response for their states. Utilities are required to provide each customer a time‐based rate schedule and a time‐based meter upon customer request. The Energy Independence and Security Act (EISA) of 2007 was the key legislation for modernizing the nation’s electricity transmission and

d­ istribution system. It authorized the Department of Energy (DOE) to establish the Federal Smart Grid Task Force as the main platform to implement and coordinate national smart‐grid policies. The DOE is now also required to establish smart‐grid technology research, development, and demonstration projects to leverage existing smart‐grid deployments. The National Institute of Standards and Technology (NIST), a major standards developing federal agency, is directed to develop a smart‐grid interoperability framework that provides protocols and standards for smart‐grid technologies. EISA also established a federal smart‐grid investment matching grant program to reimburse 20% of qualifying smart‐grid investments. The next important legislative effort was the American Recovery and Reinvestment Act of 2009. It accelerates the development of smart‐grid technologies by appropriating $4.5 billion for electricity delivery and energy reliability modernization efforts. Utilities and other investors can apply stimulus grants to pay up to 50% of the qualifying smart‐grid investments. As of 2011, the Smart Grid Investment Grant authorized under this Act had 99 recipients, with a total public investment of $3.5 billion[19]. State and Local Efforts Building on the policy directions set by federal legislation, state and local activities also form an important part of the nation’s overall grid modernization efforts. The scope and pace of smart‐grid deployments naturally vary according to the diverse needs, regulatory environments, energy resources, and legacy systems of different states. Decentralized policy efforts provide local flexibility and stimulate experimentation and innovation in policy design and implementation; thus, it is useful to examine smart‐grid policies developed at the state and local level[20]. In this section, four US states are selected for in‐depth investigation: California, Georgia, New York, and Texas. These states have a wide range of carbon footprints, from 9 and 10 metric tons of CO2 per capita in New York and California, respectively, to 17 and 24 metric tons in Georgia and Texas. The percent of their electricity generation coming from renewable resources ranks similarly, with only 5% and 7% renewables in Georgia and Texas, but fully 22% and 29% in New York and California[21]. The results show that four types of policies are widely implemented in these pioneering states: net metering policies, ­interconnection standards and rules, smart metering targets, and dynamic pricing policies.

134  Advances in Energy Systems

Net Metering Policies Net metering allows customers to use a single meter to measure both the inflow and outflow of electricity, thus enabling them to install and interconnect their own generators with utility grids. With net metering, customers can use the electricity generated from their on‐site facilities to offset their electricity consumption and sell excess generation to the utility typically at a retail price, thereby encouraging the deployment of customer‐owned distributed energy systems. By allowing utilities to buy back surplus electricity, net metering helps overcome financial barriers faced by distributed renewable facility owners. The buy‐ back price, together with the cumulative generating capacity is determined by utility regulators; therefore, they often differ across regions (see Table 8.1). Eligibility criteria are commonly defined by sectors (e.g. residential, commercial, and industrial), types of renewable resources (e.g. solar, wind, and combined heat and power [CHP]), and generating capacity (e.g. less than 10 kW or up to 1 MW). Net metering rules are often updated by policy makers to meet the needs and priorities of the market. In general, the trend is

to increase the system capacity cap, as in the cases in New York and Massachusetts and to broaden the ­eligible renewable resources[22].

Interconnection Standards and Rules Interconnection standards establish uniform processes and technical requirements for utilities when connecting DG systems to the electric grid. It allows DG developers to predict costs and time, and ensure the safety and reliability of interconnection processes. Technical requirements often include ­ protocols and standards that guide how generators interconnect with the grid, ranging from system capacity limits and qualifying generators, to the types of interconnection equipment required for ­reliability purposes. Interconnection policies can also specify connection and operation procedures, which can reduce uncertainties and prevent time delays for approving grid connections. Interconnection ­standards are often available to certain generation facilities, depending on their generating capacity, sector, and technology type (see Table 8.2).

Table 8.1  Net metering policies in four US states. Qualifying facilities Eligible technologies

System capacity limit

CA Solar, wind, biogas, fuel cell 1 MW TX Renewable energy sources 50 kW GA PV, wind, fuel cell 10 kW (residential); 100 kW (nonresidential) NY Wind, solar, fuel cell, micro‐ 10 kW–2 MW CHP, micro‐hydroelectric, farm waste

Cumulative generating capacity (% of utility’s aggregate peak demand)

Buy‐back rate

5% No limit 0.2%

Retail rate Avoided energy cost A predetermined rate

0.3% (wind); Retail rate (wind, solar, farm 1% (solar, biogas, micro‐CHP and waste); fuel cell combined) avoided energy cost (micro‐ CHP, fuel cell)

Table 8.2  Interconnection standards and rules in four US states.

CA

TX

GA

NY

Main provisions

Targeted systems

• Standard interconnection, operating, and metering requirements • Application and evaluation procedures, fees, and costs • Requirements for generators and network interconnection of distributed generation • Requirements for control, protection, and safety equipment • Customers required to meet applicable interconnection requirements, such as the National Electrical Code and the National Electrical Safety Code • Interconnection procedures • Requirements for the design and operation of distributed generation facilities • Application procedures, fees, and maximum expenses

Facilities to be connected to utility distribution systems Facilities with capacity ≤10 MW and connection voltage ≤60 kV Residential (≤10 kW) and Commercial (≤100 kW) facilities that use photovoltaic, wind, and fuel cells Facilities ≤25 kW; 25 kW ≤ facilities ≤2 MW

Smart‐Grid Policies: An International Review  135

Smart Metering Targets A smart meter reader is a device that can measure real‐time electricity consumption and communicate the information back to utilities. A smart meter, on the other hand, communicates back to both the utility and the consumers. Smart metering targets typically establish smart meter reader deployment plans for utilities, covering the timeline, and the type and number of smart meters to be installed. Sometimes, utilities are required to conduct cost–benefit analysis (CBA) of the proposed smart metering programs. Many states have set smart metering targets to be implemented by utilities (see Table 8.3). Georgia, in particular, aims to install smart meter readers for every consumer by 2012. Dynamic Pricing Policies Dynamic pricing is a market‐driven approach to boost demand response in electricity markets. The fundamental idea is to provide accurate price signals to customers, and let them decide whether to continue

consumption at higher prices or to cut electricity usage during peak times. It is currently available in many sectors, and is most widely used in commercial and industrial sectors (see Table  8.4). Under dynamic pricing schemes, utilities charge different rates for electricity based on time, generating cost, and conditions of the grid; hence, customers are exposed to some level of electricity prices variability The most common dynamic pricing policies include time‐of‐use pricing (TOU), critical peak pricing (CPP), and RTP. • TOU sets and publishes electricity prices for different time periods in advance. Electricity prices in peak periods are higher than off‐peak, which encourages customers to shift electricity consumption to a lower cost period and reduce the peak demand. The rates for each time block are usually adjusted two or three times each year to reflect changes in the wholesale market; however, TOU pricing does not address unforeseen weather c­ onditions or equipment failures.

Table 8.3  Smart metering targets in four US states. Targets Utility/agency CA Pacific Gas and Electric San Diego Gas and Electric Southern California Edison TX Center Point Oncer AEP Texas GA Georgia Power NY Public Service Commission

Deployment timeline

Electric meters (million)

Gas meters (million)

2007–2011 2007–2011

5.1 1.4

4.2 0.9

2008–2012

5.3

–

2009–2012 2008–2012 2009–2013

2.1 3.4 1.1

– – –

2008–2012 2006

Every customer will be provided with a free smart meter reader. Utilities must file proposals for integrating smart meter readers into their systems.

Table 8.4  Dynamic pricing policies in four US states. Targeting systems Types of rates CA

TX GA NY

CPP RTP TOU TOU TOU RTP TOU RTP

Residential sector √ √ √ √

Commercial and industrial sectors

Agricultural sector

Electric vehicles

√ √ √

√ √ √

√ √

√ √

√

√

√ √

CPP, critical peak pricing; RTP, real‐time pricing; TOU, time‐of‐use pricing.

136  Advances in Energy Systems

• CPP is similar in rate structure to TOU pricing, but it adds one more rate that can vary with the wholesale market. Electricity prices during a limited number of hours of the year, which refer to the “critical peak hours,” rise to levels designed to recover the full generation cost, whereas electricity prices during other times are lower than the critical periods. There can be a number of CPP event days in a year, and utilities usually will notify customers of the events and rates ahead of time. • RTP reflects the hourly or an even smaller time‐ interval marginal cost of electricity, which can be announced at the beginning of the time period or in advance. RTP can capture most of the true variation in the wholesale market, but it gives customers little time to react to price changes[23]. Technology innovations of the last decade have enhanced customers’ ability to respond to real‐time prices, and eliminated the conflicting issues between increased advanced price notification and more accurate price signals, enabling the greater use of RTP[23]. Besides the four types of policies described above, access to real‐time metered data is illustrative of the new issues requiring public regulation. US states are beginning to set requirements regarding data security and privacy of smart meters. Texas, for example, has determined that all meter data, including data used to calculate charges for service, historical load data, and any other proprietary customer information, will belong to a customer; however, customers can allow retail electric providers to access the data under rules and charges established by the Public Utility Commission of Texas. The ownership of renewable energy credits (RECs) from customer‐owned renewable facilities is another issue that is only now being clarified. The issue is important because RECs have significant economic value, and clear rules and regulations regarding their ownership could help reduce confusion and uncertainties associated with smart‐grid investment. This policy issue is also contentious as it involves the design and consideration of several policy regimes, including renewable electricity standards, net metering, interconnection policies, and utility subsidies for renewable projects. The four case studies show that the goals and design of smart‐grid policies are highly variable across states. Although most US states have net metering and interconnection standards, the s­ pecifics of these policies vary widely (e.g. eligible technologies and customers, application and evaluation procedures). States also are in different stages of smart‐meter deployment; however, there is a growing

consensus that smart meters are an essential enabler of grid modernization. Numerous different types of dynamic pricing rates have emerged over the past decade, typically starting with large industrial customers, followed by commercial and large nonresidential customers. Research has shown that dynamic pricing can not only remove subsidies embodied in flat rates but also reduce peak demand[24, 24]. Variability among dynamic pricing rates also reflects the differences in policy goals of cost recovery and demand response programs. Despite this wide‐ranging policy variability, some policy principles are emerging. Cost allocation rules need to ensure the recovery of smart‐grid costs and to facilitate investment in new smart‐grid infrastructures. CBA and evaluation metrics are also becoming essential, and some government agencies are beginning to require the collection of such information.

­SMART‐GRID POLICIES OF THE EUROPEAN UNION The EU is the second largest energy market in the world, with over 450 million customers[26]. The objective of EU energy policies in the twenty‐first century is to achieve a sustainable, competitive and secure energy supply[27]. The deployment of smart grids is an essential part of the EU’s climate change and clean energy initiatives, as it can transform traditional electricity markets and networks. The breadth of EU smart‐grid policies is illustrated in Table 8.5. Smart‐Grid Policies in Italy Italy emitted 416 Mt CO2 in 2010, a 2% increase above 2009 but lower than its peak of 467 in 2006. Its carbon intensity (at 7.0 metric tons of CO2 per capita) exceeds that of China but is lower than the other four countries examined here[28]. Its modest carbon footprint is achieved in part by its significant investment in renewable power, which represents 27% of its total electricity generation. Modernization and expansion of the electricity transmission and distribution networks has been a critical step in the successful integration of renewables in Italy’s energy system[29]. The Italian government plans to increase the country’s renewable consumption from 5.2% of total energy consumption in 2005 to 17% in 2020[30]. Efforts at various levels of governments have been made to accelerate energy infrastructure optimization. In 2007, the European Commission approved the Operational Program  –  “Renewable Energy Sources

Smart‐Grid Policies: An International Review  137

Table 8.5  Smart Grid legislation and regulations in the European Union. Policy emphases interconnection standards Directive 2001/77/EC Directive 2003/54/EC Green Paper (2005) Green Paper (2006) Directive 2006/32/EC COM (2007) 723 final Directive 2009/72/EC Conclusions of the European Council of February 4, 2011 Commission Recommendation on Preparations for the Rollout of smart metering systems (C/2012/1342) EC standardization mandate for smart meters (M/441) EC standardization mandate for electric vehicles (M/468) EC standardization mandate for smart grids (M/490)

Smart meters

Demand response and dynamic pricing

√

√ √

√

√

√ √

Electric vehicles

√ √

and Energy Saving” in southern Italian regions (Apulia, Campania, Calabria, and Sicily) with a budget of €1.6 billion ($2.0 billion)[31]. One priority of this program is to improve the infrastructure of transmission networks to promote renewable energy and small‐/microscale cogeneration, which receives €100 million ($123 million) from European and Italian state funds[29]. Within this context, the Italian Ministry of Economic Development and Italy’s largest power company, Enel Distribuzione, together launched a €77 million ($95 million) “Smart Medium Voltage Networks” project in southern Italy to make medium voltage distribution networks more favorable to PV systems with installed capacity between 100 kW and 1 MW[32]. In addition, the Italian Regulator Authority for Electricity and Gas has awarded eight tariff‐ based financial projects on active medium voltage networks, to demonstrate at‐scale advanced network management and automation solutions necessary to integrate DG[32]. Italy has one of the largest and most extensive smart metering programs in the world. In Regulatory Order 18 December 2006 no. 292/06 and Regulatory Order 26 September 2007 no. 235/07, Italian legislators introduced mandatory installation of smart meters for all household and nonhousehold low‐ voltage customers starting in 1 January 1 2008, and minimum performance standards for the meters were also provided[33]. Italy’s smart metering deployment

√ √

√

√

√ √ √

has an emphasis on the distribution system operators, which is designed to support the liberalization of the energy market and prevent electricity theft. Enel Distribuzione is the major player in Italy’s smart meter deployment. By 2011, Enel had installed smart meters for 32 million customers in its electrical distribution system and provided advanced services enabled by smart meters, such as the hourly based tariff system[34]. The intent was to install smart meters in its gas distribution grid and extend the smart metering system to its distribution grids in Spain, installing 13 million smart meters between 2010 and 2015[34]. Enel also launched the E‐mobility Italy program in three Italian cities (Rome, Milan, and Pisa) in 2008[35], designed to deliver 100 electric vehicles to selected drivers in the three cities with 400 intelligent electric vehicle recharging stations. To encourage renewable DG, the Italian government guarantees priority access of electricity generated from renewable to the grid, and provides feed‐in tariffs to solar PVs. The Fourth Conto Energia (feedin tariff) approved by the Ministry for Economic Development in 2011, provided a differentiated incentive system for solar PV, including a specific expense budget designed for large PV plants between 2011 and 2012, and preestablished half‐yearly expense budgets provided to all PV plants between 2013 and 2016[36]. However, no incentives will be awarded to PV plants entering into operations after 2016.

138  Advances in Energy Systems

Smart‐Grid Policies in the United Kingdom The United Kingdom was responsible for 532 Mt CO2 from energy use in 2010, down from before the 2008 economic downturn but 2% above 2009[37]. On a per‐ capita basis, the United Kingdom and Italy are similar, and both have less than half the carbon footprint of the United States. The British government sets a firm, long‐term, and legally binding framework to cut carbon emissions by at least 34% by 2020 and 80% by 2050 – below the 1990 baseline[38]. A critical part of this framework is to commit the country to generate 15% of its energy from renewables by 2020[39]. Only 7% of the United Kingdom’s current electricity generation comes from renewable resources. The British power system is owned and operated by three transmission network operators and nine distribution network operators (DNOs), which makes it difficult to achieve a countrywide smart‐grid vision. Stakeholders in the system, including generators, suppliers, traders, and customers, act according to the British Electricity Trading and Transmission Arrangements (BETTA) to ensure supply and demand are balanced at all times[40]. Integration of diverse generation from low‐carbon technologies inevitably requires transformation of the electric system. To modernize and reduce the carbon footprint of electric grids, one major initiative of the United Kingdom is to encourage energy efficiency through smart‐ meter deployment. The British government expects full penetration of smart meters by 2020, with a total financial investment of £8.6 billion ($13.5 billion) and total benefits of £14.6 billion ($22.9 billion) over the next 20 years[41]. Early government action can be traced back to the 2008 Energy Act, which allowed the secretary of state to take measures to install or facilitate the installation of smart meters[42]. Energy Bill 2010–2011 provided financial incentives to encourage smart‐meter installation by householders, private landlords, and businesses[43]. In July 2010, the Department of Energy and Climate Change (DECC) and Office of the Gas and Electricity Markets (OFGEM) published the “Smart Metering Implementation Program: Prospectus,” which sets design requirements, central communications, data management, and the rollout plan for the deployment of smart meters to all homes and small businesses in Great Britain[44]. The Government’s Response to Prospectus Consultation in 2011 required energy suppliers to provide smart meters that meet specific technical standards[45]. It also included selection and regulatory procedures for a new, licensed Data and Communications Company to manage smart metering data. With respect to privacy protection, customers will be able to choose how

their consumption data is used and by whom, except when data are required for regulatory purposes. Multiple incentives are designed to encourage innovative decarbonizing initiatives in the United Kingdom’s power system. The 2008 Energy Act introduced feed‐in‐tariffs (FITs) for low‐carbon electricity generation facilities with a generating capacity less than 5 MW[42]. Eligible technologies include anaerobic digestion, solar PVs, hydroelectric power, wind, and micro‐CHP systems. Total capacity of solar PV registered in FITs has reached 1 GW by 2012, with the FIT payment rates ranging from 8.9 to 21.0 kWh−1 ($0.14 ∼ $0.33 kWh−1)[46]. Between 2010 and 2015, OFGEM was committed to providing a £500 million ($785 million) low‐carbon network (LCN) fund to help DNOs develop trial projects of new technologies and commercial arrangements that enhance energy security and combat climate change[47]. British energy regulators believe that the up scaling of many critical components of smart grids, such as demand‐side management, DG, and electric vehicles can only be achieved through significant changes of the distribution networks. A £6 million ($9.4 million) Smart Grid Demonstration Fund is also in place to facilitate the development of smart‐grid technologies, focusing primarily on the supply chain and the regional integration of alternative energy sources[41]. The British government has designed multiple institutions and platforms to increase fundraising for smart‐grid development. A typical example is the Energy Technologies Institute, a partnership between the British government and industrial sectors. It allows for a variable mix of public and private funding to accelerate the development of low‐carbon technologies, including energy storage, building energy management, and DG[48].

­SMART‐GRID POLICIES OF EAST ASIA Smart‐Grid Policies in Japan As a country that is only 16% energy self‐sufficient, Japan is the world’s largest importer of liquefied natural gas (LNG), second largest importer of coal, and the third largest net importer of oil[49]. Its energy use was responsible for approximately 1160 Mt CO2 in 2010, and one‐third of these emissions were produced by its industrial sector. On a per capita basis, Japan is half the carbon intensity of the United States (9.2 versus 18.1 metric tons of CO2 per capita)[50]. Japan aims to reduce its carbon emissions by 30% by 2030 compared to the 1990 level, and to have 70% of its electricity generated from zero‐emission sources by 2030, whereas in 2010, only 1% of Japan’s total energy

Smart‐Grid Policies: An International Review  139

consumption was from nonhydro renewables[49, 49]. To achieve these goals, major changes must take place in the energy system. The official goal is to build “the world’s most advanced next‐generation interactive grid network,” to realize “smart grids and smart communities,” and to promote “the development, installation of smart meters and relevant energy management systems” as early as possible in the 2020s[51]. The 2011 Fukushima nuclear incident, which put the country in an unprecedented energy crisis, has greatly accelerated government’s investment in electric grid infrastructure. It is estimated that smart‐grid market value in Japan would increase from $1 billion in 2012 to $7.4 billion in 2016[52]. Japan’s power industry is dominated by 10 regional monopolies, which accounts for 85% of the country’s total installed generating capacity[49]. Tokyo Electric Power Company (TEPCO), the largest utility company in Japan serving over 28 million customers, plans to install 17 million smart meters by 2019[52]. The FIT scheme  –  “New Purchase System for Photovoltaic Electricity”—launched in 2009 is a key government incentive for renewables[53]. Surplus electricity generated from solar PVs is purchased at ¥48 kWh−1 ($0.59 kWh−1) for residential sector, and ¥24 kWh−1 ($0.30 kWh−1) for industries, businesses, and schools. The buyback prices will decrease each year based on the innovation and price trends of solar PV technologies. Japan’s Ministry of Economy, Trade, and Industry (METI) is the major government agency responsible for smart‐grid development. Its objectives are to enable further integration of renewable energy, facilitate the development of electric vehicles, including the charging infrastructure, and create new services using smart meters and information and communications technology (ICT) networks[54]. METI has implemented demonstration projects at both regional and international levels to facilitate the penetration of smart‐grid technologies, including a $73 million investment on community grid systems (Remote Island Smart Grid Project, Smart Charge Project, and Smart House Project), $1.1 billion on four smart‐grid technology pilot projects (Kansai Science City, Yokohama City, Kitakyushu City, and Toyota City), and four smart community demonstration projects located in the State of New Mexico (the United States), Hawaii (United States), Lyon (France), and Malaga (Spain)[55, 56]. There has been increasing cooperation and collaboration between Japan’s public and private sectors in smart‐grid deployment. For instance, the Japan Smart Community Alliance established by the New Energy and Industrial Technology Development Organization (NEDO) in 2010 provides a platform for the

p­articipation of a wide range of smart‐grid stakeholders[57]. The concept of “smart community,” which refers to a new, intelligent, and sustainable way of living, not only stimulates changes in the electricity market, but also motivates innovations in automobiles, telecommunications, and home appliances industries. Toshiba Corporation, Tokyo Electric Corporation, and TEPCO are also working together to launch a venture into the commercialization of smart meters[58]. Smart‐Grid Policies in South Korea South Korea imports 97% of the energy it consumes and is highly dependent on imported petroleum and LNG. Its energy system emitted 579 Mt CO2 in 2010, representing a steady increase from 484 in 2006, mirroring its economic growth. Its per capita CO2 emissions have also been on the rise. It reached 11.9 metric tons in 2010, reflecting a growing carbon intensity. Renewable energy only accounts for 1% of its electricity generation, which is the lowest among the six countries examined here[59]. Korea doubled its CO2 emissions between 1990 and 2010, the fastest‐growth among Organization for Economic Cooperation and Development (OECD) countries[60]. By 2035, its carbon emissions are expected to increase 35% from the 2002 base line, compared to less than 15% for all the OECD countries[59]. Although as a non‐Annex I Party, South Korea is not obliged to reduce its carbon emissions under the Kyoto Protocol, the Korean government sets a voluntary goal of reducing its greenhouse gas emissions by 30% below the BAU case by 2020[59]. Reducing the nation’s energy dependence and carbon intensity is one of the top priorities of the Korean government and a mandatory cap‐ and‐trade system was to be operating by 2015[61]. The electric system of South Korea is more reliable and efficient than many other developed countries[62]. Korea Electric Power Corporation (KEPCO) was created in 1961 to supply electricity to the entire economy. KEPCO is responsible for the generation, transmission, and distribution of electricity which comprises six power generation companies, four subsidiaries, and four affiliated companies[63]. The deployment of smart‐grid technologies started in 2005 when Korea launched the Power IT National Program in order to develop digital, environmentally friendly and intelligent electric power devices and systems, and advance Korean electric power and electrical industries[64]. In August 2008, President Lee Myung‐bak announced “Korea’s National Strategy for Green Growth,” which proposed a total investment of 107 trillion won ($101 billion) between 2009 and 2013[59]. The deployment of smart‐grid technologies

140  Advances in Energy Systems

was a key part of this five‐year plan. Among the 27 core green technologies listed in its national plan, more than one‐third related to the development of smart grids and smart cities. Korea’s “Smart Grid Road Map 2030” is another key step[65]. The roadmap will be implemented in five sectors: smart power grid, smart consumers, smart transportation, smart renewables, and smart electricity services. By 2030, a nationwide smart grid and 27 140 power charge stations for electric vehicles will be built; and the penetration rate of smart meters and AMI will reach 100% by 2020. In addition, Korea will have 11% of its energy from renewables, and achieve a maximum of 10% power reduction by 2030. The annual blackout time per household will be reduced from 15 minutes in 2012 to 9 minutes in 2030, and the power transmission and distribution loss rate will decrease from 3.9% in 2012 to 3.0% in 2030. A total of 27.5 trillion won ($25.85 billion) will be allocated for the technology development and infrastructure construction in this plan. As a first step to implement the Road Map, the Korean government started a pilot program on Jeju Island in June 2009, which consists of a fully integrated smart‐ grid system for 6000 households, wind farms, and four distribution lines[66]. A total of $50 million public funds and $150 million private funds were to be invested between 2009 and 2013. More than 100 companies from automobile, renewable, power, telecommunication, and home appliance industries participated in the program. KEPCO participated in all five sectors of the Jeju Island pilot program. KEPCO is also committed to develop green technologies such as export‐ ready nuclear power plants, electric vehicle charging infrastructure, integrated gasification combined cycle (IGCC), and carbon capture and storage (CCS) technologies[63]. The s­ econd stage includes the expansion into metropolitan areas. The last stage expands to the nationwide intelligent grid networks. The anticipated effect is to generate 50 000 new jobs every year and reduce a total of 230 million tons of greenhouse gases by 2030[65]. Smart‐Grid Policies in The People’s Republic of China Since the 1980s, China’s energy consumption has been growing at an unprecedented rate due to rapid economic development. Its CO2 emissions first eclipsed the United States in 2007 at 6184 Mt CO[2], and its economic growth has catapulted China’s emissions to 8321 Mt CO2 in 2010[16]. Between 1990 and 2010, China’s electricity generation increased from 621 to 4206 Terawatt‐hours (TWh)[67], with annual

growth rates of electricity demand ranging from 10% to 15%[68]. In 2010, 19% of China’s electricity generation came from renewable resources, second only to Italy among the six countries examined here. China has experienced several major power outages since 2005, and the shortfall in electricity has started to hurt China’s economy[68]. To meet the increasing demand and secure economic growth, the Chinese government will invest 286 billion yuan ($45 ­billion) in smart‐grid deployment between 2011 and 2015[69]. The country’s transition to a high‐tech and high value added manufacturing and service economy also directs enormous government support to the new energy industry and transport system. Promoting the development of clean energy and smart grids is among the top priorities of the government. The Amendment of the Renewable Energy Law (2009) urges utilities to develop and apply smart‐grid and energy storage technologies to improve grid operation and management, and facilitate interconnection of distributed renewable energy[70]. The Twelfth Five‐year Plan, a series of major social and economic initiatives, sets separate targets for energy intensity (16% reduction by 2015), non‐fossil‐fuel energy (11% of the total primary energy consumption by 2015), and carbon intensity (17% reduction below 2011 by 2015)[71]. Smart grids and clean energy technologies are seen as effective approaches to achieve these targets. New sources of electric power and vehicle propulsion are two of the seven strategic emerging industries to receive financial and regulatory support from the government. By 2015, several long‐distance ultra‐high voltage (UHV) transmission lines and 200 000 km of transmission lines (333 kV and above) will be constructed. The Plan also proposes the “Rural Electricity Supply Project” to upgrade rural electric grids and meet the increasing demand of rural areas. Some of the targets include developing 1000 PV demonstration villages, 200 green energy counties, 300 hydropower and rural electrification counties, and 10 000 MW of small hydropower. The Ministry of Science and Technology released the “Special Planning of 12th Five‐Year Plan on Smart Grid Major Science and Technology Industrialization Projects” in May 2012. It identified nine key tasks, including large‐scale grid‐connected intermittent renewable energy technology, grid technology for supporting electric vehicles, large‐scale energy storage systems, intelligent distribution technology, intelligent grid operation and control, intelligent transmission technology and equipment, grid information and communication technologies, flexible power transmission technology and equipment, and smart‐grid integrated comprehensive demonstrations[72]. Resource

Smart‐Grid Policies: An International Review  141

allocation optimization, clean energy development, power system reliability, diverse customer needs, energy efficiency improvement, and technology innovation are the major drivers for smart‐grid deployment in China[72]. State Grid Corporation of China (SGCC), the largest power company and the major smart‐grid policy implementer in China, provides services to over one billion customers and covers 88% of the national territory[73]. In May 2009, SGCC announced a plan for developing a “strong and smart grid” in China by 2020[74]. UHV transmission and highly efficient distribution transformer that enables the expansion of transmission and distribution capacity and reduces line losses are key technologies to be developed and deployed. SGCC’s smart‐grid development plan is distinct in its focus on the transmission, rather than the distribution side, due to the fact that major power generation sources in China, such as coal and hydropower are located in remote areas, and there are huge disparities among power generation in different regions. Other reasons for the focus on transmission might be the relatively primitive structure at the ­distribution ends, and the unique asset ownership and management structure of utilities and electric markets[75]. With an emphasis on power generation and transmission, the Chinese electricity market still has a long way to develop an effective interaction mechanism between customer and utility companies, such as dynamic electricity prices and demand response programs[76].

­INTERNATIONAL COLLABORATION The Smart Grids European Technology Platform was established in 2004, with an aim to enhance the level of coherence between the European, national, and regional efforts addressing smart grids. One important role of this platform was to cooperate with other countries, especially North America and Japan, to ensure international development paths for smart grids are complementary and consistent with the development of commercial products[50]. The IEA Implementing Agreement on Electricity Networks Analysis, Research and Development (ENARD) was developed by 14 IEA member countries in July 2006. Its mission is to provide comprehensive and unbiased information, data and advice to key stakeholders and policymakers of the issues relating to current and anticipated developments in electricity transmission and distribution networks[77]. Some of the work programs that are closely linked to smart grids include Annex II (DG system integration),

Annex III (infrastructure asset management), and Annex IV (transmission system issues). ENARD is currently focusing its activities within the IEA member countries; however, it is open to participation by non‐IEA member countries, private sectors, and nongovernmental organizations. Established in April 2010, the Global Smart Grid Federation (GSGF) brings together the key smart‐grid stakeholders around the world, including the United States, GridWise Alliance, Australia, Canada, Ireland, Korean Association, India, and Japan[78]. Its goals are to facilitate the collaboration of governments and nongovernmental organizations, to support the development of smart‐grid technologies, and foster knowledge sharing. The International Smart Grid Action Network (ISGAN) was launched at the first Clean Energy Ministerial in Washington, D.C. in July 2010 to accelerate the development of smart‐grid technologies at the global level. ISGAN focuses on five principal areas, including policy, standards, and regulation; finance and business models; technology and systems development; user and consumer engagement; and workforce skills and knowledge[79]. It includes four projects: the global smart‐grid inventory, smart‐grid case studies, benefit–cost analyses, and toolkits and synthesis of insights for decision makers.

­CONCLUSIONS AND RECOMMENDED FUTURE POLICY DIRECTIONS Along with the recent introduction of smart‐grid technologies has emerged a new generation of regulations and fiscal policies to ensure that the public’s interests are protected. Current smart‐grid policies address many of the barriers that hinder deployment and are aligned with many key drivers (see Table 8.6). Countries are in different stages of smart‐grid deployment and have set various targets for future development. Table  8.7 summarizes the energy and climate change targets of five nations and regions, and describes the drivers and focuses of their smart‐grid policies. Although smart‐grid policies vary across US states, most states have implemented net metering policies and interconnection standards. Many utilities are installing smart meters using funds from American Recovery and Reinvestment Act appropriations, and dynamic pricing programs are widely used in industrial and commercial sectors. Smart‐grid programs are critical components of the EU’s low‐carbon agenda. The British regulators have been very active not only in the rollout of smart meters and modernized distribution networks but also in innovative low‐carbon technologies. Japan and Korea are both

142  Advances in Energy Systems

Table 8.6 Smart‐grid policies to tackle barriers and leverage drivers. Barriers

Smart‐grid policies Net metering Interconnection standards and rules Dynamic pricing Smart metering targets Renewable energy subsidies and regulations International smart‐ grid collaboration Smart‐grid demonstration projects

Access to capital

Technical risks

× ×

× ×

×

Drivers

Regulation and Incomplete and monopoly imperfect structure information × ×

×

× ×

× ×

Privacy and security

Increasing electricity demand

Rising energy prices and reliability concerns

×

× ×

× ×

× ×

× ×

× ×

× ×

× ×

×

× ×

×

×

×

×

×

×

×

Deployment of Climate renewable power change and and electric clean air vehicles

×

×

×

×

×

×

×

×

×

×

×

×

×

×

×

×

Economic development and business opportunity

Table 8.7

Status, targets, policy drivers, and emphases by country. CO2 emissions Metric tons of CO2 per capitaa

US

EU

US: 18.1

CO2 emissions targets

17% below 2005 level by 2020 US: 11

CA—10

CA—29

GA—17

GA – 5

TX – 5880 MW by 2020

NY—9 TX—24 Italy: 7.2

NY – 22 TX—7 Italy: 27

NY – 29% by 2015 GA—none 20% by 2020

9.2

South Korea 11.9

China

Renewable electricity targets US: None. (80% clean energy by 2035) CA – 33% by 2020

20% below 1990 level by 2020

UK: 8.5

Japan

Renewable electricity Percent renewable generationb

6.3

UK: 7

30% below 1990 by 2030

10

30% below BAU by 2020

1

Carbon intensity: 17% below 19 2011 by 2015

Policy drivers

• Technical and operational standards

• Renewable energy and energy efficiency • Economic revitalization

• Smart meters

• Renewable energy and energy efficiency • Carbon emissions reduction

• Technical and operational standards

70% zero‐emission power by 2030 • Energy security • Carbon emissions reduction • Enhancing competitiveness of domestic industries 11% by 2030 • Energy Security • Carbon emissions reduction • Enhancing competitiveness of domestic industries

11% of total energy consumption by 2015

Policy emphases

• Power system reliability

• Reducing power generation disparities between regions • Reducing energy/carbon intensity

• Dynamic pricing and demand response programs

• Competitive retail market • Smart meters • Transmission and distribution networks modernization • Smart community • Smart meters • Solar photovoltaic generation • Smart power grid • Smart consumers • Smart transportation • Smart renewables • Smart electricity services • Ultra‐high voltage regional transmission • Upgrading and modernizing urban and rural electric grid

• Strategic economic restructuring • Renewable Energy Source of data for countries: US Energy Information Administration, International Energy Statistics (2010): http://www.eia.gov/cfapps/ipdbproject/iedindex3.cfm?tid=90&pid=44&aid=8&cid=regions&syid=2006&eyid=2010&unit=MTCDPP. Total renewables for country data include hydroelectric, geothermal, wind, solar, tide and wave, biomass, and waste. Sources of data for individual US states: US Energy Information Administration, State Energy Data System (2009): http://www.eia.gov/state/seds. Source: US Energy Information Administration, International Energy Statistics (2010): http://www.eia.gov/cfapps/ipdbproject/iedindex3.cfm?tid=2&pid=alltypes&aid=12&cid=regions&syid=2006&eyid=2010&unit=BKWH For data on individual US states, renewable energy includes geothermal, hydroelectric, solar thermal, PV, wind, wood and wood‐derived fuels, and other biomass. Source: US Energy Information Administration, Electricity (2010): http://www.eia.gov/electricity/data.cfm. Source: https://explore.data.gov/Energy‐and‐ Utilities/Electricity‐Generation‐by‐State‐by‐Type‐of‐Produce/rhyi‐ndfk.

a b

144  Advances in Energy Systems

focusing on innovation and export of smart‐grid technologies to build competitive advantages of domestic industries. Power shortages, following the 2011 nuclear incident in Japan, also accelerated the country’s investment in smart‐grid infrastructure, with the aim to integrate variable energy sources. China, the largest developing country in the world, sees smart grids as essential for renewable deployment and strategic energy industries. It also plans to close the power generation gap between regions by constructing high‐voltage direct current transmission lines. Evidence from the past decade suggests that the rapid and widespread deployment of smart‐grid technologies will not occur without supporting policies. This review of emerging smart‐grid policies in the United States, European Union, Japan, Korea, and China suggests that considerable progress has been made to develop effective policy frameworks. Nevertheless, further advances are needed to harmonize policies across nations, states, and localities, and to learn from recent experiences with this new generation of electric grid technologies[20]. As the interoperability of technologies is essential for a large‐scale and integrated deployment of smart grids, development of standards at the national and global level will be particularly important in the future. Establishment of lead agencies to coordinate efforts at various levels of governments would facilitate the standardization process, as well as address cybersecurity issues across all sectors. The electric power industry is facing tremendous opportunities and becoming increasingly important in the emerging low‐carbon economy. The costs required for the full deployment of smart grids are large. Currently, government is still the key player in smart‐grid investments. This suggests the need for a policy framework that attracts private capital investment, especially from renewable project developers and communication and information technology companies. A competitive electricity market that encourages variable business models could enhance the flexibility of electricity systems and support an increasing penetration of renewable generation technologies. Reforming the rate design mechanisms that are currently discouraging utilities’ investment in advanced technologies, and ensuring that costs and benefits are shared among all stakeholders, are also important future directions. Regulatory changes that remove barriers to a competitive energy market could also optimize overall operations and costs, hence increasing the net social benefits from smart grids. As the deployment of smart grids progresses, demand response and DG may significantly reduce

peak demand and make some generation facilities redundant. This requires sophisticated resource planning and CBA at the early stages of smart‐grid deployment. Smart‐grid customer policies, such as dynamic pricing and customer protection, require an understanding of customer behavior. New policies should be developed based on social science studies of consumer feedback and response to smart‐grid technologies and regulations. Collaboration on smart‐grid standards and sharing experiences from demonstration projects can reduce repetition and overlap in smart‐grid deployment efforts. Disseminating best practices can be particularly beneficial to those developing countries, where electricity infrastructure is expanding rapidly.

­ACKNOWLEDGMENTS The authors would like to thank the anonymous reviewer and associate editor for their invaluable comments. We are grateful to Professor Miroslav Begovic and Joy Wang (Georgia Institute of Technology) for their comments on an earlier version of this article. We are also indebted to Yeong Jae Kim (Georgia Institute of Technology) for his feedback on Korean smart‐grid policies and his help in compiling energy and carbon statistics for the countries and states examined in this chapter. Any errors or omissions are the responsibility of the authors.

­REFERENCES 1. The European Smart Grid Task Force (2010). Task Force Smart Grids—Vision and Work Programme. Brussels, Belgium: EC. 2. US Department of Energy (2009). Smart Grid System Report. Washington, D.C.: US Department of Energy. 3. The Electricity Advisory Committee (2008). Smart Grid: Enabler of the New Energy Economy. Washington, D.C.: US Department of Energy. 4. Electric Power Research Institute (2011). Estimating the Costs and Benefits of the Smart Grid—A Preliminary Estimate of the Investment Requirements and the Resultant Benefits of a Fully Functioning Smart Grid. Palo Alto, CA: EPRI. 5. Natural Resources Defense Council (2012). Removing Disincentives to Utility Energy Efficiency Efforts. New York: NRDC. 6. National Energy Technology Laboratory (2007). The NETL Modern Grid Initiative Powering our 21st‐Century Economy: Barriers to Achieving the Modern Grid. Washington, D.C.: Office of Electricity Delivery and Energy Reliability, US Department of Energy. 7. Brown, A. and Salter, R. (2010). Smart Grid Issues in State Law and Regulation. Chicago, IL: Galvin Electricity Initiative.

Smart‐Grid Policies: An International Review  145

8. Electric Power Research Institute (2011). Environmental and Health Issues Related to Radio Frequency Emissions from Smart Grid Technologies. Palo Alto, CA: EPRI. 9. International Energy Agency (2011). Energy Technology Perspectives 2010. Paris: OECD/IEA. 10. US Department of Energy. (2008). The smart grid: an introduction. Available at: http://www.oe.energy.gov/ DocumentsandMedia/DOE_SG_Book_Single_Pages. pdf. Washington, D.C.: US Department of Energy. (Accessed June 21, 2011). 11. The Brattle Group, Freeman Sullivan & Co., and Global Energy Partners, L. (2009). A National Assessment of Demand Response Potential. Federal Energy Regulatory Commission Staff Report. Washington, D.C.: FERC. 12. US Energy Information Administration (2010). International Energy Outlook 2010. Washington, D.C.: US Department of Energy. 13. US Energy Information Administration. Emissions of greenhouse gases in the U.S. Available at: http://www. eia.gov/environment/emissions/ghg_report/ghg_ overview.cfm. Washington, DC: US Department of Energy; 2011. (Accessed June 23, 2011). 14. US Energy Information Administration (2011). International Energy Outlook 2011. Washington, DC: US Department of Energy. 15. National Renewable Energy Laboratory (2012). Renewable Electricity Futures Study: Office of Energy Efficiency and Renewable Energy. Washington, DC: US Department of Energy. 16. US Energy Information Administration. International Energy Statistics—Total Carbon Dioxide Emissions From the Consumption of Energy. Available at: http:// w w w. e i a . g ov / c f a p p s / i p d b p r o j e c t / I E D I n d ex 3 . cfm?tid=90&pid=44&aid=8. Washington, DC: US Department of Energy; 2012. (Accessed July 20, 2012). 17. US Energy Information Administration. International Energy Statistics—Per Capita Carbon Dioxide Emissions From the Consumption of Energy. Available at: http://www.eia.gov/cfapps/ipdbproject/iedindex3. cfm?tid=90&pid=45&aid=8&cid=regions&syid=2006 &eyid=2010&unit=MMTCD. Washington, DC: US Department of Energy; 2012. (Accessed July 20, 2011). 18. Executive Office of the President of the United States (2011). A Policy Framework for the 21st Century Grid: Enabling our Secure Energy Future. Washington, D.C: White House. 19. Office of Electricty Delivery & Energy Reliability. American Recovery & Reinvestment Act—Recovery Act Overview. Available at: http://www.oe.energy.gov/ american_recovery_reinvestment_act.htm. Washington, DC: US Department of Energy; 2011. (Accessed June 21, 2011). 20. Brown, M.A. and Sovacool, B.K. (2011). Climate Change and Global Energy Security: Technology and Policy Options. Cambridge, MA: MIT Press. 21. US Energy Information Administration. Renewable & Alternative Fuels—State Renewable Electricity Profiles. Available at: http://www.eia.gov/renewable/state. Washington, D.C.: US Department of Energy; 2012.

22. Doris, E., Busche, S., and Hockett, S. (2009). Net Metering Policy Development in Minnesota: Overview of Trends in Nationwide Policy Development and Implications of Increasing the Eligible System Size Cap. Golden, CO: National Renewable Energy Laboratory. 23. Borenstein, S., Jaske, M., and Rosenfeld, A. (2002). Dynamic pricing, advanced metering and demand response in electricity markets. In: The Center for the Study of Energy Markets (CSEM) Working Paper Series. Berkeley, CA: University of California Energy Institute. 24. Faruqui, A. and Hledik, R. (2009). Transitioning to Dynamic Pricing. San Francisco, CA: The Brattle Group. 25. Faruqui, A. and Sergici, S. (2009). Household Response to Dynamic Pricing of Electricity—A Survey of the Experimental Evidence. San Francisco, CA: The Brattle Group. 26. Commission of the European Communities (2005). Green Paper on Energy Efficiency or Doing More with Less. (COM(2005) 265 final). Brussels, Belgium: Euoprean Commission. 27. Commission of the European Communities (2006). Green Paper: A European Strategy for Sustainable, Competitive and Secure Energy. Brussels, Belgium: European Commission. 28. European Environmental Agency. Greenhouse gas profiles—GHG trends and projections in Italy. Available at: http://www.eea.europa.eu/themes/climate/ghg‐country‐ profiles. Copenhagen, Denmark: European Environmental Agency; 2011. (Accessed June 21, 2011). 29. Italian Ministry for Economic Development (2010). Italian National Renewable Energy Action Plan. Rome, Italy: Ministero dello Sviluppo Economico. 30. Directorate‐General for Energy and Transport. Italy Renewable Energy Fact Sheet. Available at: http://www. energy.eu/renewables/factsheets/2008_res_sheet_italy_ en.pdf. Brussels, Belgium: Europe’s Energy Portal; 2008. (Accessed June 28, 2011). 31. European Commission. Development Programmes— Operational Programme Renewable Energy and Energy Efficiency. Available at: http://ec.europa.eu/regional_ p o l i c y / c o u n t r y / p r o r d n / d e t a i l s _ n ew. c f m ? g v _ PAY = B G & g v _ r eg = A L L & g v _ P G M = 1 0 4 0 & g v _ defL=4&LAN=7. Brussels, Belgium: European Commission; 2007. (Accessed July 5, 2011). 32. International Energy Agency (2011). Technology Roadmap— Smart Grids. Paris: OECD/IEA. 33. Renner, S., Albu, M., Hv, E. et  al. (2011). European Smart Metering Landscape Report. Vienna: Austrian Energy Agency. 34. Enel Distribuzione. (2011a). Smart Metering System. Available at: http://www.enel.com/en‐GB/innovation/ project_technology/zero_emission_life/smart_ networks/smart_meters.aspx?it=0. Potenza, Italy: Enel Distribuzione. (Accessed July 7, 2011). 35. Enel Distribuzione. (2011b). E‐mobility. Retrieved July 5th, 2011, from http://www.enel.com/en‐GB/innovation/ project_technology/zero_emission_life/mobile_ sustainability/e‐mobility.aspx?it=‐2. Potenza, Italy: Enel Distribuzione. (Accessed July 7, 2011).

146  Advances in Energy Systems

36. McDermott Will & Emery (2011). Italy Issues Fourth Conto Energia: New Feed‐in Tariffs for Production of Photovoltaic Energy in 2011–2016. Chicago, IL: McDermott Will & Emery. 37. DECC. (2012). Statistical Release  –  2010 UK Greenhouse Gas Emissions, Final Figures. Retrieved from www.decc.gov.uk/assets/decc/11/stats/climate‐ change/4282‐statistical‐release‐2010‐uk‐greenhousegas‐ emissi.pdf. London, UK: Department of Energy & Climate Change. (Accessed August 10, 2012). 38. HM Government (2011). Carbon Plan. London, UK: Department of Energy & Climate Change. 39. Redpoint Energy & Trilemma UK (2010). Electricity Market Reform‐Analysis of Policy: Options. London, UK: Redpoint Energy. 40. Energy and Climate Change Committee. (2010). The future of Britain’s electricity networks. Available at: http://www.publications.parliament.uk/pa/cm200910/ cmselect/cmenergy/194/19404.htm. London, UK: The House of Commons. (Accessed July 9, 2011). 41. DECC. (2009). Smarter Grids: The Opportunity. Retrieved from www.decc.gov.uk/assets/decc/what%20 we%20do/uk%20energy%20supply/futureelectricitynet works/1_20091203163757_e_@@_smartergridsop­ portunity.pdf. London, UK: Department of Energy & Climate Change. (Accessed August 10, 2012). 42. The UK Parliament (2008). Energy Act 2008. London, UK: The UK Parliament. 43. The UK Parliament. (2011). Energy Bill [HL] 2010–11 Retrieved July 5th, 2011, from http://services. parliament.uk/bills/2010–11/energyhl.html. London, UK: The UK Parliament. 44. Department of Energy and Climate Change, & Office of the Gas and Electricity Markets. (2010). Smart Metering Implementation Programme: Prospectus. Retrieved from www.decc.gov.uk/assets/decc/Consultations/smart‐meter‐ imp‐prospectus/220‐smartmetering‐prospectus‐condoc.pdf. London, UK: DECC & OFGEM. (Accessed July 14, 2011). 45. Department of Energy and Climate Change, & Office of the Gas and Electricity Markets. (2011). Smart Metering Implementation Programme: Response to Prospectus Consultation. Retrieved from www.decc.gov.uk/assets/ decc/Consultations/smart‐meterimp‐prospectus/1475‐ smart‐metering‐imp‐responseoverview.pdf. London, UK: DECC & OFGEM. (Accessed July 14, 2011). 46. Feed‐In Tariffs Ltd. (2012). Feed‐In Tariffs – Eligibility & Tariffs Retrieved June 30, 2012, from www.fitariffs. co.uk/eligible/levels. London, UK: Ownergy Plc. 47. Office of the Gas and Electricity Markets (2012). Low Carbon Networks Fund Governance Document v.5. London, UK: OFGEM. 48. Energy Technologies Institute. (2012). How we operate. Retrieved July 9th, 2012, from www.eti.co.uk/about/ how_we_operate. Loughborough, UK: ETI. 49. US Energy Information Administration. (2012). Country Analysis Briefs – Japan. Retrieved from http://www.eia. gov/cabs/Japan/pdf.pdf. US Department of Energy. Washington, DC: US Department of Energy. (Accessed August 11, 2012).

50. Greenhouse Gas Inventory Office of Japan. (2012). The GHGs Emissions Data of Japan (1990–2010). Retrieved from http://www‐gio.nies.go.jp/aboutghg/nir/nir‐e.html. Ibaraki, Japan: GIO; 2012. (Accessed August 10, 2012). 51. Ministry of Economy Trade and Industry. (2010). The Strategic Energy Plan of Japan. Retrieved July 6th, 2011, from http://www.meti.go.jp/english/press/data/ pdf/20100618_08a.pdf. Tokyo, Japan: METI. 52. Zpryme (2012). Japan: Tsunami Wakens the Smart Grid Zpryme Smart Grid Insights. Austin, Texas: Zpryme Research & Consulting, LLC. 53. Ministry of Economy Trade and Industry. (2009). The New Purchase System for Photovoltaic Electricity will be launched November 1, 2009 Retrieved July 7th, 2011, from http://www.meti.go.jp/english/press/ data/20090831_02.html. Tokyo, Japan: METI. 54. Ito S. (2009). Japan’s Initiative on Smart Grid  –  A Proposal of “Nature Grid”. In: Presentation to the European Commission. Information Economy Division, Ministry of Economy Trade and Industry (METI). Available at: http://documents.eu‐japan.eu/seminars/ europe/other/smart_grid/presentation_ogawa.pdf. Tokyo, Japan: METI. (Accessed July 10, 2011). 55. Ministry of Economy Trade and Industry. (2010). Announcement of master plans for the Demonstration of Next‐Generation Energy and Social Systems Retrieved July 6th, 2011, from http://www.meti.go.jp/english/ press/data/20100811_01.html. Tokyo, Japan: METI. 56. New Energy and Industrial Technology Development Organization. (2011). NEDO Projects  –  News List. Retrieved June 9, 2011, from http://www.nedo.go.jp/ english/archives2011_index.html. Kawasaki, Japan: NEDO. 57. Japan Smart Community Alliance. (2010). What’s JSCA. Retrieved July 5th, 2011, from https://www. smart‐japan.org/english/tabid/103/Default.aspx. Kawasaki, Japan: JSCA. 58. Yogasingam A. (2009). Toshiba partners with two companies to commercialize smart meters, EE Times. Retrieved from http://www.eetimes.com/electronics‐ news/4182604/Toshiba‐partners‐with‐two‐companiesto‐ commercialize‐smart‐meters. London, UK: UBM plc. (Accessed July 15, 2011). 59. United Nations Environment Programme (2010). Overview of the Republic of Korea’s National Strategy for Green Growth. Nairobi, Kenya: UNEP. 60. Yale Center for Environmental Law & Policy (2012). Climate Policy & Emissions Data Sheet: South Korea. New Haven, CT: Yale University. 61. Han, S. (2012). South Korea Moves Closer to Setting Limits on Carbon Emissions. New York: Bloomberg Businessweek. 62. Bae, H. and Wheelock, C. (2010). Smart Grid in Korea‐ Smart Power Grid, Consumers, Transportation, Renewable Energy, and Electricity Service: Market Analysis and Forecasts. Boulder, CO: PikeResearch. 63. KEPCO (2011). Korea Electric Power Corporation 2011 Annual Report. Seoul, Korea: Korea Electric Power Corporation.

Smart‐Grid Policies: An International Review  147

64. Ministry of Knowledge Economy and Korea Smart Grid Institute. (2005). 10 Power IT Projects. Retrieved July 7th, 2011, from http://www.smartgrid.or.kr/Ebook/10POWER_ IT_PROJECTS.PDF. Gwacheon‐si, Korea: MKE & Seoul, Korea: Korea Smart Grid Institute. 65. Ministry of Knowledge Economy and Korea Smart Grid Institute (2010) Korea’s Smart Grid Roadmap 2030: Laying the Foundation for Low Carbon, Green Growth by 2030. Retrieved 7 Jul 2011, from http://www. greentechmedia.com/images/wysiwyg/News/SG‐Road‐ Map.pdf. Gwacheon‐si, Korea: MKE & Seoul, Korea: Korea Smart Grid Institute. 66. Korea Smart Grid Institute. (2010). Korea’s Jeju Smart Grid Test‐bed Overview. Retrieved June 9, 2011, from http://www.smartgrid.or.kr/10eng3–1.php. Seoul, Korea: Korea Smart Grid Institute. 67. BP. (2011). BP Statistical Review of World Energy June 2011 Retrieved July 12th, 2011, from http://www.bp. com/sectionbodycopy.do?categoryId=7500&conten tId=7068481. London, UK: BP. 68. Austin, A. (2005). Energy and Power in China: Domestic Regulation and Foreign Policy. London, UK: The Foregin Policy Centre. 69. Smart Grid China Summit. (2011). China’s leading smart grid gathering for interaction and knowledge sharing Retrieved July 10th, 2011, from http://www. smartgridchinasummit.com/smartgrid. Shanghai, China: Smart Grid China Summit. 70. The National People’s Congress Standing Committee. (2009). Renewable Energy Law of the People’s Republic of China (Amended). Retrieved from http://www. chinagreenlp.com/files/RE_Lawpre.pdf. Beijing, China: NPCSC. (Accessed July 17, 2011). 71. Xinhua News Agency. (2011). The 12th Five‐year Plan for National Economy and Social Development of the People’s Republic of China Retrieved July 2nd, 2011,

from http://www.gov.cn/2011lh/content_1825838.htm. Beijing, China: Xinhua News Agency. 72. Ministry of Science and Technology of the People’s Republic of China. (2012). Special Planning of 12th Five‐Year Plan (2011–2015) on Smart Grid Major Science and Industrialization Projects. Retrieved from www.most.gov.cn/tztg/201205/t20120504_94114.htm. Beijing, China: MOST. (Accessed August 11, 2012). 73. State Grid Corporation of China. (2012). Corporate Profile Retrieved July 27, 2012, from www.sgcc.com. cn/gsjs/gsjj/default.shtml. Beijing, China: SGCC. 74. State Grid Corporation of China. (2011). The Acceleration of Smart Grids. Retrieved July 5th, 2011, from www.sgcc.com.cn/ztzl/newzndw/cyfz/03/243983. shtml. Beijing, China: SGCC. 75. Li, J. (2009). From Strong to Smart: The Chinese Smart Grid and its Relation with the Globe. Hongkong/ Singapore: Asia Energy Platform. 76. State Grid Coperation of China. Strong and Smart Grid: the Worldwide Attention On The China model; 2009. Available at: www.sgcc.com.cn/ztzl/tgyzl/gjhy/196859. shtml. Beijing, China: SGCC. (Accessed July 27, 2012). 77. Baker, J.N. (2011). International Energy Agency Implementing Agreement on Electricity Networks Analysis, Research and Development. Paris, France: Programme of Work, International Energy Agency. 78. Global Smart Grid Federation. About the Global Smart Grid Federation; 2011. Available at: http://www. globalsmartgridfederation.org/about.html. Washington, DC: GSGF. (Accessed July 1, 2011). 79. The International Smart Grid Action Network. Fact Sheet: International Smart Grid Action Network; 2011. Available at: http://www.cleanenergyministerial.org/pdfs/factsheets/ C E M 2 _ Fa c t _ S h e e t _ I S G A N _ 0 7 A p r i l 2 0 1 1 . p d f . Washington, D.C.: Clean Energy Ministerial. (Accessed July 2, 2011).

9

A View of Microgrids Joao A. P. Lopes1, Andre G. Madureira1 and Carlos Moreira2  Power Systems Unit, Institute for Systems and Computer Engineering of Porto (INESC Porto), Porto, Portugal

1

 Department of Electrical and Computer Engineering (DEEC), Faculty of Engineering of Porto University (FEUP), Porto, Portugal

2

Large‐scale integration of distributed energy resources in low‐voltage distribution grids will have a serious impact on power system operation. The development of the microgrid concept is presented as a solution to overcome some of the negative impacts of massive microgeneration deployment. It has paved the way for an active network management approach within the smart grid paradigm. The microgrid concept is able to address the integration of geographically dispersed energy resources, thus avoiding significant technical problems that may affect the security of operation. ­INTRODUCTION A progressive integration of distributed energy resources (DER) to low‐voltage (LV) systems will have a significant impact on planning and operation of the electrical power system. This will require a major change in the operational paradigm, evolving from an outdated passive perspective toward a future grid active management approach. To enable a massive deployment of DER, espe­ cially distributed generation (DG) based on renew­ able energy sources (RES), it is necessary to develop

advanced monitoring and novel control functional­ ities that will enable the mitigation of the technical impacts ­resulting from large‐scale integration of these units, while simultaneously trying to exploit the ­benefits brought forward by their presence in LV ­distribution systems. In this context, the microgrid concept appears as an intelligent solution to enable large‐scale integration of DER in LV networks without compromising the robustness of operation of the overall power sys­ tem. The microgrid behaves like an active cell that can be regarded as a single controlled entity from the ­perspective of the upstream medium voltage (MV) system that provides flexibility to the operation of the whole electrical distribution system. The full deployment of the microgrid concept requires the development of a dedicated smart ­metering infrastructure that will provide the means for communication between all the players involved (microgeneration units, consumers, and system opera­ tors) to fulfill the smart grid vision. In addition, the microgrid will be able to oper­ ate in normal mode, i.e. interconnected to the upstream MV network or in emergency conditions,

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

150  Advances in Energy Systems

isolated from the main power system (electrical island). The islanded mode is particularly impor­ tant because it brings additional benefits regarding security of operation given the vulnerability of the power system to external attacks, resulting either from man‐made threats (such as terrorist actions) or from natural disasters (the occurrence of hurricanes, storms, etc.). ­DISTRIBUTED ENERGY RESOURCES

Medium voltage (V > 110 kV) P, Q

Load

P, Q

Load Medium voltage

Distribution

(45 kV < V ≤ 110 kV)

(1 kV < V ≤ 45 kV) P, Q

Load

(V ≤ 1 kV)

Load Figure 9.1  Organization of conventional electric power systems.

High voltage

Very High voltage

P, Q

Low voltage

Transmission Generation

The traditional organization of electrical power sys­ tems dated from the 1950s followed a hierarchical structure with three different levels: generation, trans­ mission, and distribution. The generation level was characterized by large generators that relied mostly on three types of technologies: hydro units, thermal units based on fossil fuels (burning fuels such as coal, oil, or natural gas), and nuclear units. These central generators would feed electrical power through gen­ erator transformers to a high‐voltage (HV) transmis­ sion system. The transmission system, which could cover large distances at HV levels, was then used to transport the electrical power to be delivered to the final customers through distribution transformers[1, 2] as shown in Figure 9.1.

As a result, the conventional power system was characterized by unidirectional flows of energy from the generation to the distribution levels using an interconnected transmission network, result­ ing in rather straightforward planning and opera­ tion approaches. Furthermore, traditional utilities usually operated in well‐defined geographical ter­ ritories within local market monopolies under the strict supervision of regulatory bodies. These utilities owned the generation, transmission, and ­ distribution facilities within their assigned ser­ vice territories and financed the construction of the required facilities subject to the approval by the rel­ evant regulatory bodies[2]. In recent times, and particularly in the turning to the twenty‐first century, a growing interest in the development of DG has occurred, as opposed to the traditional central generation. However, this does not mean that DG is a purely new con­ cept. In fact, in the early days of electricity gener­ ation, electricity was only supplied to customers that were located in close proximity to the power units [3]. Moreover, to keep the balance between generation and consumption, local storage (typi­ cally in the form of batteries) was used in addition to small‐scale generation.

A View of Microgrids  151

Transmission Generation

The advent of DG faces considerable challenges and requires significant changes in the way the electrical power system is regarded at many levels, from planning to operation of the electrical power system, because the networks are changing from mere passive networks to fully active networks. The new organization of the electrical power system is shown in Figure 9.2. Nowadays, DG is considered within the wider con­ text of DER, which will play a key role in future power systems. According to Lopes et al.[4], DER include not only DG but also distributed energy storage devices as well as responsive loads, whereas other authors do not include storage within the DER concept[5, 6]. Through­ out this chapter, the term DER will follow the defini­ tion proposed by Lopes et al.[4]. In addition, important resources can be found in the demand side. It is considered that these resources include load management systems that are able to shift electricity use from peak periods to off‐peak periods and ensure energy efficiency options (e.g. reduce peak electricity demand, increase building efficiency, or reduce overall electricity demand). Consequently, DER are not only based on local generation on the customer’s side of the meter but also on means to reduce peak or average customer demand, which will largely influence the electricity supply from the distribution.

Active distribution network management is consid­ ered as a key factor to achieve cost‐effective solutions following DG integration in distribution grids at both the planning and operation stages of the distribution system. In fact, this is a huge step beyond the current “fit‐and‐forget” approach. However, with the advent of DG and improved control solutions by means of information and communication technologies (ICT), this approach needs to be revised, as it can severely limit the amount of allowable DG integration. In a future scenario, controllable DG and demand side management (DSM) are expected to also share the responsibility of delivering system support ser­ vices, a role that was previously reserved to central generation. In this case, DG will be able to displace effectively not only energy supplied by central gen­ eration but also its controllability, thus reducing the required central generation capacity and transmission infrastructures. To achieve this, the operating practice of distribution networks needs to change from passive to active, which will require a shift from traditional central control philosophy to a new distributed control paradigm. This scenario will require significant ICT capabilities, as well as new decision support tools to use all the information available from DG and demand

P, Q

(V > 110 kV) ?

P, Q

? P, Q

Load

(45 kV < V ≤ 110 kV)

DG Distribution

­ACTIVE DISTRIBUTION NETWORKS

?

P, Q

? P, Q

Load

(1 kV < V ≤ 45 kV)

DG P, Q DG

?

? P, Q

Load

(V ≤ 1 kV)

Load Figure 9.2  Integration of distributed generation in electrical power systems.

– Maintenance actions and energy price (peak time) Maintenance actions

side. Of course, this will bring additional complexity to system operation but, with correct development, this new paradigm may facilitate more reliable, cost‐ effective systems that are able to achieve maximum utilization of all resources available. Active management of distribution networks will enable the distribution system operator (DSO) to maximize the use of existing circuits by taking full advantage of generator dispatch, control of transformer taps, voltage regulators, reactive power management, and system reconfiguration in an integrated, coordinated way. This active approach to system operation can reduce the negative impact of DG on the network, thus minimizing requirements for reinforcements[7].

CHP, combined heat and power applications; GHG, greenhouse gas; PV, photovoltaic; RES, renewable energy sources.

–

Main grid failure and power quality issues

Unplanned transition Preplanned transition

Faults (on upstream or adjacent feeders)

152  Advances in Energy Systems

The Need for Distributed Resources Management and Control As already mentioned, DG connection to the distribu­ tion system has been managed on the basis of a “fit‐ and‐forget” philosophy, in which DG is regarded as a mere passive element of the system. Although this phi­ losophy works for relatively moderate penetration of this type of sources, when considering high penetration levels there is a considerable impact on the distribu­ tion system. The main technical impacts resulting from large‐scale DG integration are mostly related to the voltage rise effect, power quality issues, branch over­ load problems, protection issues, and stability issues[4]. In this context, it is necessary to know the possi­ ble solutions to cope with the problems resulting from large‐scale DG connection to the MV level or to the LV level (microgeneration). Considering relatively low levels of DG and microgeneration integration, present distribution networks are able to accommodate this generating capacity without major operational issues ­occurring. However, when the main aim is to maximize the penetration levels of these sources (to cope with the European Union’s Climate and Energy Policy)[8], the impacts on the distribution system are no longer negligi­ ble. Therefore, to face the challenges posed by a massive deployment of DG and microgeneration, while simulta­ neously obtaining the potential benefits of these units, it is imperative to develop coordinated and efficient control strategies for the operation and management of these resources. These solutions should rely on ad­ vanced control and management algorithms that may be integrated as software modules to be installed in distri­ bution network control centers. In fact, regarding coor­ dinated control, an analogy can be drawn between data and energy storage for the transfer of concepts of sys­ tem integration from computer to power engineering. In particular, the method of “cache control” presented by Strunz and Louie[9] addresses coordination between storage systems to support effective system integration.

Furthermore, the connection of small DG sources directly to the LV level of distribution networks  – ­microgeneration – is also expected to grow rapidly in a near future, thus creating autonomous active cells called ­microgrids. A microgrid can be defined as an LV feeder with several microsources (such as microturbines, micro wind generators, and solar panels), together with storage devices and controllable loads connected on that same feeder and managed by a hierarchical control system[10]. These LV microgrids may be operated either in interconnected or islanded mode, under emergency conditions[11]. The Microgrid as a Flexible Cell in Active Distribution Networks Different DG technologies such as microturbines, photovoltaic (PV) panels, or fuel cells may have rated powers ranging from a few kW up to 100 kW and can be directly connected to the LV networks. This type of DG, integrated directly next to the customer side at the LV level, is usually called microgeneration, and the corresponding generating units are called microgen­ erators or microsources. In this context, microgenera­ tion units, located at user sites, emerge as a promising opportunity to meet growing customer needs for electric power, with an emphasis on reliability and power quality. Furthermore, considering increased levels of microgeneration integration, the distribution network (particularly at the LV level) can no longer be consid­ ered as a passive element. On the contrary, microgenera­ tion can have a significant impact on the LV network, and the focus has been on assessing how much DG can be tolerated before its collective electrical impact begins to originate problems in the distribution system in terms of stability or voltage, for instance. Therefore, an adequate control and management architecture is required to facilitate the integration of microgeneration and active load management schemes. Besides, the control and management of such a system should answer for all the benefits that may be achieved at all voltage levels of the distribu­ tion network. This means that different hierarchical control strategies need to be adopted at different net­ work levels[12]. One promising way to fully accom­ plish the emerging potential of microgeneration is to assume a systemic approach – the microgrid concept. Microgrid Concept The microgrid concept may be implemented in a variety of scales, considering a part of an LV grid, an LV feeder, or even a facility, such as a house. A general classification of possible microgrid architectures and their characteris­ tics based on type of application, ownership structure, and type of loads served is presented in Table 9.1.

A View of Microgrids  153

Table 9.1 Possible microgrid architectures and their characteristics [13] (and also personal research). Utility microgrids

Industrial/commercial microgrids

Urban networks

Rural feeders

Multi facility

Application

Downtown areas

Planned islanding

Industrial parks, university campus, and shopping centers

Technologies

PV, wind, microturbine, and CHP

Hydro, PV, and wind

Main drivers Benefits

Operating modes Unplanned transition Preplanned transition

Single facility

Remote microgrids

Commercial buildings and residential buildings Microturbine, PV, CHP and fuel cell Microturbine, PV, CHP, and fuel cell

Remote communities and geographical islands

Outage management and RES integration

Power quality enhancement, reliability, and energy efficiency

GHG reduction, supply mix, congestion management, upgrade deferral, and ancillary services

Premium power quality, service differentiation (reliability levels), CHP integration, and demand response management

Interconnected and islanded mode

Interconnected and islanded mode

Electrification of remote areas and reduction in fuel consumption Supply availability, RES integration, GHG reduction, and demand response management Islanded mode

Faults (on upstream or adjacent feeders)

Main grid failure and power quality issues

–

Maintenance actions

Maintenance actions and energy price (peak time)

–

CHP, combined heat and power applications; GHG, greenhouse gas; PV, photovoltaic; RES, renewable energy sources.

Hydro, PV, and wind

154  Advances in Energy Systems

Two of the main microgrid concepts, described in detail in the following lines, are the Consortium for Electric Reliability Technology Solutions (CERTS) microgrid approach[14] from the United States and the European approach from the EU project “MI­ CROGRIDS—Large Scale Integration of Microgen­ eration to Low Voltage Grids”[15]. The microgrid concept was originally developed within the CERTS[16]. The CERTS microgrid concept assumes an aggregation of loads and microsources operating as a single system providing both power and heat[16, 17]. According to this concept, the majority of microsources must be power electric based to provide the required flexibility to ensure operation as a single aggregated system. It is this flexibility of control that allows the microgrid to present itself to the bulk power system as a single controlled unit that meets local needs for reliability and security. This approach does not accommodate the traditional operating principle that DG must be shut down automatically if problems arise in the grid. In fact, the CERTS microgrid is designed to seamlessly separate or island from the grid and later reconnect to the grid once these problems are resolved. The European microgrid concept was devel­ oped within the framework of the European project ­MICROGRIDS. According to Lopes et  al.[11], a mi­ crogrid can be defined as an LV distribution system to which small modular systems are connected. In this sense, a microgrid corresponds to an association of electrical loads and small generation systems through an LV distribution network. This means that loads and sources are physically close so that a microgrid can cor­ respond, for instance, to the network of a small urban area, to an industry, or to a large shopping center. Apart from an LV distribution network, microgeneration devices, and controllable electrical loads, a microgrid may also include storage equipment, network control and management systems, and heat recovery systems [combined heat and power applications (CHP)]. It is also assumed that the microgrid can be oper­ ated in two main situations: • Normal interconnected mode. The microgrid will be electrically connected to the main MV network either being supplied by this network totally or partially or injecting power into the main MV grid. • Emergency mode. In case there is a failure in the main MV network, the microgrid must have the ability to operate in an isolated mode, i.e. to operate in an autonomous way similar to the power systems of geographic islands. In short, a microgrid can be defined as a new type of power system comprising LV grids with small

modular generation sources, controllable loads, and storage systems, which can be connected to the main power system or be operated autonomously. Depending on the primary energy source used, on the microgenerator dimension and on the type of power interface, these microsources can be con­ sidered as noncontrollable, partially controllable, and controllable. To the utility, the microgrid can be seen as a controlled cell of the power system. To the customer, it can be designed to meet the customer’s special needs and provide additional benefits such as improved power quality and reliability, increased efficiency (through CHP applications) and local voltage support. Regarding storage systems, several different tech­ nologies may be employed, depending on the type of application required. For short‐term storage, fly­ wheels and ultracapacitors are able to ensure a fast response to events such as sudden load or renewable generation variations, which will improve safety of operation and power quality. For medium‐ and long‐ term storage solutions, other technologies such as electrochemical batteries may also be used to allow managing power balance and ensure load supply for larger time periods. The microgrid concept is usually associated with AC systems; however, DC microgrids are also an interesting possibility because the connection of DER to the networks is generally done through power electronic interfaces that can be enhanced to perform other functions apart from power injection/ absorption[18]. Furthermore, several problems related to AC microgrid systems have pushed forward the development of DC solutions for microgrids. These problems are mainly related to the lack of quality of supply, the losses produced along the line due to the pollution caused by harmonics, unbalanced and reac­ tive powers, and line overloads due to that pollution[18]. Several papers can be found in the scientific literature addressing the topic of DC microgrids[19–21]. Microgrid Architecture The architecture of the microgrid concept is presented in Figure 9.3. The microgrid is supposed to be controlled and managed by an entity installed on the LV side of the MV/LV substation – the microgrid central controller (MGCC). The MGCC has a number of crucial functions and can be seen as the interface between the microgrid and the main distribution network. At a second hierarchical level, each microgeneration and storage device is locally controlled by a microsource controller (MC) and each electrical load is locally

A View of Microgrids  155

PV MC DC AC LV

CHP

LC

MC MV

AC AC

Load LC

Load

MC

LC

AC DC

Load

MC

Storage Device

Wind Gen LC DMS

Load

MGCC

MC

MC AC AC

AC DC

Fuel Cell

LC Load

Microturbine Figure 9.3  Microgrid architecture. Source: Reprinted with permission from Ref.[22]. Copyright 2006 IEEE. Reprinted with permission from Ref.[11]. Copyright 2003 Tecnalia.

controlled by a load controller (LC). To be able to ensure proper operation of the whole system, commu­ nication between two sets of devices is required: • The LC and MC, as interfaces to control loads (through the application of an interruptibility con­ cept) and as microgeneration active and reactive power production levels, respectively; • The MGCC, as central controller that aims at pro­ moting adequate technical and management pol­ icies and providing set‐points to both LC and MC. Simultaneously, it is expected that the MGCC will be able to establish some type of communication with the distribution management systems (DMSs), located upstream in the distribution network, thus contributing to an improvement in the management and operation of the MV distribution system. Regarding the MGCC, its main functions are as follows: • During normal interconnected mode. The MGCC collects data from microsources and loads to auto­ matically perform a number of operations such as forecasting studies, economic scheduling of micro­ generation, security assessment evaluations, DSM functions, and interface with the DMS. • In emergency mode. A change in the output power control of the microgenerators is required because they change from a dispatched power mode to a

­frequency control mode in the isolated grid. In such an event, the MGCC reacts as a secondary control loop. It is also important for the MGCC to have accurate knowledge of the type of loads in the grid (to eventually adopt interruption strategies) and to use support from storage devices. As a whole, the MGCC can also be responsible for local black start strategies[23]. The black start function ensures an important advantage of microgrids in terms of improving reliability and continuity of service, by reducing interruption times. The MC and LC are local controllers aimed at contributing to the economic scheduling activities, to local control of storage devices, to load tracking activities and to managing loads with interruption or peak shaving capabilities. At an advanced stage, the microgeneration and loads will be fully integrated in electricity markets and local controllers will be in charge of preparing selling and buying offers to com­ municate to the MGCC. As previously stated, this control architecture must rely on a communication system, the main function of which is to allow the MGCC to be able to coordinate all microsources and controllable loads, through their corresponding local controllers. In normal interconnected operation, fast communications are not required. There­ fore, a communication solution based on power line communication (PLC) may be adequate, especially given the small geographic span of a microgrid.

156  Advances in Energy Systems

An alternative to the hierarchical control architecture proposed is the use of multiagent systems (MASs). In fact, MASs have been used for some time now to facil­ itate the control of individual microgrids[24–27]. According to Dimeas and Hatziargyriou[24], the use of MAS technology can solve a number of specific operational problems, such as • Centralized control is more complex because small DG units have different owners, and, therefore, sev­ eral decisions should be taken locally. • There is a lack of dedicated communication facilities. • Given that microgrids are expected to operate in a liberalized market, controller decisions for each device concerning the market should have a certain degree of “intelligence.” In Ref.[25], the same authors propose some changes to their original framework. Here, the MAS approach was developed as a tool not only to provide intelli­ gence for the needs of complex tasks but also to facilitate the design of the algorithm. In this work,

microgrid operation and, namely, its participation in the energy market are addressed. The main idea of the algorithm presented is that every device should be able to decide what is best for itself as an individual. Of course, MAS does not aim exclusively at market participation and may also be used for other function­ alities. Therefore, the proposed MAS architecture is considered as a first step toward a more comprehen­ sive control mechanism. Test Systems and Real Experimental Microgrids Most of the concepts presented have been tested in simulation environments, using microgrid test sys­ tems. Several of these typical microgrids with dif­ ferent configurations and parameters may be found in the available scientific literature[28–30]. The mi­ crogrid test system developed in the MICROGRIDS EU project is shown in Figure 9.4. This test network comprises several different microgeneration technol­ ogies, as well as a central storage device based on a flywheel system.

20 kV

0.4 kV Single residential consumer

Other lines

– Storage Split-shaft microturbine ... Commercial consumer

Apartment building

Photovoltaics

Wind generator –

...

Apartment building

Apartment building Single-shaft microturbine –

Fuel cell

– –

Figure 9.4  Low voltage microgrid test network. Source: Reprinted with permission from Ref. [22]. Copyright 2006 IEEE.

A View of Microgrids  157

Furthermore, some concepts were also implement­ ed on a small laboratory microgrid and on real net­ works, following the work developed within several research projects in Europe, America, and Asia. For instance, in Europe, MAS techniques were applied to a small isolated power system in the Greek island of Kythnos, where loads and DG units are con­ trolled by intelligent agents that cooperate to solve a certain problem[31]. Also, an MAS platform was im­ plemented in the Am Steinweg microgrid in Germany. A centralized control approach is used in the Brons­ bergen microgrid in the Netherlands where several PV panels are connected in a residential area with the possibility of isolation and reconnection to the main power grid[32]. Several other cases are present in the United States, such as the CERTS microgrids test bed and the Boston Bar microgrid as well as in Japan with the Hachinohe project and the Kyoto Eco Energy project[32, 33]. More information on these real‐world experimental microgrids can be found in Refs.[31–33]. Microgrids: Confronting Advantages and Disadvantages The massive deployment of the microgrid concept and active distribution networks will have a significant impact on power systems and is expected to bring var­ ious technical, economic, and environmental advan­ tages, but also challenges and several management and operational issues. Some of these issues are dis­ cussed in the following sections. Technical, Economic, and Environmental Benefits The development of the microgrid concept is very promising for the electric power industry as a number of advantages can be foreseen at several levels[11, 34]: 1. Operation/investment issues. Reduction of both physical and electrical distance between generat­ ing units and loads may contribute to • Improvements of reactive support of the whole system, thus enhancing the voltage profile[35]. • Reduction of transmission and distribution feeder overload. • Reduction of transmission and distribution ­losses[36]. • Reduction/postponement of investments in the expansion of transmission and large‐scale gen­ eration systems. 2. Power quality issues. Improvement in power quality and reliability in particular is achieved because of

• Better match of supply and demand, especially when involving micro‐CHP units. • Reduction of the impact of large‐scale transmis­ sion and generation outages. • Minimization of downtimes if microsources are allowed to operate autonomously making use of the control capabilities of the microgrid that involve the management of microsources, loads, and storage[37]. • Improvement of voltage profiles, in case of under‐ and overvoltages, if microsources are allowed to regulate voltage at their connection point through their power electronic interfaces either locally or using a hierarchical control approach[35]. 3. Market issues. The following advantages can be attained • Possible development of market‐driven oper­ ation procedures of microgrids will lead to a significant reduction of market power exercised by established generation companies because the microgrid acts as an aggregator for individual loads and microgeneration units, enabling them to participate in electricity markets. • Microgrids may be used to provide ancillary ser­ vices[38], namely, regarding load‐frequency con­ trol and local voltage support. • Widespread application of modular micro­ sources may contribute to a reduction in energy price in the power market with appropriate economic balance between network investment and DG utilization. 4. Environmental issues. The environmental impact of microsources is expected to be smaller than with large conventional thermal power stations. There are two main benefits of the microgrid in this topic: • Physical proximity between consumers and microsources might help increase consumer awareness toward a more rational use of energy. • Reduction of greenhouse gas (GHG) emissions could mitigate the alleged effects of climate change due to the creation of technical conditions to increase the connection of RES at the LV level. Challenges and Drawbacks Conversely, several challenges and potential drawbacks face the development of microgrids as follows[11, 34]: • Technical issues. These technical barriers are mostly related to the relative lack of experience and technical knowledge to operate and control a significant number of microsources, which require extensive real‐time and off‐line research on issues such as management, protection, and control of

158  Advances in Energy Systems

­ icrogrids. Also, specific telecommunication infra­ m structures and communication protocols need to be developed to help manage, operate, and control the microgrids. However, some of these technical ­difficulties are in the way of being overcome as more research and demonstration projects are being set up across Europe, the United States, and Asia[31]. • Cost issues. The high installation cost for ­microgrids is a big disadvantage that may be reduced if some form of subsidies from government bodies is ob­ tained as a way to encourage investment, at least for a transitory period, given the current official envi­ ronmental and carbon capture goals. • Standardization issues. As this is a comparatively recent area, standards are not yet available for ­addressing power quality, operation, and protection issues, for instance. This constitutes a serious obstacle to the massive deployment of microgrid technologies. • Administrative/legal issues. In some countries, there is a lack of legislation and regulations for the operation of microsources. However, in Portu­ gal, for instance, there is already specific legisla­ tion addressing the connection of microgeneration to the grid that establishes the tariffs to be paid to microgeneration, adopting an avoided cost strategy leading to subsidized tariffs[39]. Implementation and Deployment Costs The successful development of the microgrid con­ cept depends very much on the regulatory structure defined, which should create an adequate general framework for microgeneration and microgrids[40, 41]. Addressing economic regulation, it must be kept in mind that microgeneration and the microgrid concept are in an unfavorable position to compete with some already‐established technologies, which have benefited from mass production and learning effects for some time now[42]. In this situation, incentive schemes are used to foster the development of these technologies to facilitate competition between established technol­ ogies and microgeneration technologies. The most common incentive mechanisms used in Europe are feed‐in tariffs and quota systems[43]. The identification of costs and benefits resulting from the deployment of the microgrid concept is the first step of the process of establishing an incentive mechanism and corresponding financing source[42]. A  relatively large number of costs and benefits that DG and microgeneration induce have been identified in the scientific literature[11, 42, 44, 45]. The subsequent step is the definition of the principles to quantify and share those costs and benefits. These issues are ad­ dressed in the following section.

Cost and Benefit Sharing Following the expected increase in microgeneration connection to LV distribution grids, it is necessary to quantify the main costs and benefits resulting from the deployment of these units and the development of the microgrid concept. To quantify these costs and benefits, a division must be made between mi­ crogrid players (microgenerator and consumers) and microgrid business participants (DSO, suppliers, etc.) as proposed in Ref.[42]. According to Ref.[42], the main benefits identified related to microgeneration integration in distribution grids are the electricity value (related to the value of the generated electricity), avoided losses (resulting from network power flow reductions), avoided emis­ sions (resulting from displaced electricity generation and avoided losses), networks investment deferral (related to the expenditures to acquire and install new assets or upgrade existing ones), and generation ade­ quacy (related to the need for ensuring adequate gen­ erating capacity)[46]. Moreover, the benefits envisioned for the microgrid concept are increased reliability for microgrid participants (resulting from the possibility of islanded operation), general reliability improve­ ments (possibility of supporting network reconfigura­ tion), and network investment deferral and generation adequacy. Concerning the costs identified for microgenera­ tion and microgrids, they should include microgrid development costs and DSO costs. The first ones result from the investment in controllers, protec­ tion systems, storage devices, and from operation and maintenance expenditures. The second group of costs is related to some additional capital and opera­ tion costs to the DSO from the connection of micro­ generation to the distribution grids, such as potential investments to overcome technical problems such as inadmissible voltage profiles or line overloading[42]. As it happens with DG, the costs and bene­ fits resulting from microgeneration and microgrid deployment tend to be asymmetrically captured by different entities (e.g. consumers, microsources and microgrid owners, system operators), which results in additional difficulties to their development. Therefore, the identified costs and benefits should be shared among all involved agents when estab­ lishing microgrids so as to ensure benefits to each one of them. Therefore, the funding scheme for the incentives tends to be a combination of financial contributions from different entities. The individual contribution of each entity should be based on the share of total costs and benefits resulting from ­microgrid deployment[42].

A View of Microgrids  159

­ ECHNICAL REQUIREMENTS FOR T ­MICROGRID OPERATION Given the characteristics of microgeneration technol­ ogies, it will not be common to find fully controllable synchronous generators within a microgrid, which are units normally responsible for voltage and frequency control in conventional power systems. Indeed, most technologies are not suitable for direct connection to the electrical network due to the characteristics of the energy produced, as happens with PV panels or microturbines, which produce DC power or high frequency AC power, respectively. Therefore, power electronic interfaces (either DC/AC or AC/DC/AC converters) are required. This emphasizes the importance of inverter control in microgrid operation, especially in emergency con­ ditions[47]. As proposed in Refs.[48, 49], a control scheme based on droop control for inverters can be used to enable islanded operation. In the case of an islanding of the microgrid system, and if no synchronous gen­ erators can be found in the microgrid, it is necessary to exploit inverter control capabilities. Such an approach will enable not only islanded operation of the ­ microgrid system but also the participation in black start strategies. Furthermore, one of the most interesting prospects for microgrids is the possibility of participating in ancillary services provision such as voltage control and supply of reserves. These operational issues will be discussed in the following sections. Ancillary Services Provision In Ref.[50], the authors state that one of the most excit­ ing prospects of the distribution system of the future will be its ability to provide ancillary services. If one wants to be exhaustive in listing the possible ancillary services that microgrids can offer, the possibility of islanded operation and black start services should be included, although they should be classified as a spe­ cial type of ancillary service. In the near future, these ancillary services will be supplied in response to market signals and may be contracted over the internet. It is considered that both loads and DG will be able to supply these services because supplying the services locally is usually more efficient than supplying them from distant generating units. In addition to an intelligent distribution system, an automated market system will also be necessary to make this happen. The possible development of market‐driven operation may lead to the reduction of market power of already established generation companies and to the possible contribution of the

­ icrogrid to the provision of some ancillary services. m The main benefits that the microgrid could offer to the distribution system are congestion relief, postpone­ ment of new generation, response to load changes, and local voltage support[16]. In fact, many publications can be found that address the opportunity for microgrids in providing ancillary services[11, 13, 16, 34, 45, 51]. Therefore, microgrids might participate in open market as both suppliers and customers of electricity services, leading to overall improvement in resource utilization. This would bring significant benefits to the main power utility, whereby the central genera­ tors would be able to generate electricity freely with­ out having to provide the ancillary services[34]. Some ancillary services that may be provided by microgrids include[22, 23, 34, 45]: • • • •

Reactive power and voltage control Frequency response and supply of reserves Regulation and load following Black start

Still, there is no generalized consensus over the services that may be provided by a microgrid. First, because the microgrid cannot be regarded only as a controlled load, able to control its power demand and power factor, but rather as an entity that is able to sell power to the main grid and provide several valuable ancillary services to the utility, with appropriate payment. This possibility is especially interesting under stressed operation. For example, the possibility of controlling the load leads to excellent control of the customer voltage ­profile[34]. Thus, the deployment of capacitors for reactive power control at the customer end may be avoided if power is supplied through the microgrid. Most ancillary services deal with real‐time energy balance between microsources and loads, whereas black start is especially meant for the microgrid itself for sustaining its major loads without any exchange of power with main utility grid. In this case, a major challenge in providing these services is the communica­ tion system and its reliability and speed. Islanded Operation As previously seen, it is possible to utilize a frequency droop control, which allows DG units to communicate without an explicit communication system. When a voltage source inverter (VSI) is interconnected with a stiff AC system, characterized by an angular fre­ quency ωgrid (and terminal voltage Vgrid), the frequency (and voltage) reference is externally imposed[48]. In

160  Advances in Energy Systems

ω ωmax

ω02 ω01

ωgrid ωmin

Pmin1 Pmin2

P1

P2 Pmax1

Pmax2

P

Figure 9.5  Frequency versus active power droop.

this case, the desired output power P1 (and Q1) can be ­obtained in the VSI output by adjusting the idle values of the angular frequency ω01 (and voltage V01), as shown in Figure  9.5 (and similarly for a voltage/ versus reactive power droop). In the case of islanded operation where no synchronous generator is directly connected to the grid, depending on the DG operation mode, two con­ trol strategies can be implemented for operating the inverter[22, 52]: • PQ control. The inverter is controlled to meet a desired active and reactive power setpoint. • Voltage/frequency (V/f) control. The inverter con­ trol scheme allows an independent regulation of terminal voltage and frequency; to achieve these control characteristics, active power/frequency and reactive power/voltage droops are used. If a cluster of microsources is operated within a microgrid interconnected with the upstream grid, all the inverters can be operated in the PQ mode because the voltage and frequency references are defined by the main system. In this case, a sudden disconnection of the main power supply would lead to the loss of the microgrid, since there would be no possibility for load/generation balancing, and therefore for fre­ quency and voltage control. However, by using a VSI, operated under the V/f control logic to provide a reference for voltage and frequency, it is thus possible to operate the microgrid in islanded mode, and a smooth moving to islanded operation can be performed without changing the con­ trol mode of any inverter. This is possible since the VSI is able to operate in parallel with other voltage sources (for instance, the upstream MV network dur­ ing normal operating conditions, or with other VSI during islanded operation)[22].

Black Start Being an autonomous entity, the microgrid can also develop a local black‐start action under specific con­ ditions. If a system disturbance causes a general blackout such that the microgrid is not able to isolate and continue operating in islanded mode and if the MV system is unable to restore operation in a pre­ defined time, a first step in system recovery can be to perform a local black start using the microgrid. In conventional systems, tasks related to power res­ toration are usually carried out manually by system operators, according to predefined guidelines. These tasks must be completed as fast as possible, in real‐ time basis, and under extreme stressed conditions. In a microgrid, the whole restoration procedure is expected to be much simpler because of the small number of control variables (loads, switches, and microsourc­ es). However, the characteristics of most microgen­ erators (such as primary energy source response time constants) and the control characteristics of power electronic interfaces require the identification of very specific restoration sequences[23]. In the case of faults in the main grid, the microgrid may be disconnected from the upstream MV network and will continue to operate with as many micro­ sources as possible. The strategy to be followed will involve the MGCC and the local controllers (LC and MC), using predefined rules to be embedded in the MGCC software. Exploiting microgrid capabilities to provide fast service restoration functionalities at the LV level is an innovative aspect that will enable faster restoration times to final consumers, thus improving reliability and reducing customer interruption times. During grid reconnection, the issue of out‐of‐ phase reclosing needs to be carefully considered. The development of local controllers in close coordination with the MGCC functions needs to be evaluated from

A View of Microgrids  161

the dynamic operation point of view. The strategies to be followed to deal with these problems (black start and grid reconnection) will be embedded in the microgrid local controllers as a set of rules activated by the environmental conditions (characterized by the electrical variables voltage and frequency) and following orders from the MGCC. The main steps to be considered include building the LV network, connecting MS, controlling voltage and frequency, connecting controllable loads, and MG synchroniza­ tion with the upstream motor vehicle (MV) network, when it is available[53].

­MICROGRID DEPLOYMENT ROADMAP As previously seen, the deployment of DG, and ­microgeneration in particular, is expected to undergo several stages with different time horizons. The integration of microgeneration in LV grids is already taking place, mainly through the connection of PV panels and micro wind generators to the grid, typi­ cally in southern countries and with CHP applications in the countries in the north of Europe. In general, the current level of integration of these sources in the grid is low, involving only the development of simple rules of thumb, with little technical requirements that will not allow the development of the ­microgrid concept as described previously. In this case, mainly protec­ tion requirements for disconnecting microsources in case of abnormal operating ­conditions are consid­ ered, as the DSO is not expected to control these units (directly or indirectly). At a second stage, we may assume that microgen­ eration integration may increase, which will require some level of control leading to the microgrid con­ cept. Some basic control functionalities may be included, namely, for the power electronic devices (inverters) connected to microgeneration units in the form of local control actions. These actions can be the inclusion of a droop function for active power/ frequency control in islanded mode or a similar function for preventing overvoltages in LV networks by reducing the output generation of microsources, mostly those based on RES. In the long run, a considerable number of gener­ ating units will be connected to the LV distribution system, which will then require an active approach to network management. In this case, advanced control functionalities must be employed to miti­ gate the adverse effects resulting from large‐scale microgeneration integration. This will require the development of a dedicated control infrastructure, hierarchical control scheme presented such as the ­

previously. Furthermore, the advent of electric ve­ hicles (EVs) to be connected to the distribution sys­ tem will require an even more elaborate approach to deal with the integration of these devices and the impacts they have on the electrical power system[54]. EV batteries will be an additional resource; their charging could ­conceivably be controlled within the microgrid concept[55]. These issues are addressed in the EU ­project Mobile Energy Resources in Grids of Electricity (MERGE)[56, 57]. However, the high investment required for setting up a control and management infrastructure should be carefully considered. One possible alternative is exploiting telemetering schemes to build a smart metering infrastructure that could sustain the development of the microgrid concept. The Microgrid Under the General Smart Grid Concept Following the change of paradigm in electrical power systems, there has been a growing awareness, within the electricity supply industry, of the need to reinvent electricity networks. With the advent of new technol­ ogies for generation, networks, energy storage, load efficiency, control, and communications, as well as with the arrival of liberalized markets and environ­ mental challenges, it is necessary to have a shared and strategic vision for the electrical power system. This is seen as the way to ensure that the networks of the future can meet the future needs of customers and have a broader range of stakeholders. These active networks will efficiently link small‐ and medium‐scale power sources with demand, thus enabling efficient decisions on how best to operate in real time. The level of control required to achieve this aim is significantly higher than that found in the current transmission and distribution systems. Power flow assessment, voltage control, and protec­ tion require cost‐competitive technologies and new communication systems with more devices such as sensors and actuators than are currently used in distri­ bution systems. To manage active networks, the vision of grid computing should be adopted, which assures universal access to resources. An intelligent grid infrastructure will provide more flexibility concerning demand and supply, providing at the same time new tools for optimal and cost‐effective grid operation. Intelligent infrastructure will enable sharing of grid and ICT resources including ancillary services, balancing, and microgrids behaving as a virtual power plant (VPP)[58]. According to Pudjianto et  al.[59], the microgrid and  the VPP concepts can be regarded as vehicles

162  Advances in Energy Systems

to ­facilitate cost‐efficient integration of DER into the ­existing power system. Through aggregation, DER access to energy markets is facilitated and DER‐based system support and ancillary services can be provided[12]. Following these developments, the need arose for a coherent approach to the topic of smart grids and, in 2005, the SmartGrids European Technology Platform for Electricity Networks of the Future was established to meet the challenges seen by network owners, operators, and particularly users across the European Union[60]. According to Ref.[61], a smart grid is an electricity network that can intelligently integrate the actions of all users connected to it  –  generators, consumers, and those that do both – to deliver efficiently sustain­ able, economic, and secure electricity supplies. Smart grids deployment must not only include technology, market and commercial considerations, environmental impact, regulatory framework, standardization usage, ICT, and migration strategy, but also include societal requirements and governmental edicts. In the future, the operation of power systems will be shared between central generation and DG. Control of DG could be aggregated to form microgrids or VPP to facilitate their integration both in the physical system and in the market. As seen above, a microgrid can be regarded, within the main grid, as a controlled entity operated as a single aggregated load or generator and, given attractive remuneration, as a source of power or of ancillary services supporting the main network. Of course, there are still significant technical and commercial challenges that have to be addressed to achieve active distribution network operation and its coordinated control with the upstream conventional networks. Smart Metering as an Enabling Technology As previously stated, the deployment of smart meter­ ing can be seen as a means of pushing forward the development of microgrids and smart grid concepts by providing the infrastructure to support advanced control and management functionalities within the distribution system. Several projects across Europe have dealt with the deployment of a smart metering infrastructure. In Portugal, the InovGrid project focused on the development of a fully active distribution network based on a smart metering infrastructure, following the need to introduce more intelligence to manage and control distribution networks with large‐scale integration of microgeneration and responsive loads[62]. The city of Evora is the first Portuguese InovCity[63] and a full‐scale deployment of the technologies for smart metering is expected in a near future.

This approach goes well beyond automatic meter reading (AMR), which is a passive approach using single‐flow data communication from the meters to operation centers. In fact, smart metering exploits automated meter management (AMM) as an intelli­ gent metering service using two‐way data communi­ cation between customers, suppliers, and DSO. This will turn the smart meter into a gateway for providing various services with many potential benefits for all agents involved[64]. The communication infrastructure can utilize dif­ ferent technical solutions that may coexist, such as PLC, GPRS, or WIFI solutions. Advanced Architectures for Distribution Systems: Multimicrogrids The new operation paradigm in electrical power s­ ystems involves a growing penetration of microgen­ eration in LV networks based on the development and extension of the microgrid concept. Furthermore, MV distribution grids of the future will include a massive penetration of DG and microgrids (which can operate as active cells) that should be managed under a coor­ dinated and hierarchical control approach. This new type of system is known as a multimicrogrid system. The concept of multimicrogrids consists of a high‐ level structure, formed at the MV level, in which LV microgrids and DG units are directly connected to the MV level on several adjacent feeders[65]. For the purpose of grid control and management, microgrids, DG units, and MV loads under active DSM control can be considered as active cells in this new type of network. In this new scenario, the capability of some MV loads to be responsive to control requests under a load curtailment strategy may also be regarded as a way to get additional ancillary services. Consequently, a large number of LV networks with microsources and loads, which are no longer passive elements, must operate in a coordinated way. This means that the system to be managed increases mas­ sively in complexity and dimension, thus requiring a new control and management architecture. This con­ cept was developed within the framework of the EU project Advanced Architectures and Control Concepts for MORE MICROGRIDS[66]. Logically, this new scenario will involve the adaptation of existing DMS tools, as well as the development of new functionalities, that are able to deal with such demanding operating conditions. An effective management of this type of system requires the development of a hierarchical control architecture, where intermediary control will be exercised by a new controller  –  the Central Autonomous Management

A View of Microgrids  163

DMS

HV

HV

RTU CAMC

MV

DG

CAMC

RTU

MV

MGCC MGCC LV MGCC

LV

DG

LV

Figure 9.6  Control and management architecture of a multimicrogrid system.

Controller (CAMC)  –  to be installed at the MV bus level of a HV/MV substation, under the responsi­ bility of the DSO, which will be in charge of each multimicrogrid. In this way, the complexity of the system may be reduced by sharing tasks and respon­ sibilities among several control entities. The CAMC will behave like a mini DMS that is able to tackle the scheduling problem of generating units (both DG and microsources) and other control devices installed in the system, under normal and emergency operating conditions. The architecture foreseen for this type of system is presented in Figure 9.6. Nowadays, the DMS is wholly responsible for the supervision, control, and management of the whole distribution system. In the future, in addition to this central DMS, there may be two additional management levels: • The HV/MV substation level, where a new management agent – the CAMC – will be installed as illustrated in Figure  9.6. The CAMC will

accommodate a set of local functionalities that are n­ ormally assigned to the DMS (as well as other new functionalities) and will be responsible for interfacing the DMS with lower‐level controllers. • The microgrid level, where the MGCC, to be housed in MV/LV substations will be responsible for managing the microgrid, including the control of the microsources and responsive loads. The main issue when dealing with control strategies for multimicrogrid systems is the use of individual controllers, which should have a certain degree of autonomy and be able to communicate with each other to implement certain control actions. A partially decentralized scheme is justified by the tremendous increase in both dimension and complexity of the sys­ tem so that the management of a multimicrogrid sys­ tem requires the use of a more flexible control and management architecture. Consequently, the CAMC plays a key role in a mul­ timicrogrid system, as it will be responsible for the

164  Advances in Energy Systems

data acquisition process, for enabling the dialogue with the DMS located upstream, for running specific network functionalities, and for scheduling the differ­ ent resources in the downstream network. This new controller will also have to deal with technical and commercial constraints and contracts to manage the multimicrogrid both in HV grid‐connected operating mode and in emergency operating mode. ­CONCLUSIONS RES and electricity consumers are, in general, geo­ graphically distributed across the territory. The press­ ing need to exploit these available resources to generate electricity and the need to ensure a larger involvement from the customer side on the management and opera­ tion of the electric power system requires the adoption of a completely new set of concepts, among which microgrids play an important role. In fact, the microgrid concept is capable of ­responding to the need to accommodate in a safe and efficient way more DG in electrical grids and dealing with active management of the consumption, in particular of controllable loads, including a new type of highly flexible load that will result from a future massive deployment of EV. It is important to stress that the future deployment of a smart metering infrastructure will be crucial to support the development and widespread use of the microgrid concept as well as of the upper‐level con­ trol and management concepts associated to the management of MV distribution grids. Such a smart metering infrastructure should not be designed only to provide AMR solutions, but should instead be capa­ ble of supporting additional technical requirements. In addition, real‐time constraints over geographically dispersed areas require new secure communication protocols and infrastructures, which constitute a criti­ cal issue that needs to be tackled effectively. This new vision of power systems brought forward by the microgrid concept creates many new chal­ lenges that are multidisciplinary in nature. A robust and secure operation of such a complex and distrib­ uted system requires novel theoretical approaches regarding sensing, control, computational intelli­ gence, software, and communication. The multiple software layers that the management and control of the distribution grid will require need also to be trust­ worthy, robust, flexible, user friendly, and seamlessly integrated with the enormous databases that will be created. In summary, these issues are at the forefront of the research agenda in many disciplines and will need to be integrated to conceptualize the foundations of what is now called the smart grid.

­REFERENCES 1. Jenkins, N., Allan, R., Crossley, P. et  al. (2000). Embedded Generation. London: The Institution of Electrical Engineers. 2. Puttgen, H.B., MacGregor, P.R., and Lambert, F.C. (2003). Distributed generation: semantic hype or the dawn of a new era? IEEE Power Energy Mag. 1: 22–29. 3. Pepermans, G., Driesen, J., Haeseldonckx, D. et  al. (2005). Distributed generation: definition, benefits and issues. Energy Policy 33: 787–798. 4. Lopes, J.A.P., Hatziargyriou, N., Mutale, J. et al. (2007). Integrating distributed generation into electric power systems: a review of drivers, challenges and opportuni­ ties. Electr. Power Syst. Res. 77: 1189–1203. 5. Ackermann, T., Andersson, G., and Soder, L. (2001). Distributed generation: a definition. Electr. Power Syst. Res. 57: 195–204. 6. International Energy Agency, Development OfECa (2002). Distributed Generation in Liberalised Electricity Markets. Paris: OECD/IEA. 7. Djapic, P., Ramsay, C., Pudjianto, D. et  al. (2007). Taking an active approach. IEEE Power Energy Mag. 5: 68–77. 8. Commission E. EU Climate and Energy Package. 2008. 9. Strunz, K. and Louie, H. (2009). Cache energy control for storage: power system integration and education based on analogies derived from computer engineering. IEEE Trans. Power Syst. 24: 12–19. 10. Tsikalakis AG, Hatziargyriou ND. Centralized control for optimizing microgrids operation. In: IEEE Power and Energy Society General Meeting. San Diego, CA: IEEE; 2011. 11. Lopes JAP, Saraiva JT, Hatziargyriou N, et  al. Management of microgrids. In: International Electrical Equipment Conference. Bilbao, Spain: Tecnalia; 2003. 12. Vasquez, J.C., Guerrero, J.M., Miret, J. et  al. (2010). Hierarchical control of intelligent microgrids. IEEE Ind. Electron. Mag. 4: 23–29. 13. Driesen, J. and Katiraei, F. (2008). Design for distributed energy resources. IEEE Power Energy Mag. 6: 30–39. Consortium for Electric Reliability Technology 14. Solutions (CERTS). Available at: http://certs.lbl.gov. (Accessed June 2011). MICROGRIDS—Large 15. Scale Integration of Microgeneration to Low Voltage Grids. Available at: http://www.microgrids.eu/micro2000/index.php. (Accessed June 2011). 16. Lasseter R, Akhil A, Marnay C, et  al. The CERTS MicroGrid Concept. 2002. Available at: http://www. westernsunsystems.comorhttp://www.gosolarcalifornia. org/research/notices/2002‐05‐02_WORKSHOP_SUPP. PDF (Accessed June 2011). 17. Lasseter RH. MicroGrids. In: IEEE Power Engineering Society Winter Meeting. 2002. Available at: http:// ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber= 985003. 18. Oyarzabal J, Sanchez E, Santiago FJ. Novel Concepts for μGrids: DC Networks. 2007. Available at: http:// www.microgrids.eu/documents/645.pdf (Accessed June 2011).

A View of Microgrids  165

19. Kakigano H, Nishino A, Miura Y, et al. Distribution volt­ age control for DC microgrid by converters of energy storages considering the stored energy. In: IEEE Energy Conversion Congress and Exposition (ECCE). 2010. 20. Lie, X. and Dong, C. (2011). Control and operation of a DC microgrid with variable generation and energy stor­ age. IEEE Trans. Power Deliv. 26: 2513–2522. 21. Guerrero, J.M., Vasquez, J.C., Matas, J. et  al. (2011). Hierarchical control of droop‐controlled AC and DC microgrids: a general approach toward standardization. IEEE Trans. Ind. Electron. 58: 158–172. 22. Lopes, J.A.P., Moreira, C.L., and Madureira, A.G. (2006). Defining control strategies for microgrids islanded opera­ tion. IEEE Trans. Power Syst. 21: 916–924. 23. Moreira, C.L., Resende, F.O., and Lopes, J.A.P. (2007). Using low voltage microgrids for service restoration. IEEE Trans. Power Syst. 22: 395–403. 24. Dimeas A, Hatziargyriou N. A MultiAgent system for Microgrids. In: IEEE Power Engineering Society General Meeting. 2004. 25. Dimeas, A.L. and Hatziargyriou, N.D. (2005). Operation of a multiagent system for microgrid control. IEEE Trans. Power Syst. 20: 1447–1455. 26. Dimeas AL, Hatziargyriou ND. Agent based control for Microgrids. In: IEEE Power Engineering Society General Meeting. Tampa, FL; 2007. 27. Chatzivasiliadis SJ, Hatziargyriou ND, Dimeas AL. Development of an agent based intelligent control system for microgrids. In: IEEE Power and Energy ­ Society General Meeting. Pittsburgh, PA; 2008. 28. Moreira C. Identification and development of microgrids emergency control procedures. PhD Thesis. Department of Electrical and Computer Engineering, Faculty of Engineering of Porto University; 2008, 279. Available at: http://repositorio‐aberto.up.pt/bitstream/10216/11250/2/ Texto%20integral.pdf. (Accessed June 2011). 29. Resende F. Contributions for microgrids dynamic model­ ling and operation. PhD Thesis. Department of Electrical and Computer Engineering, Faculty of Engineering of Porto University; 2007, 293. Available at: http:// repositorio‐aberto.up.pt/bitstream/10216/11590/2/ Texto%20integral.pdf. (Accessed June 2011). 30. Madureira A. Coordinated and optimized voltage man­ agement of distribution networks with multimicrogrids. PhD Thesis. Department of Electrical and Computer Engineering, Faculty of Engineering of Porto University; 2010, 234. Available at: http://repositorio‐aberto.up.pt/ bitstream/10216/58358/1/000143265.pdf. (Accessed June 2011). 31. Hatziargyriou, N., Asano, H., Iravani, R. et  al. (2007). Microgrids: an overview of ongoing research, develop­ ment, and demonstration projects. IEEE Power Energy Mag. 5. 32. Lidula, N.W.A. and Rajapakse, A.D. (2011). Microgrids research: a review of experimental microgrids and test sys­ tems. Renewable Sustainable Energy Rev. 15: 186–202. 33. Barnes M, Kondoh J, Asano H, et al. Real‐world micro­ grids—an overview. In: System of Systems Engineering, 2007. SoSE ’07. IEEE International Conference on. 2007.

34. Chowdhury, S., Chowdhury, S.P., and Crossley, P. (2009). Microgrids and Active Distribution Networks. London: The Institution of Engineering and Technology. 35. Madureira, A.G. and Lopes, J.A.P. (2009). Coordinated voltage support in distribution networks with distributed generation and microgrids. IET Renew. Power Gener. 3: 439–454. 36. Costa, P.M. and Matos, M.A. (2009). Avoided losses on LV networks as a result of microgeneration. Electr. Power Syst. Res. 79: 629–634. 37. Costa, P.M. and Matos, M.A. (2009). Assessing the con­ tribution of microgrids to the reliability of distribution networks. Electr. Power Syst. Res. 79: 382–389. 38. Gomes, M.H. and Saraiva, J.T. (2010). Allocation of reactive power support, active loss balancing and demand interruption ancillary services in MicroGrids. Electr. Power Syst. Res. 80: 1267–1276. 39. Ministry of Economy and Innovation (MEI). Decree‐Law 363/2007 (in Portuguese). Available at: http://dre.pt/ pdf1s/2007/11/21100/0797807984.pdf. (Accessed June 2011). 40. Pudjianto D, Strbac G, van Oberbeeke F, et  al. Investigation of regulatory, commercial, economic and environmental issues in microgrids. In: International Conference on Future Power Systems. Amsterdam, The Netherlands; 2005. 41. Marnay, C., Asano, H., Papathanassiou, S. et al. (2008). Policymaking for microgrids. IEEE Power Energy Mag. 6: 66–77. 42. Costa, P.M., Matos, M.A., and Lopes, J.A.P. (2008). Regulation of microgeneration and microgrids. Energy Policy 36: 3893–3904. 43. Menanteau, P., Finon, D., and Lamy, M.‐L. (2003). Prices versus quantities: choosing policies for promot­ ing the development of renewable energy. Energy Policy 31: 799–812. 44. Vasiljevska J, Lopes JAP, Matos MA. Multi‐microgrid impact assessment using multi criteria decision aid meth­ ods. In: IEEE Powertech. Bucharest, Romania; 2009. 45. Abu‐Sharkh, S., Arnold, R.J., Kohler, J. et  al. (2006). Can microgrids make a major contribution to UK energy supply? Renewable Sustainable Energy Rev. 10: 78–127. 46. Costa, P.M. and Matos, M.A. (2010). Capacity credit of microgeneration and microgrids. Energy Policy 38: 6330–6337. 47. Zamora, R. and Srivastava, A.K. (2010). Controls for microgrids with storage: review, challenges, and research needs. Renewable Sustainable Energy Rev. 14: 2009–2018. 48. Engler, A. (2005). Applicability of droops in low voltage grids. Int. J. Distrib. Energy Resour. 1: 3–15. 49. Lasseter R, Piagi P. Providing premium power through distributed resources. In: 33rd Hawaii International Conference on System Sciences. Maui, Hawaii; 2000. 50. Kueck, J.D. and Kirby, B.J. (2003). The distribution ­system of the future. Electricity J. 16: 78–87. 51. Katiraei, F., Iravani, R., Hatziargyriou, N. et al. (2008). Microgrids management. IEEE Power Energy Mag. 6: 54–65.

166  Advances in Energy Systems

52. Lopes, J.A.P., Moreira, C.L., Madureira, A.G. et  al. (2006). Control strategies for MicroGrids emergency operation. Int. J. Distrib. Energy Resour. 2: 211–231. 53. Moreira, C.L. and Lopes, J.A.P. (2010). Microgrids operation and control under emergency conditions. Intell. Autom. Soft Comput. 16: 255–272. 54. Lopes, J.A.P., Soares, F.J., and Almeida, P.M.R. (2011). Integration of electric vehicles in the electric power sys­ tem. Proc. IEEE 99: 168–183. 55. Lopes, J.A.P., Polenz, S.A., Moreira, C.L. et al. (2010). Identification of control and management strategies for LV unbalanced microgrids with plugged‐in electric vehicles. Electr. Power Syst. Res. 80: 898–906. 56. Mobile Energy Resources for Grids of Electricity (MERGE). Available at: http://www.ev‐merge.eu. (Accessed June 2011). 57. Hatziargyriou N, Lopes JAP, Soares FJ, et  al. Mobile energy resources in grids of electricity: the EU MERGE project. In: 2nd European Conference SmartGrids and E‐Mobility. Brussels, Belgium; 2010. 58. Pudjianto, D., Ramsay, C., and Strbac, G. (2007). Virtual power plant and system integration of distributed energy resources. IET Renew. Power Gener. 1: 10–16. 59. Pudjianto, D., Ramsay, C., and Strbac, G. (2008). Microgrids and virtual power plants: concepts to support the integration of distributed energy resources. J. Power Energy 222: 731–741.

60. SmartGrids: European Technology Platform. Available at: http://www.smartgrids.eu. (Accessed June 2011). 61. SmartGrids ETP. Strategic Deployment Document for Europe’s Electricity Networks of the Future. 2008. 62. Cunha LV, Reis J, Lopes JAP. InovGrid project—distri­ bution network evolution as a decisive answer to new electrical sector challenges. IET Seminar Digests 2008, 2008:16. Available at: http://digital‐library.theiet.org/ getabs/servlet/GetabsServlet?prog=normal&id=IEESE M002008012380000016000001&idtype=cvips&gifs= yes&ref=no. (Accessed June 2011). 63. Evora InovCity ‐Smart Energy Living. Available at: http://www.inovcity.pt/en/Pages/homepage.aspx. (Accessed June 2011). 64. Vasconcelos J. Survey of Regulatory and technological developments concerning smart metering in the European union electricity market. EUI RSCAS Policy Papers, San Domenico di Fiesole, Italy; 2008. Available at: http://cadmus.eui.eu/handle/1814/926. (Accessed June 2011). 65. Madureira, A.G., Pereira, J.C., Gil, N.J. et  al. (2011). Advanced control and management functionalities for multimicrogrids. Eur. Trans. Electr. Power 21: 1159–1177. 66. Advanced Architectures and Control Concepts for MORE MICROGRIDS. Available at: http://www. microgrids.eu/index.php. (Accessed June 2011). 67.

10

New Electricity Distribution Network Planning Approaches for Integrating Renewables Fabrizio Pilo, Gianni Celli, Emilio Ghiani and Gian G. Soma University of Cagliari, Cagliari, Italy

The electricity distribution business is experiencing a tremendous and challenging transformation. The use of renewable energy sources is moving generation from the top to the bottom of power systems, where traditionally only loads existed. Active demand, distribution energy‐ storage devices, and electric vehicles are going to change even more drastically the way the distribution system will be operated. Finally, several stakeholders will share the responsibility for system operation while they often pursue opposite objectives. In contrast to conventional approaches, modern distribution planning algorithms should emulate the new environment to produce expansion and strategic plans for guiding the evolution of system in times of financial restrictions. Probabilistic methods are necessary to capture the intrinsically stochastic behavior of renewable generation, whereas the multiobjective programming is recognized to be the most effective way for planning, transparently and objectively, the system evolution, taking into account the multiple needs of different stakeholders. Finally, the integration of smart grid operation within planning algorithms is the key point for a proper distribution planning that allows integrating renewable resources and minimizing the cost for new electrical infrastructures.

­INTRODUCTION The contemporary political drives and environmental ­ concerns determined a situation in which the electricity distribution industry is facing the problem of the connection and integration of renewable and other generation plants to networks that have traditionally carried electricity from transmission systems to final consumers in one direction only, according to a passive network management scheme. If up to now, the planning of distribution networks has had the objective of defining the best expansion plan of the system for ensuring adequate capacity (e.g. substation transformers, medium voltage (MV)/low voltage (LV) secondary substations and feeders) to meet the load forecasts within the planning horizon, today the key challenge of the distribution planning is the integration of all distributed energy resources (DERs) (i.e. generators, energy storage, electric vehicles, and controllable loads) that are going to be connected to the system. In this context, new models and methodologies for planning have been proposed in the scientific literature and are still under improvement to take into

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

168  Advances in Energy Systems

account all the possibilities offered by the automation and communication technology that is going to be implemented in distribution networks according to the Smart Grid (SG) paradigm. This chapter describes the challenges in electricity distribution network planning for integrating renewables and proposes up‐to‐date ideas and methodologies to cope with them, even though it has not been possible to venture into deep detail in each issue pointed out. The list of references provided is not exhaustive, but covers the relevant works in the field of modern electricity distribution network planning.

­DISTRIBUTION PLANNING IN THE SG ERA Power distribution planning is a complex task in which planners must ensure that several objectives are met within the planning horizon. In 1997, the review paper of Khator and Leung[1] summarized the following main topics in the traditional planning of a power distribution system: • • • • • •

Optimal location of substations Optimal location of feeders Optimal individual feeder design Optimal allocation of load Optimal allocation of substation capacity Optimal mix of transformers by substation

Planning engineers were committed to find an optimal solution in which all operational issues were solved at planning level by designing the system (i.e. network solutions) so that the worst cases might be successfully handled (fit‐and‐forget policy) and few or no operation systems were necessary. Since 1998, electric power distribution has been facing tremendous changes (i.e. liberalization, unbundling, generation integration, etc.), making planning and operation more challenging and risky. In fact, in many countries, governments and regulators have been leading initiatives (e.g. incentives, feed‐in tariffs, subsidies, etc.) aimed at sustaining the “green economy” and at increasing the amount of renewable generation. As a consequence, distribution network operators (DNOs) have been overwhelmed by an increasing number of investors interested to connect distributed generation (DG) to medium and low voltage networks. Moreover, DNOs are committed to integrate, at reasonable costs, all this new generation into their distribution networks without jeopardizing the existing quality of service. In the early stage of this era, DNOs continued to apply the aforementioned fit‐and‐forget policy, which normally requires significant capital Expenditures

(CAPEX) that, depending on the regulatory framework, can lead to shallow, shallowish, or deep DG connection charges[2]. If shallowish or shallow connection costs are used, the civil society (CS) recognizes the added value of DG and renewable energy sources (RES) and is available to sustain partially or totally the system investments necessary for the integration of modern and efficient generation technologies. Instead, with the third option, the connection costs are fully paid by private investors and might become an insurmountable barrier to RES integration. Independently, by the connection charge mechanism, there is a general consensus that the fit‐and‐forget approach is no longer suitable to modern distribution because of the high and not necessary expenditures caused by its application. For this reason, plenty of SG demonstration and deployment projects are ongoing worldwide[3–8] for showing that with suitable operation systems DER do not entail CAPEX as much as the fit‐and‐forget does. Indeed, SG is conceived to favor the integration of DG and, in particular, of RES[9, 10]. Furthermore, the full integration of stationary energy‐ storage devices, plug‐in electric vehicles, and active demand through the SG communication facilities has the ambition to change the way the power system is operated, by partly abandoning the classic “load ­following” paradigm to adopt the new “generation following” paradigm that, by shaping the load demand, is essential for managing intermittent and uncertain power generation (Figure 10.1). The SG revolution aims at applying at distribution level techniques that have been used for decades in the transmission system. The future availability at the ­distribution level of an operation system is changing the objectives of system planning that will be mostly oriented to the maximum exploitation of existing assets and infrastructures, by operating them much closer to their physical limits than in the past. The hosting capacity for RES can then be increased with less network investments because operational issues can be fixed with the so‐called “no‐network” solutions. The novel planning approach is intrinsically capable to favor the integration of RES[11, 12] by incorporating “no‐network” solutions as valid planning alternatives. Table  10.1 shows the most common issues in the distribution system arising from the current high integration of renewable generation. It is interesting that each issue, for example, voltage regulation, can be faced with traditional network solutions and/or with innovative “no‐network” solutions, and that some of the most innovative solutions require the cooperation among not regulated players (e.g. the DER owners) and the DNO (the committed entity responsible for the operation of the distribution

New Electricity Distribution Network Planning Approaches for Integrating Renewables  169

Distribution management system (DMS)

Router Communication link Monitoring and control

nal

ectio

igit

al

co m

mu

nic ati

on

Bidir

ion

nicat

mmu

l co digita

Bid

ire

ctio

na

ld

Renewable energy sources

Storage

Active demand and electric vehicles Renewable energy sources Figure 10.1  Graphical representation of modern concepts in smart distribution, with the pervasive role of bidirectional digital communication systems that opens new perspectives to the distribution planning and development.

Table 10.1  Network and no‐network solutions in modern distribution planning with high shares of renewable energy sources installed. Challenge

Current solution

Future alternatives

Voltage rise

Reinforcement Operational p.f. 0.95 lagging Generation tripping

Voltage drop

Reinforcement Fixed capacitor banks

Network capacity

Reinforcement

Network power factor

Limits/bands for demand and generation

Sources of reactive power

Transmission network Fixed capacitor banks

Network asset loss of life

Strict network designs specifications based on technical and economic analyses

Volt/VAR control Storage Generation curtailment Online reconfiguration Volt/VAR control Storage Demand side response Online reconfiguration Storage Generation curtailment Demand side response Online reconfiguration Storage Unity power factor generation Storage SVC Volt/VAR control Dynamic protection settings Asset condition monitoring

170  Advances in Energy Systems

n­ etwork). For this reason, these innovative concepts in distribution operation and consequently in distribution planning require a suitable regulatory environment that will transform the current DNO into a distribution system operator (DSO), by replicating the typical functions of the transmission system operator at the distribution level. In comparison with the traditional planning topics previously cited, additional themes arise from the new SG concept and the planning of the future active distribution network (ADN): • Strategic planning for RES • Optimal integration between electricity distribution and information and communication technology (ICT) systems • Optimal choice of network topology • Optimal planning of multienergy hubs • Definition and development of new planning tools for ADN and RES integration • Optimal management of aging electricity distribution systems All these topics are important for correct implementation of the SG paradigm and are briefly discussed later in this chapter. In the rest of the present chapter, greater attention has been given to the characteristics of the modern planning tools for ADN because the availability of these tools is a prerequisite for the solution of all other issues. Strategic Planning for RES In strategic system planning, politicians and decision makers have to be aware about the impact on the system and the related investments. Distribution planners can provide decision makers with information on the areas where RES would be less harmful or more useful for the network. The integration of planning methodologies within Geographic Information Systems allows finding the most promising areas for RES or the areas where RES should be promoted, according to the regulatory environment and the incentive schemes[11]. Optimal Integration Between Electricity Distribution and ICT Systems Planning the evolution of the system without planning the integration of the enabling technologies can lead to diseconomies or impede RES development. The analysis in Table  10.1 implies that novel planning tools should integrate operation models within planning as well as the models of the communication system. Lacking in this point leads

to unreliable plans making the transition toward ADNs too expensive. For instance, the distribution management system (DMS), the core of active networks operation, relies heavily upon control capabilities and automation, which are generally absent in current distribution networks. The lack of experience, the increase in complexity, and the use of novel communication systems are perceived by DNOs weaknesses of ADNs that slowdown their acceptance of the new distribution system paradigm. However, few analyses on the expected reliability of SG have been published. In particular, Refs[13, 14] are focused on the expected reliability of SG and perform the calculations by applying a technique based on a pseudo‐sequential Monte Carlo (PSMC) simulation to the distribution system, typically used in the classical composite generation–transmission system. Meteorological models may be implemented with the PSMC simulation to reproduce fluctuating renewable generation caused by the moving of cloud patterns and/or to simulate local meteorological conditions that can degrade the performance of ICT wireless communication networks[14]. Ref.[15] is also oriented on reliability calculations but it follows an approach usually a­ dopted for the telecommunication system analyses (a combination of the fault‐tree analysis and the partial state‐space evaluation). Current planning algorithms and software tools have significant shortcomings in dealing with these emerging issues. Considering that the communication facility will not be an add‐on of the power system, but an essential part of active distribution, responsible for the reliability of SG, the cosimulation of both power and communication systems will be essential as illustrated by papers[13–17]. The above papers pioneered the development of reliability assessment of SG and, even though the data on SG are not easily available and exact quantitative calculations are not yet possible, the developed tools can be successfully used for comparative studies of different planning solutions (qualitative analysis) and help the planners to harmonize the development of SCADA, databases, communication, and automation facilities with the electric power system (protection, machinery, lines, cables, etc.)[11]. This is a very promising area of research from both academic and industrial perspective. Optimal Choice of Network Topology Active networks and RES may benefit from the adoption of meshed topology for the operation in opposition to the dominant radial scheme  [18–20]. Meshed networks may have positive effects on power losses, voltage regulation, and network reliability and on

New Electricity Distribution Network Planning Approaches for Integrating Renewables  171

the exploitation of lines and substations, particularly if high shares of DG have to be integrated. It should be recognized that the more DG is installed in distribution networks, the fewer are the reasons to keep using radial operated networks and the greater is the motivation to adopt weakly meshed systems. Compared with the radial ones, meshed networks require more complex operations, but the level of complexity is not higher than in radial ADNs or SG. On the contrary, with the automation and communication level of ADNs, the operation of meshed systems is greatly simplified. The major concerns are related to the protection system that might become inadequate to cope with the increased short circuit level. Furthermore, the coordination of overcurrent relays for selectivity is another serious issue that could be solved with the SG approach. CIGREWG C6.11[11] found that meshed operated distribution networks should be taken into account in planning for active distribution systems with efficient network automation and proactive protections. The use of short‐circuit limiters in meshed networks is a valid planning alternative that allows avoiding the refurbishment of existing breakers. The optimal allocation of short circuit limiters might be solved with algorithms similar to the ones used for the optimal allocation of capacitor banks. Not only the network architecture but also the network reconfiguration (flexibility) is fundamental to SG and can favor RES integration. A flexible network topology allows the reduction of energy losses, by minimizing changes in the DG‐scheduled power production and achieving a more reliable system. Network portions with a good balance between production and demand may be formed for autonomous functioning according to intentional islanding operation[21, 22] or as a microgrid[23–25]. Meshed operated distribution networks, if adopted, make flexible topology easier to be applied. The problem of online network reconfiguration has been studied for many years and good results have been achieved. However, those studies have small possibilities to be practically used because most of them are strongly based on passive distribution systems and do not care about DG, RES, and active management. Trying to overcome such limitations, Pilo et al.[26] proposed an optimization algorithm that could be easily integrated in planning tools to take into account benefits of online network reconfiguration running into a DMS for ADNs. Optimal Planning of Multienergy Hubs The integration of planning techniques with other energy or utility service planning is a very promising field of study. The transformation, conversion, and

storage of diverse types of energy in centralized units called energy hubs, and the combined transportation of different energy carriers over longer distances, is one vision for the next generation of energy systems[27]. Planning for energy hubs means integrating all energy sources and storage types in energy distribution networks[28]. The impact of electric vehicles, cogeneration systems, storage devices, and flexible demand on the electrical system has to be assessed with suitable tools. Plans should consider these elements and the integration with other infrastructures (e.g. the natural gas network is essential to exploit microcogeneration and cogeneration)[11]. Definition and Development of New Planning Tools for ADN and RES Integration Current distribution planning relies on deterministic models and the fit‐and‐forget policy. The majority of DNOs still adopt deterministic models even with high shares of RES in place, and does not take account of SG opportunities in planning studies[12, 29–32]. If used also to govern the transition toward ADN and the SG paradigm, this traditional planning approach might lead to costly and underutilized assets. In fact, the new distribution planning scenario characterized by a massive penetration of RES introduces a plethora of uncertainties related to several aspects: public opinion attitude to sustainable development, political drivers (i.e. subsidies or penalties), regulatory environment, fuel cost and energy price, intermittent power production (i.e. renewables), DER availability, and SG success. Therefore, new planning tools are needed to correctly deal with all these uncertainties. First of all, a probabilistic approach has to be used for the network calculations[11]. Then, the risk management should be clearly integrated. Indeed, making decisions in uncertain scenarios causes risks, and planning algorithms, particularly in ADNs, need to explicitly deal with risks, by allowing planners to make objective and transparent decisions. Planning in the SG era is then a decision‐making process applied in an uncertain scenario and decision‐theory‐based algorithms can be successfully applied[11, 33, 34]. Finally, with the liberalization of the electricity market and the unbundling of the vertically integrated utilities, new players with contrasting goals emerged in the distribution system. Therefore, contrary to the traditional single objective function (OF) (costs), the multiobjective (MO) optimization often appears more suitable for ADN planning. For instance, SG and RES may cause an increase in losses because of the combination of investment deferment with the nonhomothetic growth of power production and load demand. To minimize

172  Advances in Energy Systems

this risk, both DNO and CS would like to have truly dispersed generators for reduction of the losses. On the contrary, producers strive to fully exploit the local energy resources and do not care about the negative impact on the power delivery efficiency. Multicriteria programming has been successfully proposed to solve the tensions between DNO and producers[35, 36]. Optimal Management of Aging Electricity Distribution Systems Distribution systems in developed countries are reaching the end of their life cycles because the majority of electrical infrastructures were built in the 1960s. Repetitive equipment failure associated with the aging of distribution infrastructures is a major issue that could negate the reliability improvements achieved in recent years with the combination of asset management, predictive and corrective maintenance, novel planning strategies, and network automation. In the long term, the active operation could amplify aging problems due to the deferment of investments, but this is a general topic not specifically related with RES integration[11].

­ ODERN DISTRIBUTION PLANNING TOOLS M FOR RES INTEGRATION The ongoing revolution in the distribution sector toward the SG poses several issues because of the vague picture of the future scenario. What level of active management will be put into practice? Which regulatory framework will govern this system? Moreover, in contrast to the traditional transmission system planning, the new power generation installed in the distribution system has not been designed together with the power system requirements. The drivers for DG success are external to the power system needs, and DNOs have to plan their networks with the goal of increasing the hosting capacity at minimum cost

to accept all the received connection requests without jeopardizing both network security and the quality of supply. Unfortunately, the majority of this new generation is unpredictable (RES), and increasing levels of uncertainty are faced by DNO/DSO. Despite this planning scenario revolution, utilities still follow traditional steps of the typical planning process. The reduced willingness of DNOs to use modern planning techniques has been recently confirmed by the CIGRE WG C6.19[12]. The reasons for this situation are the lack of planning tools able to deal correctly with the future scenario and the lack of ad hoc business cases that prove the benefits of the new planning procedure in comparison with the traditional one. In this section, the main requirements for new planning methodologies are analyzed, whereas some applications of these tools to real cases are presented in the following section and they can be considered the first step toward effective business cases. Generally speaking, new planning approaches should have the following characteristics[37]: • Deal with uncertainty and explicitly with risks; • Use a probabilistic approach in load and generation modeling; • Consider time‐varying load and generation; • Are based on MO or multicriteria optimization frameworks; • Are able to select multiple, diverse, trade‐off solutions; • Are decision focused and use mathematical decision techniques; • Use appropriate time schedules and planning horizons; • Integrate the operation within the planning process; • Enable reuse of model, data, and solutions; • Make use of appropriate PC based and graphical tools. In comparison with the traditional distribution‐ planning philosophy, the most significant changes that should characterize the new planning tools for ADN can be summarized in Table 10.2. All these aspects are

Table 10.2  The most important features of modern planning with particular reference to RES integration compared with traditional planning techniques. Aspects of the planning procedure Data modeling Network calculations Network constraint violations Planning alternatives Objective function

Traditional planning approach

Modern planning approach

Worst cases of operation condition (fit and forget) Deterministic None

Time‐series models

Only network solutions

Mix between network and no‐network (active management) solutions Multiobjective approach

Single objective (costs)

Probabilistic Risk acceptance

New Electricity Distribution Network Planning Approaches for Integrating Renewables  173

discussed in the following with reference to the impact on RES integration. Data Modeling Traditionally, techniques of load analysis and synthesis were applied to generate characteristic load profiles used to categorize customers into different classes (this approach is still used by the majority of the distribution utilities). The results of the load analysis were used to produce suitable approximations of the peak demand, starting from the yearly energy consumption. This traditional representation was used in distribution network planning studies by assuming unique yearly values of demand and generation. Peak values were calculated for classifying worsen operation conditions and used, together with a constant yearly growth rate, in the deterministic fit‐and‐forget approach to plan network expansion for a predefined planning period. Average values (if necessary) were considered for Joule losses estimations and reliability analyses. The main drawback of this deterministic approach is that the distribution network is designed assuming the worsen conditions as certain, even if actually they have a very low probability of occurrence.

These simplified representations are unsuitable for the planning studies of the future ADNs. To capture the operational aspects that can affect the planning stage, the time variability of demand and generation has to be explicitly represented in the planning calculations. The challenge is to find to what extent operational aspects need to be modeled in planning. For instance, although fine granularities (e.g. minute by minute) are capable of capturing detailed operational aspects, when it comes to planning calculations, they often prove to be unnecessary and can also lead to time‐consuming processes, especially in medium‐ to long‐term horizons. On the contrary, simplistic representations will not bring the benefits from operational strategies to the planning studies, affecting the quality of the results. The most suitable solution is to adopt typical daily patterns to represent the behavior of distribution network customers in a year. These days are then divided into elementary intervals (one hour) and the network calculations are repeated sequentially for each of them. For capturing the uncertainties of the demand and the generation, loads and DG are modeled with suitable probabilistic density functions (pdfs) (e.g. normal, beta, etc.) as shown in Figure 10.2.

Load P

Wind generation P

:Uncertainty band

Costant standard deviation 𝜎P

𝜇P

0

12

24

0

Hours

Biomass generation P

Photovoltaic generation P

Variable mean power output 𝜇P and standard deviation 𝜎P

𝜎P = 0

𝜇P

0

24 Hours

24 Hours

0

24 Hours

Figure 10.2  Load and generation modeling suitable for active distribution network planning with high shares of RES.

174  Advances in Energy Systems

By so doing, active management schemes can be taken into account to solve potential constraints violations or to improve network performance. It is worth noting that the use of daily profiles preserves the chronological sequence of demand and generation states and it allows representing eventual chronological aspects related to the active management (e.g. storage scheduling). Probabilistic Approach in Electric Power Distribution Planning The several uncertainties that mark the current and future electrical distribution system suggest the use of probabilistic models to represent the typical planning data and the introduction of the risk concept in the choice of the planning alternatives. One of the main sources of uncertainty in MV distribution network calculations is the renewable generation because of the unpredictability of the primary energy sources (e.g. wind speed, solar radiation, and water flow). Also, the load exhibits natural random variations that can increase in presence of residential photovoltaic installations and uncontrolled recharge of electric vehicles[38]. These uncertainties can be modeled by suitable pdfs,[30–32] if the probabilistic data for the input variables are available. Depending on the stochastic distributions assumed (i.e. Gaussian, Beta, Rayleigh, etc.), network calculation can be performed by

specific probabilistic load flow algorithms or the more g­eneral Monte Carlo simulation approach. Instead, when the probabilistic data are unknown, the planner can draw out possible scenarios based on experience and knowledge or, for instance, using fuzzy set theory[39, 40]. The results of these calculations are the stochastic representation of the nodal voltage and branch current variables, through which the technical constraints can be verified with a relative confidence (acceptable risk of violation). For instance, the planner may define an acceptable probability of overload occurrence (e.g. 10%) and, by using the probability distribution of the branch current, he can determine the value of current that has this probability to be surmounted. If this value is greater than the thermal limit, the corresponding branch has to be upgraded. With a more‐risky choice (maximum overload occurrence of 10%), this refurbishment could be avoided (Figure 10.3). The importance of the probabilistic distribution network design is exemplified in the histogram of ­Figure  10.4. The data refer to the network planning calculations of the simple distribution system of Figure 10.3. The test considers the presence of three RES generators connected at the beginning of the planning period. Because of the unpredictable generation, in the traditional planning studies, the presence of the RES is often disregarded (case 1). This assumption leads to more expensive planning solutions. An alternative hypothesis, sometimes adopted in the

I23

3

I12 I12 with Gaussian distribution

μ12 = 120 [A] σ12 = 10 [A]

1

2

I24

ITL = 140 [A]

4

Assuming to accept a 10% of risk: 90°

I12 = 120 + 1.3 ∙(10) = 133 [A] < ITL

90°

I12

Assuming, instead, a 1% of risk: 99°

I12 = 120 + 2.3 ∙(10) = 143 [A] < ITL 99°

I12

–3σ

–σ

μ

+σ

+3σ ITL

Figure 10.3  Probabilistic network design, based on the concept of acceptable risk. Hypothesis: all nodal currents are represented with Gaussian pdfs. Consequently, also the branch current, after probabilistic load flow calculations, presents the same distribution (ITL – thermal limit of the branch conductor).

New Electricity Distribution Network Planning Approaches for Integrating Renewables  175

Primary substation MV/LV trunk node MV/LV lateral node RES generator Emergency connection

4000 CAPEX comparison

3500 3000 2500 2000 1500 1000 500 0

Case 1

Case 2

Case 3

Figure 10.4  Comparison of different approaches in the distribution network design calculation. Case 1: uncertain ­variables are disregarded. Case 2: uncertain variables are considered certain with their expected values. Case 3: uncertainties are treated with a probabilistic approach.

deterministic planning calculations, is to consider the RES power production certain and equal to its expected value (case 2). Obviously, this deterministic approach leads to the cheapest or to the most expensive network arrangement depending on the size and position of RES. In general, deterministic calculations can easily lead to unreliable schemes with a deterioration of the service quality. The probabilistic approach, instead, is more reliable and it better deals with the uncertainties that characterize the network planning problem (case 3). The examples proposed in Figures  10.2–10.4 highlight the importance of proper probabilistic calculations that are essential for a proper analysis of RES in planning. It should be recognized that probabilistic calculation are sometimes complex and cumbersome and it is often not easy to define proper modeling of distribution systems, particularly when strong correlations do exist among the stochastic variables and external correlations are imposed by market or by operation control centers. In these cases, the use of sequential or PSMC is very promising, particularly if parallel computing capabilities are exploited. The uncertainties characterize not only the behavior of loads and RES, but also their evolution during the planning period, for instance, in a long‐term planning (10–20 years), the growth rates of loads or of photovoltaic plants is normally influenced by many exogenous drivers. Therefore, decision support tools have to  be applied to assign a fitness value to a planning

solution representing the overall goodness or risk of its implementation in the different scenarios. MO or Multicriteria Decision‐Making The liberalization of the electricity market has broken the monopoly of the players involved in the power system, adding new players and stakeholders to the electric utilities (committed to network management and expansion as well as to minimize the overall system cost) like the regulator, which represents the interest of the CS and would like to favor the integration of RES at reasonable costs, and the DG owners that wish to maximize the profit from their investments, but also aggregators of active demand and small generation. The need to find compromising solutions for the conflicting goals of system stakeholders, and the difficulty of defining a unique OF, leads to MO approaches. In fact, by using MO programming, trade‐off solutions in a set of acceptable solutions, the Pareto set, can making be identified by applying suitable decision‐­ techniques. A solution belongs to the Pareto set if no improvement is possible in one objective without worsening in any other objective. It is crucial that in the absence of preference information, all nondominated solutions are considered equivalent. Thus, an MO analysis provides very useful information not only by finding single particular solutions that are nondominated, but also by deciphering the shape, extension, and correlation of the trade‐offs between objectives.

176  Advances in Energy Systems

In the literature, the MO methods are divided into two main groups[41, 42]. The first group makes use of single‐objective technique and a priori information. By changing the master OF, several solutions of the Pareto set are identified. The use of single‐objective optimization methods is known as the “classical approach” to MO optimization. The classical approach asks the user to perform an a priori decision‐making, by assigning preferences to the objectives under consideration, such that the final product is the solution that matches with those specifications. The weighted‐sum and the ε‐constrained methods are the ones that are most widely used in this category and provide a single least‐cost solution[43]. A “master” OF is optimized, whereas the remaining OFs are considered as constraints to be complied with. Alternatively, all objectives are aggregated into a single OF to be minimized or maximized according to the problem in hand. Deep knowledge of the problem is required to define adequate master objectives and constraint levels or aggregation method and weights, respectively. These procedures can be very useful to find single solutions when information is known a priori. On the contrary, several solutions of the ­ Pareto set can be found by changing the aggregation function or the master objective iteratively. These methods have their limitations: the weighted‐sum method may require a long operating time with a large number of objectives and the solutions found strongly depend on the shape of the Pareto frontier. Similarly, the ε‐constrained method requires strong a priori knowledge of the problem and it is not suitable for a large number of objectives. However, this can prove to be very time consuming, and the solutions depend on the shape of the Pareto frontier and the aggregation method[43]. The authors have proposed a MO version of their DG allocation tool, based on the ε‐constrained method, but they had to recognize that the need to make some a priori decisions was a drawback of the procedure because it was, in some cases, too much dependent on the planning engineering subjective point of view[34, 44]. The complexity of the future distribution system, with multiple players sharing the responsibility for the proper operation, suggests the use of “true” MO algorithms that produce a set of Pareto optima solutions without the use of subjective weights. These algorithms fall in the group of MO optimization methods based on evolutionary algorithms (EAs)[45]. EAs manage sets of possible solutions simultaneously, and permit identification of several solutions of the Pareto front at once. During the past 20 years, a large number of multiobjective evolutionary algorithms (MOEAs) have been developed. The main classification of these

algorithms is in first‐generation or second‐generation MOEA. The second‐generation MOEA is characterized by the use of elitism. At present, two of the most recognized algorithms of the second generation are the nonsorting genetic algorithm II (NSGA‐II),[46] and the strength Pareto evolutionary algorithm 2[47]. These algorithms allow finding an accurate, diverse, and well‐spread Pareto front, and they guarantee production of useful information for the subsequent decision‐making process. With specific formulation and modifications, however, many authors have proposed MO approaches for the DER planning optimization problem. In the recent literature, the different MO optimization techniques adopted to solve DER planning problem have been compared. They have been classified on the basis of authors (or research groups), and this can be seen as the “schools” from which this thinking on DER planning optimization is emerging[41, 42]. The existing approaches have been analyzed in detail and their limitations from both a theoretical and an empirical standpoint have been described. As a conclusion, the recent NSGA‐II demonstrates the capacity to generate a rich set of trade‐offs between the examined objectives, does not require a priori preference articulation, and develops concave ­portions of the Pareto approximate front[48]. Recent MO frameworks presented the integration of stochastic and controllable DER in the distribution grid,[49] the inherent time‐varying behavior of demand and DG (particularly when renewable sources are used),[50] the fact that load models can significantly affect the optimal location and sizing of DER in distribution systems,[51] and strategies to achieve an integration of DG units in LV and MV distribution grids, while optimizing several relevant objectives[52]. Nevertheless, in literature, a host of pioneering studies into MO DER planning has been pursued, each lacking important features of contemporary optimization theory. Most of all, they do not offer a suitable tool to lead the DER planner in the formalization of a new optimization problem. Finally, MO programming is a powerful tool, but one of its strengths could also be interpreted as a weakness, providing more than one solution, which leaves their interpretation open to the subjectivity of the planner rather than to objectivity and transparency. For this reason, this MO technique should be combined with the application of decision‐making techniques that assign a fitness value to each planning alternative in the Pareto set by assessing the overall expected goodness or the risk of its implementation. These tools can be based on the probability choice method[29, 33] that minimizes the expected costs of each

New Electricity Distribution Network Planning Approaches for Integrating Renewables  177

alternative in all envisaged scenarios, or on the risk analysis[29, 33, 53] that identifies the preferred solution as the one that minimizes the regret felt by a decision maker after verifying that the decision he had made was not optimal, given the futures that in fact has occurred. A third approach has also been proposed in the literature that, resorting to the stability areas concept, combines the two aforementioned ones and helps to recognize the best solution when it is difficult to assign a probability of occurrence to each ­scenarios[54]. Distribution Network Planning with Active Operation The operation influences the planning stage of the modern distribution system and more attention has already been paid to operation strategies that bridge the transition from passive to active/smart networks[55–57]. Active management enables the DSO to maximize the use of the existing circuits by taking full benefit from generator dispatch, demand side integration, control of transformer tap changers, reactive power management, and system reconfiguration in an integrated manner. All these ways to control and integrate DERs impact the system operation, but they also have a significant role even in the optimal development of the system or, in other terms, the active network operation significantly contaminates the planning. For instance, the generation curtailment of renewable generators and/or the load shedding of responsive loads can help to relief network congestions and therefore they are valuable planning alternatives to the classical network reinforcement/expansion. For these reasons, it is crucial that the planning tools for the ADNs integrate network operation practices (Table  10.1) in the set of feasible planning alternatives, to identify the best technical and economic balance between the innovative active management (that tends to maximize the utilization of existing assets in distribution system) and the traditional network expansion. Obviously, for an accurate comparison of the planning options, the costs of the active management implementation should be defined. CAPEX and operational expenditures (OPEX) depend on the ICT and on the regulatory environment (policy for refunding investments, obligation to serve, or remuneration of the ancillary services). Despite the different approaches followed by researchers and distribution engineers, a general framework can be proposed and used as the reference scheme for distribution planning. Indeed, the planning process is the research in the space of solutions of design alternatives that can optimize one or more OFs. All techniques and algorithms for

optimal distribution planning have the ambition to perform smart and suitable explorations of the space of solutions so that the computation burden remains feasible without compromising the quality of solutions. By so doing, possible planning solutions to be assessed and compared are produced by suitable algorithms. The calculations should be based on probabilistic approaches to capture the behavior of RES, and the assessment/comparison of alternatives can be based on one or more objectives. ­Figure 10.5 shows the general framework for planning as ­proposed by many authors and accepted by international organizations[11], even though not yet fully exploited in distribution companies[12] for the current lack of business cases that prove the validity of this scheme. The block that simulates the n­ etwork operation before the choice to use or not to use the traditional network solutions represents the main novelty recently proposed. Generally speaking, the macrocode for distribution planning can be based on the following steps: 1. Input data (technical, financial, economic, etc.). 2. Generate a set of possible expansion/development/ building alternatives. 3. Make each alternative technically sound. Are operation issues to be expected in the planning period? Try to solve them with operation, if this is not possible, resort to network solutions. 4. Assess each alternative according to the specific problem. 5. Choose the best alternative or the set of best options. ­APPLICATIONS OF PLANNING TOOLS FOR RES INTEGRATION Several different planning studies can be performed that involve the integration of RES in the power distribution system[58–82]. In general, they can be grouped into two optimization categories (involving two distinct planning tools): the network expansion over a planning period and the allocation of DER. The first group aims at identifying, both in brownfield and greenfield conditions, the network and no‐­network solutions that are necessary to implement on a distribution system to face the natural rise of energy demand and the connection of new customers and generators. Several methodologies have been proposed in the past years to solve this problem, often derived from the more traditional transmission‐­ expansion problem[58–62]. Many of them possess some of the key aspects previously mentioned (e.g. probabilistic calculations, MO approaches), but few consider the active operation as a valid planning solution[63].

178  Advances in Energy Systems

Planning Alternative To be evaluated in normal and emergency conditions, in each hour f of the typical day

Network presizing [Z] = [Z]PRE For existing a branches use available data, to new branches assign minimum allowable cross section

f=1

[Inode] = [I1node]

[Z] = [Zresized]

Solve probabilistic load flow [V 1node] & [I1branch]

f=f+1

[Inode] = [I1node]ANM

ANM block

Yes No f < 24?

Yes

Are there any constraints’ violations?

No

Evaluate the alternative with OF

Has the ANM been used?

No

Is the ANM allowed?

Yes

Discard alternative

No

Can network solutions be used?

Yes

No

Yes

Search for available network solutions

Figure 10.5  General flowchart for the technical validation of a network configuration that can be applied to any optimization algorithm for distribution planning with active operation in place. Active network management (ANM) is the block that simulates the operation of the network.

In the second category, it is possible to distinguish between the optimal placement of network devices that are useful to increase the hosting capacity of the distribution system (distributed energy storage  –  DES, Volt/VAR control devices, etc.) and the optimal allocation of DG. The usefulness of the first kind of applications is clear and, especially for DES, it has been gaining attention in the past years[64–66]. The usefulness of the second type of application is more questionable because of the widely adopted prohibition for distribution companies to own generators. However, in this case, there are two important applications: the definition of a reference network to calculate incentives or penalties to the generator connection charge, and the assessment of the impact that different regulatory frameworks can have on the diffusion of RES[67–69]. In the following, two examples of applications to both planning cases are presented. Both examples are based on the applications of the general concepts that characterize modern planning algorithms oriented to SG and RES integration. The first example aims at highlighting the impact of active management on network expansion costs. Ref[63] deals with the application of one of these modern

planning tools to the distribution network expansion problem, with the comparison of various levels of active management implementation. The test network considered was derived from a real distribution network and characterized by spurious open‐loop architecture, with long overhead laterals (rural area), and a pure open‐loop structure of underground cables (urban/industrial area). Figure  10.6 depicts only the network in the industrial area. The total demand is 16.4 MW with a constant growth rate of 3% per year (study period is three years long). Six new generators have to be connected at the beginning of the planning period: three in the industrial feeder (two 1.5 MVA CHPs (combined heat and powers) and a 1 MVA biomass gas turbine) and three in the rural feeder (two small 1 MVA wind turbines and a 9 MVA biomass gas turbine). Table 10.3 summarizes the results of the different optimizations. The fit‐and‐forget solution requires high capital investments for building new lines and upgrading the primary substation. The implementation of the generation ­curtailment solves overvoltage issues that appear in the rural area during high generation and low demand hours, but it is unable to avoid the feeder overload in the industrial area feeders caused by the

New Electricity Distribution Network Planning Approaches for Integrating Renewables  179

1

1 8

24 25

2

3 4 5

9

10 11

30 31

19

33

CHP

34 36

17

39

34

CHP

36 37 38 39

• Existing • DSR • GC + DSR • P&Q + DSR

20

17

10 11 12 13

29

19

32

34

CHP

36 37 38

20

16 17 18

GT

35

21

23

39

22

15

14

33

18

21

23

30 31

16

GT

35

22

19

32

33

18

21

23

30 31

16

7 9

CHP

28

15

14

3 4 5

6

27

13

29

2

26

12

28

GT

35 37 38

20

10 11

27 15

14

32

3 4 5

9

8

24 25

7

26 13

29

2

CHP

12

27 28

6

8

24 25

7

CHP

26

1

6

22

• P&Q

• Fit and forget • GC

Figure 10.6  Example of MV distribution network referred to an existing industrial area (pure open loop structure of underground cables).

Table 10.3  Network costs comparison in the whole planning period with different active management degree of ­implementation. Planning alternatives Fit and forget Generation curtailment (GC) Demand side response (DSR) P&Q generation control (P&Q) GC + DSR P&Q + DSR

CAPEX (k€)

OPEX (k€)

Total (k€)

Cost of EENS (k€)

Global cost (k€)

727 387 321 324 228 321

706 878 688 662 861 409

1433 1265 1009 986 1089 730

67 67 36 54 36 36

1500 1332 1045 1040 1125 766

high demand. On the contrary, the resort to active demand limits the investments on the industrial area to some branch upgrades. It is worth noticing that active demand is exploited only during emergency ­reconfigurations (after a fault), and it is used for only 7 min yr−1. Volt/VAR regulation with DG producing also reactive power reduces the CAPEX and improves voltage profile on the rural area without limiting the production of “green” active power. Finally, the integrated operation of DER (e.g. generator and responsive loads) leads to the best compromise between network expansion and the exploitation of RES. In conclusion, it is worth noticing that by taking into account active management in planning, the hosting capacity of existing circuits is increased with less economic investments (relevant to DSO and c­ ustomers

who pay for the distribution system expansion). Increasing the hosting capacity with small CAPEX is particularly relevant to DG owners (smaller connection charges, smaller delays from design to production, etc.) but also to the CS that has a green attitude and is looking for a low‐carbon future. From these remarks, it clearly emerges that distribution planning is MO, and players have contrasting objectives. The following example[81] ­ shows how MO can be applied in distribution planning to take into account the perspectives of different players. In this case, the idea is to perform a high‐level strategic planning of the system to simulate the impact on RES deployment of different compensation mechanisms to regulate the services offered by DERs with the active participation to

180  Advances in Energy Systems

Green certificate market

Green certificate purchase to preserve minimum price

Green certificate earned Electricity market

Refund for ancillary services

DG investors CAPEX OPEX

Connection and UoS charges

Po we (re r qu wa ali rd ty c & o pe nst na ra lty ints )

Income from energy produced

Feed-in tariff

Civil society

Refunds (if allowed): • for Joule losses • for DSO CAPEX • for ADN management

DSO CAPEX OPEX

Figure 10.7  Financial exchanges among distribution system stakeholders.

s­ystem operation. The CS, the DSOs, and the DG owners (investors) are the players considered ­(Figure 10.7). The CS is mainly interested in preserving the environment, and it favors the DG and the integration of RES at reasonable costs. The DSO strives to minimize the OPEX related to the distribution services. The minimization of investments might also be a DSO’s goal because of the budget restrictions or financial costs. Finally, RES investors also make decisions comparing costs (e.g. DG capital and OPEX, connection, and use of systems charges) with incomes for energy selling and incentives. The planning tool performs the optimal placement of RES in a given reference distribution network with a MO approach and probabilistic load flow calculations. Some possible different scenarios (Table 10.4) have been assumed, based on different implementation levels of active management and different regulatory scenarios. Type A scenarios are based on the current Italian regulation, whereas the B scenario is similar to the English regulation. For both scenarios, the fit‐and‐forget policy is used as benchmark, being the business as usual case. In C.1 and D.1 scenarios, each DG unit has to accept a maximum amount

0004228659.INDD 180

of energy curtailment per year as a sort of indirect use of system charge, whereas in C.2 and D.2 scenarios, investors may decide to help the DSO manage the network, getting paid for the services offered and the sharing of responsibilities. Table  10.5 shows the average value of the OFs; with reference to the integration of RES, it should be noticed that with the incentives for PV, all ­scenarios are positive for RES integration, and the payback time (PBT) is generally acceptable. Instead, if the RES incentives are removed (A.2 scenario), the profitability of the RES investments is drastically reduced and the PBT is almost doubled. Moreover, even if the DG penetration remains equal, the probability to achieve that level of integration is drastically smaller because the number of optimal configurations obtained by the MO optimization (i.e. the Pareto set) is halved. This means that only when generators are connected in specific optimal locations does the investment become acceptable. All scenarios with active management of the distribution systems allow a higher RES integration (30% more, roughly). In particular, the DG reactive power control (scenario D.1) allows the highest RES integration, preserving high incomes to the investors.

1/23/2019 10:57:34 AM

New Electricity Distribution Network Planning Approaches for Integrating Renewables  181

Table 10.4

Different scenarios for active service remuneration.

Scenario

Active management

Investor responsibility

Connection charge

Use of system charge

Incentives for RES

A.1 A.2 B C.1 C.2 D.1 D.2 E F G

No No No GC GC P&Q P&Q DSR Online RCF P&Q + DSR + RCF

No No No Committed Remunerated Committed Remunerated Remunerated No Remunerated

Shallow Shallow Shallowish Shallow Shallow Shallow Shallow Shallow Shallow Shallow

No No Yes Energy curtailment No Energy curtailment No No No No

Yes No Yes Yes Yes Yes Yes Yes Yes Yes

GC, generation curtailment; P&Q, active and reactive DG power control; DSR, demand‐side response; RCF, online network reconfiguration.

182  Advances in Energy Systems

Table 10.5

Average values in the optimal Pareto set of the OFs and of some significant planning parameters.

Scenarios OFDSO (M€) OFInv (M€) CS costs (M€) DG penetration (%) Net DSO CAPEX (k€) EL (MWh) Payback time (mean value) (yr)

A.1

A.2

1.4 51.1 4.1 140 9.5 2.52 1.8

1.4 12.8 0 140 3.2 2.52 3.1

B 2.2 48.3 5 132 105.8 2.23 1.8

C.1 0.6 33.6 13.8 174 34.2 6.76 2.0

C.2

D.1

D.2

1.1 37.5 16.7 171 34.2 5.8 2.1

0.6 37.3 14.7 175 33.7 6.7 1.9

1.2 38.6 17.3 171 34.6 5.13 2.0

E 1.3 53.6 4.8 145 4.9 2.67 1.8

F 1.3 50.5 4.2 139 6.6 2.68 1.7

0.9 41.2 13.8 171 6.2 6.19 2.0

New Electricity Distribution Network Planning Approaches for Integrating Renewables  183

In fact, by assuming a fixed RES configuration and changing the generator power factor, the same result can be obtained with a smaller reduction of the active power that is the basis for assessing the income from incentives. It should be recognized that in this case, the power converters must be oversized with an increment of CAPEX for DG investors. C.1 (D.1) scenario seems to be a good compromise for the three stakeholders. The small reduction of RES investors’ incomes is accompanied by a significant cut of expenditures for the CS. A concluding remark is that the active operation of the system is fundamental to reducing medium‐ term investments for network upgrading without posing barriers to the integration of RES or, in other words, the hosting capacity is increased with less expenditures. The active management permits balancing the advantages and disadvantages to the system players related to a higher RES integration in the distribution network. The scenarios without active management remuneration are preferable, because the reward penalizes the regulator too much, whereas the lower income obtained by investors due to the lack of the active control compensation is still acceptable.

­CONCLUSIONS Liberalization, unbundling, and RES have changed the power distribution business in the last 10–15 years. RES and DG are the most important cause of development in the distribution systems. The success of politics that aimed at incentivizing RES and efficient DG requires the development of plans that allow the DSO to cope with the technical issues arising by the intermittency and unpredictability of many DER limiting as much as possible the resort to expensive and complex network solutions. The future will be even more challenging because the SGs are going to leave the academic papers and the demonstration projects, and they are going to be fully deployed in real distribution systems with real customers to be served. Distribution planning is getting even more important than in the recent past, as it is demonstrated by the amount of research work carried out in academia and marginally by the industry so far. Some points can be considered as already consolidated: distribution planning should be probabilistic, multienergy and multicriteria, related to ICT facilities and operation controls, and should take into consideration life‐cycle costs and externalities (particularly important for RES). The analysis of the scientific production showed that academic researchers have intercepted these trends and many good solutions are

0004228659.INDD 183

now ready to be implemented in companies’ decision processes. Starting from these concepts, future works should be focused on real‐world applications, on the full exploitation of the data gathered by smart meters for a better load characterization, and on the full integration of tools for communication and power systems cosimulation.

REFERENCES 1. Khator, S.K. and Leung, L.C. (1997). Power distribution planning: a review of models and issues. IEEE Trans. Power Syst. 12 (3): 1151–1159. https://doi. org/10.1109/59.630455. 2. Knight RC, Montez JP, Knecht F, et al. ELEP—distributed generation connection charging within the European Union. Rev Curr Pract Fut Opt Eur Pol Recomm; 2005. Available at: http://www.iee‐library.eu. (Accessed October 2012). 3. European Commission. Directorate General Joint Research Centre, A view on Smart Grids from Pilot Projects: Lessons learned and current developments; 2011. Available at: http://ses.jrc.ec.europa.eu. (Accessed December 2012). 4. Council of European Energy Regulators (CEER). CEER status review of regulatory approaches to smart electricity grids. Available at: http://www.energy‐regulators.eu. 2012. (Accessed June 2012). 5. DOE Electricity Advisory Committee. Smart Grid: enabler of the new energy economy; 2008. Available at: http://energy.gov/oe/downloads/smart‐grid‐enabler‐ new‐energy‐economy. (Accessed January 2013). 6. European Commission. European technology platform Smart Grids: vision and strategy for Europe Electricity Networks of the Future; 2006. Available at: www.smart grids.eu. (Accessed June 2012). 7. Smart Grid Energy Research Center Project. Available at: http://smartgrid.ucla.edu. (Accessed June 2012). 8. OFGEM. Long‐Term Electricity Network Scenarios (LENS)—Final Report Reference number: 157/08; 2008. 9. Ackermann, T., Andersson, G., and Söder, L. (2001). Distributed generation: a definition. Electr. Power Syst.  Res. 57: 195–204. https://doi.org/10.1016/j. bbr.2011.03.031. 10. International Energy Agency (IEA) (2002). Distributed Generation in Liberalised Electricity Markets. Paris, France: OECD. 11. CIGRE (2011). WG SC6.11 Development and operation of active distribution networks. CIGRE Technical Brochure. ISBN: 978–2‐85873‐146‐6. 12. El Bakari K, Carter‐Brown C, Jupe S, et al. Survey on methods and tools for planning of active distribution networks. Proc CIRED Workshop; 2012, Lisbon (Portugal), 29–30 May, 2012. 13. Celli G, Ghiani E, Pilo F, et al. Active distribution network reliability assessment with a pseudo sequential Monte Carlo method. Proc Power Tech; 2011, Trondheim (SW), 19–23 June, 2011.

1/23/2019 10:57:35 AM

184  Advances in Energy Systems

14. Celli G, Ghiani E, Pilo F, et  al. Impact of ICT on the reliability of active distribution networks. Proc CIRED Workshop; 2012, Lisbon (Portugal), 29–30 May, 2012 15. Venemans P, Schreuder M. A method for the quantitative assessment of reliability of smart grids. Proc CIRED Workshop; 2012, Lisbon (Portugal), 29–30 May, 2012. 16. Lin H, Sambamoorthy S, Shukla S, et al. Power system and communication network co‐simulation for smart grid applications. IEEE PES Conf Innovat Smart Grid Technol; 2011, Manchester (UK), 5–7 December, 2011. https://doi.org/10.1109/ISGT.2011.5759166. 17. Godfrey T, Mullen S, Dugan RC, et al. Modeling smart grid applications with co‐simulation. IEEE Int Conf Smart Grid Commun; 2010, Gaithersburg (USA), 4–6 October, 2010. 291–296. https://doi.org/10.1109/ SMARTGRID. 2010.5622057. 18. Celli G, Pilo F, Pisano G, et  al. Distribution network interconnection for facilitating the diffusion of distributed generation. Proc CIGRE; 2005, Turin (Italy), 6–9 June, 2005. 19. Celli G, Pilo F, Pisano G. Meshed distribution networks to increase the maximum allowable distributed generation capacity. Proc 2005 CIGRE Symposium, Athens (Greece), 13–16 April, 2005. 20. Alvarez‐Herault, M.‐C., Picault, D., Caire, R. et  al. (2011). A novel hybrid network architecture to increase DG insertion in electrical distribution systems. IEEE Trans. Power Syst. 26 (2): 905–914. https://doi. org/10.1109/TPWRS.2010.2071887. 21. Pilo F, Celli G, Mocci S. Improvement of reliability in active networks with intentional islanding. Proc 2004 DRPT Conf 2004, Hong Kong (China), 5–8 April, 2:474– 479. https://doi.org/10.1109/DRPT.2004.1338023. 22. Bollen M, Sun Y, Ault G. Reliability of distribution networks with DER including intentional islanding. Proc 2005 Int Conf Future Power Syst; 2005, Amsterdam (The Netherlands), 16–18 November 498–503. https:// doi.org/10.1109/PMAPS.2006.360214. 23. Hatziargyriou, N., Asano, H., Iravani, R. et  al. (2007). Microgrids. IEEE Power Energy Mag. 5 (4): 78–94. https://doi.org/10.1109/MPAE.2007.376583. 24. Lasseter, R.H. (2011). Smart distribution: coupled microgrids. Proc. IEEE 99 (6): 1074–1082. https://doi. org/10.1109/JPROC.2011.2114630. 25. Ghiani E, Mocci S, Pilo F. Optimal reconfiguration of distribution networks according to the Microgrid paradigm. Proc 2005 Int Conf Fut Power Syst; 2005, Amsterdam (The Netherlands), 16–18 November. https://doi.org/10.1109/FPS.2005.204290. 26. Pilo, F., Pisano, G., and Soma, G.G. (2011). Optimal coordination of energy resources with a two stage on line active management. IEEE Trans. Ind. Electron. (99): 4526–4537. https://doi.org/10.1109/TIE.2011. 2107717. 27. Geidl, G., Köeppel, G., Favre‐Perrod, P. et  al. (2007). Energy hubs for the future. IEEE Power Energy Mag. 5 (1): 24–30. https://doi.org/10.1109/MPAE.2007.264850. 28. Chicco, G. and Mancarella, P. (2009). Distributed multi‐generation: a comprehensive view. Renewable

0004228659.INDD 184

Sustainable Energy Rev. 13 (17): 535–551. https://doi. org/10.1016/j.rser.2007.11.014. 29. Miranda, V. and Proenca, L.M. (1998). Probabilistic choice vs risk analysis – conflicts and synthesis in power system planning. IEEE Trans. Power Syst. (3): 1038–1043. https://doi.org/10.1109/59.709095. 30. Celli G, Cicoria R, Mocci S, et al. Probabilistic optimization of MV distribution network in presence of distributed generation. Proc 2002 PSCC Conf, Sevilla (Spain), 24–28 June, 2002. https://doi.org/10.1109/ PICA.2001.932323. 31. Carrano, E.G., Guimaraes, F.G., Takahashi, R.H.C. et al. (2007). Electric distribution network expansion under load‐evolution uncertainty using an immune ­system inspired algorithm. IEEE Trans. Power Syst. 22  (2): 851–861. https://doi.org/10.1109/TPWRS. 2007.894847. 32. Herman, R. and Gaunt, C.T. (2008). A practical probabilistic design procedure for LV residential distribution systems. IEEE Trans. Power Delivery 23 (4): 2247–2254. https://doi.org/10.1109/TPWRD. 2008.919041. 33. Carpinelli, G., Celli, G., Pilo, F. et al. (2003). Embedded generation planning under uncertainty including power quality issues. Eur. Trans. Electr. Power 13 (6): 381–389. https://doi.org/10.1002/etep.4450130607. 34. Carpinelli, G., Celli, G., Mocci, S. et  al. (2005). Optimization of embedded generation sizing and siting by using a double trade‐off method. IEE Proc. Gener. Transm. Distrib. 152: 503–513. https://doi.org/10.1049/ ip‐gtd:20045129. 35. Harrison, G.P., Piccolo, A., Siano, P. et  al. (2007). Exploring the tradeoffs between incentives for distributed generation developers and DNOs. IEEE Trans. Power Syst. 22 (2): 821–828. https://doi.org/10.1109/ TPWRS.2007.895176. 36. Siano, P., Ochoa, L.F., Harrison, G.P. et  al. (2009). Assessing the strategic benefits of distributed generation ownership for DNOs. IET Gener. Transm. Distrib. 3 (3): 225–236. https://doi.org/10.1049/iet‐gtd:20080235. 37. Ault, G.W., Foote, C.E.T., and McDonald, J.R. (2002). Distribution system planning in focus. IEEE Power Eng. Rev. 22 (1): 60–62. https://doi.org/10.1109/ MPER.2002.4311664. 38. Lopes, J.A.P., Soares, F.J., and Almeida, P.M.R. (2011). Integration of electric vehicles in the electric power system. Proc. IEEE 99 (1): 168–183. https://doi. org/10.1109/JPROC.2010.2066250. 39. Choi, J., El‐Keib, A.A., and Tran, T. (2005). A fuzzy branch and bound‐based transmission system expansion planning for the highest satisfaction level of the decision maker. IEEE Trans. Power Syst. 20 (1): 476–484. https:// doi.org/10.1109/TPWRS.2004.840446. 40. Silva, I.J., Rider, M.J., Romero, R. et  al. (2006). Transmission network expansion planning considering uncertainty in demand. IEEE Trans. Power Syst. 21 (4): 1565–1573. https://doi.org/10.1109/TPWRS.2006. 881159. 41. Alarcon‐Rodriguez, A.D., Ault, G.W., Curie, R.A.F. et al. (2008). Planning highly distributed power systems:

1/23/2019 10:57:35 AM

New Electricity Distribution Network Planning Approaches for Integrating Renewables  185

effective techniques and tools. Int. J. Distrib. Energy Resour. 4 (1): 35–53. 42. Alarcon‐Rodriguez, A., Ault, G., and Galloway, S. (2010). Multi‐objective planning of distributed energy resources: a review of the state‐of‐the‐art. Renewable Sustainable Energy Rev. 14: 1353–1366. https://doi. org/10.1016/j.rser.2010. 01.006. 43. Leyland G. Multi‐objective optimisation applied to industrial energy problems. PhD Thesis. Ecole Polytechnique Federale de Lausanne; 2002. 44. Celli, G., Ghiani, E., Mocci, S. et al. (2005). A multi‐ objective evolutionary algorithm for the sizing and siting of distributed generation. IEEE Trans. Power Syst. 20: 750–757. https://doi.org/10.1109/TPWRS.2005.846219. 45. Deb, K. (2001). Multi‐Objective Optimization Using Evolutionary Algorithms. Chichester, UK: Wiley. 46. Deb, K., Paratap, A., Agarwal, S. et  al. (2002). A fast and elitist multi‐objective genetic algorithm: NSGA‐II. IEEE Trans. Evol. Comput. 6 (2): 182–197. https://doi. org/10.1109/4235.996017. 47. Zitzler E, Laumanns M, Thiele L. SPEA2 Improving the Strength Pareto Evolutionary Algorithm, Technical Report 103. Computer Engineering and Communication Networks Lab (TIK), Swiss Federal Institute of Technology (ETH) Zurich, Gloriastrasse 35, CH‐8092 Zurich; 2001. 48. Berry M, Cornforth DJ, Platt G. An introduction to multiobjective optimisation methods for decentralised power planning. Power Energy Soc. General Meet; 2009. https://doi.org/10.1109/PES.2009.5275716. 49. Alarcon‐Rodriguez, A. and Ault, G.W. Multi‐objective planning framework for stochastic and controllable distributed energy resources. IET Renew. Power Gener. 3 (2): 227–238. https://doi.org/10.1049/iet‐rpg:20080045. 50. Ochoa, L.F., Padilha‐Feltrin, A., and Harrison, G.P. (2008). Evaluating distributed time‐varying generation through a multiobjective index. IEEE Trans. Power Delivery 23 (2): 1132–1138. https://doi.org/10.1109/ TPWRD.2005. 860262. 51. Singh, D. and Verma, K.S. (2009). Multi‐objective optimization for DG planning with load models. IEEE Trans. Power Syst. 24 (1): 427–436. https://doi. org/10.1109/ TPWRS.2008.2009483. 52. Kannan, S., Baskar, S., McCalley, J.D. et  al. (2009). Application of NSGA‐II algorithm to generation expansion planning. IEEE Trans. Power Syst. 24 (1): https:// doi.org/10.1109/TPWRS.2008.2004737. 53. Fang, R. and Hill, D.J. (2003). A new strategy for transmission expansion in competitive electricity markets. IEEE Trans. Power Syst. 18 (1): 374–380. https://doi. org/10.1109/TPWRS.2002.807083. 54. Caramia, P., Carpinelli, G., Russo, A. et  al. (2001). Decision theory criteria and cable sizing in non sinusoidal conditions. Int. J. Electr. Power Energy Syst. 23 (7): 507–516. 55. Liew, S.N. and Strbac, G. (2002). Maximizing penetration of wind generation in existing distribution networks. IEE Proc. Gener. Transm. Distrib. 149: 256–262. https://doi.org/10.1049/ip‐gtd:20020218.

0004228659.INDD 185

56. Department of Trade and Industry UK (DTI) (2004). Ancillary service provision from distributed generation. DTI. 57. Hidalgo R, Abbey C, Joós G. A review of active distribution networks enabling technologies. IEEE Power Energy Soc General Meet; 2010. https://doi.org/10.1109/ PES.2010.5590051. 58. Borges, C.L.T. and Ferreira Martins, V. (2012). Multistage expansion planning for active distribution networks under demand and distributed generation uncertainties. Int. J. Electr. Power Energy Syst. 36 (1): 107–116. https://doi.org/10.1016/j.ijepes.2011.10.031. 59. Soroudi, A. and Ehsan, M. (2010). A distribution network expansion planning model considering distributed generation options and techo‐economical issues. Energy 35 (8): 3364–3374. https://doi.org/10.1016/j.energy. 2010.04.022. 60. Martins, V.F. and Borges, C.L.T. (2011). Active distribution network integrated planning incorporating distributed generation and load response uncertainties. IEEE Trans. Power Syst. 26 (4): 2164–2172. https://doi. org/10.1109/TPWRS.2011.2122347. 61. Tse‐Chi Wang, D., Ochoa, L.F., and Harrison, G.P. (2011). Modified GA and data envelopment analysis for multistage distribution network expansion planning under uncertainty. IEEE Trans. Power Syst. 26 (2): 897– 904. https://doi.org/10.1109/TPWRS.2010.2057457. 62. Mac Donald RA, Ault GW, Mc Donald JR. A novel approach to the optimal planning of flexible active distribution networks. Proc Universities Power Eng Conf (UPEC); 2009, Glasgow (UK), 1–4 September, 2009. 63. Celli G, Pilo F, Pisano G, et  al. Optimal planning of active networks. 16th Power Systems Comput Conf; 2008, Glasgow (UK), 14–18 July, 2008. 64. Celli G, Mocci S, Pilo F, et  al. Optimal integration of energy storage in distribution networks. Proc Power Tech; 2009, IEEE Bucharest, 28 June–2 July, 2009. https://doi.org/10.1109/PTC.2009.5282268. 65. Carpinelli G, Mottola F, Proto D, et al. Optimal allocation of dispersed generators, capacitors and distributed energy storage systems in distribution networks. Proc IEEE Modern Electr Power Syst Conf; 2010, Wroclaw (Poland), 20–22 September, 2010. ISBN:978–839213157‐1. 66. Celli G, Pilo F, Soma GG, et al. Benefit assessment of energy storage for distribution network voltage regulation. Proc of CIRED Workshop; 2012, Lisbon (Portugal), 29–30 May, 2012. 67. Nara K, Hayashi Y, Ikeda K, et al. Application of tabu search to optimal placement of distributed generators. Proc 2001 IEEE PES Winter Meet Columbus (USA), 28 January–2 February, 2001, 918–923. https://doi. org/10.1109/PESW.2001.916995. 68. Celli G, Pilo F. Optimal distributed generation allocation in MV distribution networks. Proc 2001 22nd PICA Conf, 81–86, Sydney (Australia), 20–24 May, 2001. https://doi.org/10.1109/PICA.2001.932323. 69. Haesen, E., Espinoza, M., Pluymers, B. et  al. (2005). Optimal placement and sizing of distributed generator units using genetic optimization algorithms. Electr.

1/23/2019 10:57:35 AM

186  Advances in Energy Systems

Power Qual. Util. J. 11 (1): 97–104. https://doi.org/10.1 080/15325008.2011.64723. 70. Harrison, G.P. and Wallace, A.R. (2005). OPF evaluation of distribution network capacity for the connection of distributed generation. IEE Proc. Gener. Transm. Distrib. 152 (1): 115–122. https://doi.org/10.1016/j. epsr.2007.03.008. 71. Keane, A. and O’Malley, M. (2005). Optimal allocation of embedded generation on distribution networks. IEEE Trans. Power Syst. 20: 1640–1646. https://doi. org/10.1109/TPWRS.2005.852115. 72. Rau, N.S. and Wan, Y.H. (1994). Optimum location of resources in distribution planning. IEEE Trans. Power Syst. 9: 2014–2020. https://doi.org/10.1109/ 59.331463. 73. Mendez, V.H., Rivier, J., De La Fuente, J.I. et  al. (2006). Impact of distributed generation on distribution investment deferral. Int. J. Electr. Power Energy Syst. 28: 244–252. https://doi.org/10.1016/j. ijepes.2005.11.016. 74. Harrison, G.P., Piccolo, A., Siano, P. et  al. (2008). Hybrid GA and OPF evaluation of network capacity for distributed generation connections. Electr. Power Syst. Res. 78: 392–398. https://doi.org/10.1016/j.epsr.2007.03.008. 75. El‐Khattam, W., Bhattacharya, K., Hegazy, Y. et  al. (2004). Optimal investment planning for distributed generation in a competitive electricity market. IEEE Trans. Power Syst. 19: 1674–1684. 76. Pilo F, Celli G, Mocci S, et al. Multi‐objective programming for optimal DG integration in active distribution systems. Proc IEEE Power Energy Soc Gen Meet; 2010, Minneapolis (USA), July 2010. https://doi.org/10.1109/ PES.2010.5590087. 77. Zitzler, E., Deb, K., and Thiele, L. (2000). Multiobjective optimization using evolutionary algorithms – a comparative case study. Evol. Comput. J. 8: 173–195. 78. Repo S, Laaksonen H, Jarventausta P, et  al. A case study of voltage rise problem due to a large amount of distributed generation on a weak distribution network. Proc 2003 IEEE Bologna PowerTech Conf, Bologna

0004228659.INDD 186

(Italy), 23–26 June, 2003; https://doi.org/10.1109/ PTC.2003.1304703. 79. Celli G, Soma GG, Pilo F, et al. Comparison of planning alternatives for active distribution networks. CIRED Workshop; 2012, Lisbon (Portugal), 29–30 May, 2012. 80. Celli G, Pilo F, Soma GG, et  al. A multi‐objective approach to investigate active distribution network impact on the contrasting goals of the distribution system stakeholders. Proc CIGRE’ Symposium; 2010, Paris (France), 22–27 August, 2010. 81. Celli G, Mocci S, Pilo F, et al. Multi‐objective analysis of regulatory frameworks for active distribution networks. Proc CIRED; 2011, Frankfurt (Germany), 6–9 June, 2011. 82. Kuri B, Redfern M, Li F. Optimization of rating and positioning of dispersed generation with minimum network disruption. Proc 2004 IEEE PES Gen Meet, 2074–2078, Colorado (USA), June 2004. https://doi. org/10.1109/PES.2004.1373246.

FURTHER READING Brown, R.E. (2002). Electric Power Distribution Reliability. New York, NY: Marcel Dekker, Inc. Coello Coello, C.A., Lamont, G.B., and Van Veldhuizen, D.A. (2007). Evolutionary Algorithms for Solving Multi‐ Objective Problems. New York, NY: Springer. Goldberg, D.E. (1989). Genetic Algorithms in Search, Optimization and Machine Learning. Boston, MA: Addison Wesley. Jenkins, N., Allan, R., Crossley, P. et al. (2000). Embedded Generation. London, UK: IEE Publisher. Lakervi, E. and Holmes, E.J. (2003). Electricity Distribution Network Design. London, UK: IEE Power Series. Larson, R.E. and Costi, J.L. (1978–1982). Principles of Dynamic Programming. New York, NY: M. Decker. Willis, H.L. (2004). Power Distribution Planning Reference Book, 2e. New York, NY: Marcel Dekker, Inc. Willis, H.L. and Scott, W.G. (2000). Distributed Power Generation. New York, NY: Marcel Dekker.

1/23/2019 10:57:35 AM

11

Transmission Planning for Wind Energy in the United States and Europe: Status and Prospects Charles Smith1, Dale Osborn2, Robert Zavadil3, Warren Lasher4, Emilio Gómez‐Lázaro5, Ana Estanqueiro6, Thomas Trotscher7, John Tande8, Magnus Korpas8, Frans Van Hulle9, Hannele Holttinen10, Antje Orths11, Daniel Burke12, Mark O’Malley12, Jan Dobschinski13, Barry Rawn14, Madeline Gibescu14 and Lewis Dale15  UWIG, Reston, VA, USA  MISO, St. Paul, MN, USA 3  EnerNex, Knoxville, TN, USA 4  ERCOT, Taylor, TX, USA 5  Universidad de Castilla‐La Mancha, Ciudad Real, Castilla‐La Mancha, Spain 6  LNEG, Lisbon, Portugal 7  Statnett, Husebybakken 28 B, Oslo, Norway 8  SINTEF, Trondheim, Norway 9  EWEA, Brussels, Belgium 10  VTT, Finland 11  Energinet.dk, Denmark 12  UC Dublin, Dublin, Ireland 13  IWES, Germany 14  TU Delft, Delft, South Holland, the Netherlands 15  National Grid, London, UK 1

2

This chapter provides an overview of major transmis­ sion planning activities related to wind integration studies in the United States and Europe. Transmis­ sion planning for energy resources is different from planning for capacity resources. Those differences are  explained, and illustrated with examples from ­several regions of the United States and Europe. Trans­ mission planning for wind is becoming an iterative process consisting of generation expansion planning, economic‐based transmission planning, ­system reli­ ability analysis, and wind integration studies. A brief look at the policy environment in which this activity is  taking place is provided. A set of coherent and

collaborative transmission planning, siting, and per­ mitting policies, and cost allocation method must be developed to achieve the intended objectives. The scale of transmission development envisioned for this purpose will require unprecedented cooperation across multiple jurisdictional boundaries.

­INTRODUCTION At the beginning of 2011, nameplate wind capacity in the United States had exceeded 40 GW[1], whereas that in Europe had risen to 86 GW[2]. More than 35 GW

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

188  Advances in Energy Systems

of wind capacity were added globally in 2010, and in spite of the global economic slowdown following the 2008−2009 global recession, the prospects for continued development remain bright. However, one cloud on the horizon is the lack of sufficient transmis­ sion capacity to move the wind energy from the best wind resource areas, most of which are remote to the distant load centers. A critical conundrum has been recognized in the transmission planning area, and is being dealt with at the regional, national, and interna­ tional levels[3]. This is the situation wherein it may take 5–10 years to plan, permit, and construct a transmission line, whereas a wind project can be planned, permit­ ted, and constructed in 2–3 years. A remote wind proj­ ect cannot be financed until the transmission access is provided, and the transmission line cannot be built with cost recovery certainty until the need for service from the wind plant is shown, thus setting up a scheduling conflict that cannot be resolved. At the  regional level in the United States, Texas has broken the logjam with the establishment of a Competitive Renewable Energy Zone (CREZ) process, which allows transmission to be built and paid for in advance of the construction of the wind plants. This model is being applied to other parts of the United States and is beginning to be explored in Europe, for example, for accessing offshore wind power resources with the planned high‐voltage direct current (HVDC) voltage  source converter offshore “sockets” that the German transmission system opera­ tors (TSOs) have been legally required to install for off­ shore wind power development zones in Germany[4, 5].

­ RANSMISSION PLANNING FOR ENERGY T RESOURCES

Much of the United States wind generation was installed in response to legislative requirements established through a state Renewable Portfolio Stan­ dard (RPS), whereas much of the continued growth of wind power in Europe has been driven by the suc­ cess of various types of support schemes in different countries, notably the successful feed‐in tariff system, linked to achieving mandatory renewables targets set by European legislation. Wind is a nondispatchable energy resource, as opposed to the more traditional dispatchable capacity resource. As a renewable energy resource, its value is in displacing higher‐priced fossil fuels and reducing carbon emissions, as opposed to providing for ­system reliability requirements. As such, tradi­ tional capacity‐based transmission planning methods need to be modified in recognition of the different attributes of this energy source. Remote wind loca­ tions may require substantial transmission with significant associated costs. In the capacity planning world, transmission does not have to be able to pay for itself for capacity delivery requirements. In the energy planning world, the RPS or other policy direc­ tives require that a certain amount of wind energy be delivered. Wind energy also creates a large pool of low‐cost energy that may require transmission that must be able to pay for itself to be able to deliver the wind energy. Once generation is built or contracted, only the cost of producing energy is considered for operation of the generation. Generation is dispatched from the lowest‐cost energy‐producing generators first, then the next, and so on in a merit order of cost of produc­ tion, with wind energy having an assumed production cost of zero.

Traditional Transmission Planning Before deregulated markets and wind energy resources were available, generation was selected economically from a set of candidate generation types. The amount of generation of each type was chosen to produce the most economical mix of generation from the types available. A trade‐off between the capital cost of a generator and the cost to produce energy determined the amount of any one type of generation. The mag­ nitude of the total generation mix was chosen to meet the load plus some reserve margin economically[6, 7]. Transmission was planned based on meeting the peak load hour of the year, and was referred to as reliabil­ ity‐based transmission planning[8, 9]. This method solves problems associated with specific short‐term needs, but does not address the issues associated with moving large blocks of renewable energy from remote locations to load centers[10].

Transmission Planning for Large Amounts of Energy Resources – Economic Planning Transmission has an economic value in the energy market when low‐cost energy is delivered to high‐ priced areas. To justify transmission economically, the benefits from the difference in the price of energy between the low‐cost area and the high‐priced area must be greater than the annual capital and operating cost of the transmission overlay. To make this happen, usually a low cost of transmission per unit of energy delivered and a large volume of energy are required to pay for a transmission overlay. A study[11] indicates that transmission overlay designs whose benefits are greater than their costs can be developed for the United States Eastern Intercon­ nection for wind energy penetration levels from 5% to 20%. In Europe, cost−benefit considerations of

Transmission Planning for Wind Energy in the United States and Europe: Status and Prospects  189

transnational transmission development at European scale have mainly focused on increasing capacity of existing cross‐border and national transmission cor­ ridors (European Wind Integration Study [EWIS][12]], European Network of Transmission System Opera­ tors for Electricity [ENTSO‐E][13], and TradeWind[14]). Currently, a discussion on a European supergrid has started inside ENTSO‐E. A special case of the supergrid is the development of a transnational offshore grid in Northern Europe combining the functions of electricity trade and off­ shore wind power connection, which would involve the construction of new transmission highways for accessing renewable generation[15,16]. Because of the specific geographic situation where the North Sea wind resources are located surrounded by the demand markets UK‐IE, Nordic area, and Northern Europe, a substantial part of the solution for ­accessing the offshore wind power would already be provided by better interconnecting the three above‐ mentioned regions[17], as shown in Figures  11.1 (Ref.[18]) and 11.8. Several models applying optimization methods to network integration of wind power have also been proposed  –  modeling wind characteristics generally necessitates the use of stochastic programming tech­ niques. These optimization models will be inevita­ bly larger in size and complexity because of greater

diversity in power flow situations with many spa­ tially distributed and temporally fluctuating genera­ tion sources. Pragmatic modeling approaches[19] and model size limitation, combined with model decom­ position techniques[20], will help to make this class of problem more computationally tractable.

­REGIONAL PLANNING EFFORTS – STATUS AND PROSPECTS In the following Cases A–H, cost‐effective expansion alternatives are identified by comparing the difference between ideal power flows from an economic analy­ sis, or “market model” of energy resources, and the constrained power flow actually possible through the ­network. Detailed market models are used and incorpo­ rate – in addition to fuel cost minimization  e­ missions and startup costs, scheduled and unscheduled outages of plants, and operational constraints[21]. ­Different gen­ eration and exchange market designs and intervals − for example, daily, interday, and hourly − can also be con­ sidered, as in Case G of this section. The more sophisti­ cated market models are merged with a network model in Cases A, B, and E. A variety of planning approaches, including static[22] and dynamic[23, 24] can be used, includ­ ing hourly flows on an annual basis[25], as illustrated in these variety of case studies.

FINLAND

NORWAY

A

ESTONIA

S

E

North sea I

C

SWEDEN

DENMARK

B

A

L

T

LATIVA

LITHUANIA

RUSSIA POLAND IRELAND NETHERLANDS

UNITED KINGDOM

GERMANY

2030: EWEA offshore grid vision BELGIUM

FRANCE

Figure 11.1  European Wind Energy Association (EWEA) 2030 offshore grid vision.

190  Advances in Energy Systems

Figure 11.2  Joint Coordinated System Plan (JCSP) high‐voltage direct current (HVDC) overlay.

Eastern Interconnection Joint Coordinated System Plan (Case A) Transmission overlays have to be economical as well as reliable. Three west‐to‐east HVDC lines nominally scheduled at 75% of rating, with three terminals per line, cross‐linked with 765 kV a.c. for north−south connections, have been shown to form a self‐contin­ gent design that does not adversely impact the under­ lying a.c. system[11]. Over 300 constraints on the underlying system are mitigated or removed by the transmission overlay. Designing a system with a few lines is more economi­ cal and simpler to implement than upgrading 300 con­ straints simultaneously. The Eastern Interconnection Joint Coordinated System Plan (JCSP) provided for an HVDC overlay consisting of seven lines, as shown in Figure 11.2. The estimated cost for the economic transmission in the 5% wind scenario in the JCSP study is $50 billion, with a benefit to cost ratio of 1.4 : 1. The annual cost of the economic transmission is 1% of the total cost of energy delivered, consisting of annual capital, fuel,

operation and maintenance, and transmission costs. Corresponding numbers for the 20% scenario are $80 billion capital cost, with a benefit to cost ratio of 1.7 : 1, and 2% annual cost.

Eastern Wind Integration and Transmission Study (Case B) The US Department of Energy issued a report in the spring of 2008 that sketched the broad outlines of what supplying 20% of the annual electric energy demand from wind generation would look like[26]. The Eastern Wind Integration and Transmission Study was a direct follow‐up to that effort, charged with exploring many of the technical details that could not be addressed in detail in the initial summary report. The study looked at costs and transmission associ­ ated with increasing wind capacity to 20% and 30% of  retail electric energy sales in 2024 for the study area, which includes the Midwest Independent ­System Operator, the Pennsylvania‐New Jersey‐Maryland Interconnection, Southwest Power Pool, the New

Transmission Planning for Wind Energy in the United States and Europe: Status and Prospects  191

York Independent System Operator, the Independent ­ ystem Operator of New England, and the Tennessee S Valley Authority. The key transmission issues addressed by the study were an examination of the benefits from long distance transmission that moves large quantities of remote wind energy to urban markets, while access­ ing multiple wind resources that are geographically diverse. Trade‐offs between remote and local wind resources were also made. Specific findings and conclusions from development of the transmission overlays for each scenario include the following: • 800‐kV HVDC and Extra High Voltage (EHV) a.c. lines are preferred, if not required, because of the volumes of energy that must be transported across  and around the interconnection, as well as the distances involved. • The modeling indicates that significant wind gen­ eration can be accommodated as long as adequate transmission capacity is available. • Transmission offers capacity benefits in its own right, and enhances wind generation’s contribution to reliability by a measurable and significant amount. Electric Reliability Council of Texas (Case C) Texas Senate Bill 20 in 2005 was designed to break the impasse between wind generation development and transmission construction, instructing the Public Utility Commission of Texas (PUCT) to designate areas of the state as CREZ, prior to construction of wind generation resources, and to order specific transmission improve­ ments to connect these areas to major load centers[27]. Almost 3 1 years later, the PUCT designated five 2 CREZ, spanning much of West Texas from Amarillo to McCamey, and ordered $5 billion of transmission improvements to move wind generation from the CREZ to load centers, as shown in Figure 11.3. On the basis of planning studies, these transmission improve­ ments are expected to provide adequate capacity for over 18 400 MW of wind generation in West Texas. The PUCT also designated transmission companies to build these lines and set a deadline for plan comple­ tion of December 2013 – allowing the selected com­ panies less than four years to route, permit, and build over 2300 circuit miles of new 345‐kV transmission. Iberian Peninsula: Spain and Portugal (Case D) The European Council set a target of 20% share of renewable energies in European Union (EU) energy consumption by 2020. In terms of electricity in Spain,

40% should be generated by renewable energy power stations. The Spanish target by 2020 is 40 GW in onshore wind power, together with 5 GW in offshore wind power plants. Therefore, the transmission net­ work must be updated to integrate new renewable power stations. The Spanish TSO, Red Electrica de Espafia, is planning an investment of 8000 M€ during 2007–2016, as shown in Table 11.1[28]. Power system design and operation have been con­ ducted through different scenarios in 2016. The study[29] was conducted in a summer demand situation with a seasonal nonextreme peak level of 92%. Spain is divided in four zones to study the influence of wind power in the transmission system. Wind power gener­ ation is set up to 80% of the installed capacity in the studied area, whereas the wind power generation in the other three areas is fixed according to studies of statistical production data. Load flow, short circuit, and stability studies were conducted to study network contingency situ­ ations and system recovery after a disturbance[30, 31]. The planned power generation must be capable of providing mainly dynamic voltage control, given the massive penetration of these new technologies. These studies were conducted in peak demand scenarios. Other additional services are more appropriate to be analyzed in low‐demand situations such as voltage regulation control and frequency control[32, 33]. The study concludes that the planned wind power capacity can be integrated into the Spanish power system, highlighting some prerequisites such as the development of the planned transport network and compliance with the actual and proposed technical grid code requirements. Some significant challenges remain in the areas of dynamic voltage control and management of reserves. In Portugal, the transmission network planning and operation is a governmental concession to the TSO, Redes Energeticas Portuguesas, S.A. (REN). REN ­implemented the governmental targets to install 6.9 GW of wind capacity until 2020 to ensure 45% of the con­ sumed electricity by renewable energy system (RES). The TSO plan of investments for 2006–2010 includes 300 km of very high‐voltage transmission lines, construction and reinforcement of substa­ tions, and the operation of phase shift transformers. Figure  11.4 (Ref.[34]) depicts the lines driven by independent power producers, mainly wind and hydro, with a share ranging from 100% to 25%. REN followed the recommend methodologies[35] to assess the impact of the spatial distribution of the wind gen­ eration as shown in Figure 11.5[36]. In view of the large wind capacity forecasted for the Iberian Peninsula, the Portuguese TSO ­assessed

192  Advances in Energy Systems

Figure 11.3  Texas Competitive Renewable Energy Zone (CREZ) locations. Table 11.1  Planned electrical circuits in the Spanish transmission system. Planned infrastructure

Specific RES infrastructure

Lines and cables

Total

400 kV

220 kV

Total

400 kV

220 kV

New circuits (km) Refitting (km)

12 656 8308

7488 3850

5168 4458

4465 1730

3504

961

the  transient stability of the system using high‐­ probability scenarios[37], required local voltage reg­ ulation and started to operate a significant part of the wind generation using “Wind Cluster Control ­Centers” with power control capabilities.

German dena Grid Study II (Case E) The German dena Grid Study II  –  initiated by the German Energy Agency [Deutsche Energie‐­ Agentur GmbH (dena)] and published in November

Transmission Planning for Wind Energy in the United States and Europe: Status and Prospects  193

87

252

107

178

44 188

146

64

153

451

171 151 84

178 83

34 34 336

300

48 71

235 203

148 14

222

228 127 115 76

Hydro

105

14

Reversible hydro Wind

Investment –25%

–75%

–50%

–100%

Figure 11.4  RES‐driven transmission lines included in the Portuguese National Transmission Grid plan of invest­ ments (PT RNT) (2006–2010).

2010  –  focuses on the requirements for a reliable power supply system in 2020 when 39% of the gross electricity consumption is assumed to be contributed by renewable energy sources[38]. Within this scenario, onshore and offshore wind energy installations amount to about 49%, or 37 GW, and 18%, or 14 GW, of the total installed renewable energy generation capacities. The requirements for a secure grid integration of all temporally available renewable energies are identified in conjunction with a market‐driven operation of the present power plant fleet and a liberalized European energy market. Apart from the common estimation of the grid extension requirements, different transmission technologies have been evaluated. A special focus lies on investigations of flexible line management (FLM) using line ratings on the basis of actual wind speeds and conductor temperature, and high‐temperature conduc­ tors (e.g. thermal resistant aluminum alloys, or TAL) to increase the transmission capacity of overhead lines in the extra‐high‐voltage grid, as shown in Figure 11.6.

36

Wind power 215

81

Simulated in 2011 for a total of 4500 MW

Figure 11.5  Spatial distribution of the wind power to be injected in the transmission substations.

The large‐scale use of FLM and TAL are not eco­ nomically viable. However, for single sites with high average wind speeds and usually with a large number of installed wind turbines, FLM shows a high poten­ tial to increase the transmission capacity during times with high wind power feed‐in. Furthermore, TAL can be used for individual cases to overcome permanent grid congestion. Transmission Planning in the European North Sea (Case F) Europe is set to build large amounts of offshore wind power, increasing from 2.5 GW today up to 4085 GW in the year 2030; some of it will be located far from

194  Advances in Energy Systems

Routes in km

Billion €/year

7000

7

6000

6

5000

5

4000

4

3000

3

2000

2

1000

1 Basic

FLM

TAL

New overhead line routes Modification of overhead line routes (i.e, structural changes to existing routes) New underground cable routes Costs per year (annualized capital and operating costs) Figure 11.6  Grid extension and annual costs of the different technologies: basic grid with standard transmission capacity (basic); flexible line management (FLM); high‐temperature conductors (TAL). Source: Reprinted with permission from Ref.38. Copyright 2010 German Energy Agency.

shore with the need for long subsea power cables to the onshore power system. At the same time, there is a need to better integrate the power markets in ­Europe by increasing the transnational power exchange capacity. Both developments call for consideration of combining offshore wind power grid connection and interconnections between countries. In December 2010, a memorandum of under­ standing was signed by the 10 countries around the North Sea[15], represented by their energy ministries, their TSOs organized in ENTSO‐E, and their regu­ lators, organized since March 2011 in Association for the Cooperation of Energy Regulators (ACER), and the European Commission, forming together the North Seas Countries Offshore Grid Initiative. The objective of this cooperation is to coordinate efforts toward necessary investigations on technical and grid planning questions[16], as well as identifying mar­ ket and regulatory barriers, which then should be removed as far as possible. The first results on grid development have been published by ENTSO‐E[16]. As a result of ongoing European research, a new optimization tool for transmission expansion planning has been developed[39]. The tool can  –  in contrast to previous models  –  account for the stochastic prop­ erties of wind power distributed over large areas. The capacity expansion planning problem is solved

with regard to several states that together give a good statistical description of the wind power and load variations. The tool explicitly considers the benefit of transmission capacity between differently priced areas and the value of connecting offshore wind power to the grid versus the investment cost of power cables. The outcome is an optimal grid that answers the question of where to build the new transmission lines/cables and with how much capacity. This tool has been applied to a case study of the North Sea region where there exists extensive plans for both off­ shore wind development and new subsea interconnec­ tors between countries[39]. In the study, 33 prospective interconnectors were considered; Figure  11.7 shows the resulting optimal meshed grid. In the case study, wind power was modeled using Reanalysis wind velocity data and regional power curves, adapted from TradeWind. Cross‐Border Transmission in Europe: TradeWind (Case G) As Europe is targeting a 33% share of the electricity demand covered by wind power in 2030[40], cross‐ border transmission capacity needs to be significantly increased in several corridors, bringing significant economic benefits in terms of reduced operational

Transmission Planning for Wind Energy in the United States and Europe: Status and Prospects  195

60°N

3600

58°N 600

3000

600

56°N

1200

1800

6000

4200

54°N

1800 6000

4200 3600

52°N

600 600

3600 2400 2400

50°N

0°

3°E

6°E

9°E

Figure 11.7  Optimal grid example for the North Sea region. Green, optimized interconnectors; red, existing intercon­ nectors; blue triangles, major offshore wind farms.

costs of power generation. This is the conclusion of the TradeWind study[14], after simulating power flows in the European transmission network with the expected wind power capacity deployment scenarios in 2010, 2015, 2020, reaching 300–400 GW in 2030. Increasing wind power capacity in Europe was found to lead to increased cross‐border energy exchanges and more severe cross‐border transmission bottle­ necks in the future, especially with the amounts of wind power capacity in 2020 and 2030. If the 42 identified onshore and offshore cross‐ border transmission upgrades are implemented, oper­ ational costs of power generation would be reduced by €1.5 billion per year after 2030. TradeWind also evaluated the effect of improved power market rules and quantified these in terms of reduction of the oper­ ational costs of power generation. The establishment of intraday markets for cross‐border trade is found to be of key importance for market efficiency in Europe,

as it will lead to savings in system costs in the order of €1–2 billion per year as compared with a situation where cross‐border exchange must be scheduled a day ahead. Consequently, the TradeWind analy­ sis concluded that the European electricity market needs intraday rescheduling of generators and trade, a consolidation of market areas, and increased inter­ connection capacity to enable efficient wind power integration. EWIS Results (Case H) The EWIS[12] is the first time that a year‐round mar­ ket analysis, necessary to represent the effects of wind on a pan‐European basis, has been coupled with detailed representations of the networks, which is necessary to comprehensively address network performance limitations and so ensure reliability and economy. A key recommendation from EWIS is that

196  Advances in Energy Systems

Figure 11.8  Cross‐border links with strong economic expansion benefits identified in the enhanced network scenario – outlook beyond 2015 of the EWIS.

pan‐European modeling, coordinated, and adjusted by more precise regional or national models, should be further developed and used as appropriate to assess future development of the European transmission net­ work, especially as the proportion of wind generation increases. This task is being pursued by ENTSO‐E. Looking beyond the immediate measures to strengthen and make best use of existing networks, EWIS also examined the benefits of enhancing cross‐ border interconnection capacity and identified those links that are likely to have congestion reducing ­benefits that exceed the probable capital costs. These are illustrated in Figure  11.8 (Ref.[12]) and include some 30 links with a total capital cost of c. €12 ­billion. It is likely that strengthening these priority cross‐border interconnections will give fuel savings and CO2 emission benefits exceeding the reinforce­ ment capital costs.

of demand. The TSOs own and operate a grid with a length of 305 000 km lines inside five synchronous zones. According to European regulations, the TSOs are obliged to publish every second year a plan on the next 10 years’ grid development – the Ten‐Year‐Net­ work‐Development‐Plan. A first pilot plan has been published in March 2010[13]. The main drivers behind these transmission planning needs are a lot of wind power in northern Europe, some hydro power in northern and central Europe, and a lot of solar power expected in southern Europe. Additionally, some conventional power plants are being decommissioned, some new will be built, and some demand will change, resulting in chang­ ing flow patterns, leading to a need for transmission lines. Summarizing, there are three main reasons for transmission needs: security of supply, connection of renewables, and implementation of the market.

System Development in the Framework of ENTSO‐E (Case I)

­LOOKING INTO THE FUTURE

According to the EU Regulation 714/2009, a common body of the European TSOs has been installed repre­ senting 42 TSOs from 34 countries. In this area, there is 828 GW generation capacity covering 3400 TWh

In the near term in the United States, meeting the ambitious targets that have been set for renewable energy will require the upgrading of existing lines and the construction of new ones. Because of the long distances and the multiple state and regional

Transmission Planning for Wind Energy in the United States and Europe: Status and Prospects  197

boundaries that must be crossed to move the renew­ able energy to market, as well as the critical national security and long‐term environmental sustainability issues involved, it is clear that there is an appropri­ ate role for the federal government. Legislation has been introduced, which requires interconnection‐wide transmission planning to be performed, an intercon­ nection‐wide cost allocation for high‐voltage back­ bone transmission line costs, and federal backstop authority for transmission line siting. It is not clear whether or when such legislation will be passed, but it is an indication of the growing importance with which the critical need for an expanded transmission infra­ structure is being viewed. Similar discussions are being held in European countries and at the EU level. As in the United States, there is a growing consensus at the political levels that increased transmission is essential for reaching the renewables targets, and that there is a strong role for a coherent European policy. Traditionally, cross‐ border transmission planning at a European scale was linked to the development of a single internal market for electricity. More recently, the European Commission has produced a new “European Energy Infrastructure Package” to facilitate the realiza­ tion of the renewables targets of the Commission, 20% renewables by 2020, including a blueprint for an offshore grid in Northern Europe. The European Commission is implementing the Third Liberaliza­ tion Package legislation involving both a stronger cooperation of European TSOs through ENTSO‐E, as well as a stronger cooperation between European regulators in the new ACER. The overall policy framework is focused on achieving competitive, sus­ tainable, and secure electricity supply in a single electricity market. One of the very difficult issues, namely, how to finance and recover costs of trans­ national transmission against a diversified backdrop of regulatory frameworks, is part of today’s ongoing discussions between these stakeholders.

­CONCLUSIONS There is a growing recognition around the world that wind energy is different from more conventional sources of energy and requires a different approach to transmission planning. Traditional capacity‐based methods must be modified and expanded to incor­ porate the unique characteristics of wind as an energy resource with limited capacity attributes and value. Regional approaches to transmission expan­ sion planning have unleashed a number of creative approaches to planning and building transmission for

wind. Transmission has become recognized as a key enabler to reach renewable energy goals and carbon reduction goals. As a consequence of this, policy ini­ tiatives are underway in the United States and Europe to catalyze the process of transmission expansion for a sustainable energy supply at a continental level. The scale of transmission development envisioned for this purpose will require unprecedented coop­ eration across multiple jurisdictional boundaries to develop a set of coherent and collaborative transmis­ sion planning, siting, and permitting policies and cost allocation method necessary to achieve the intended objectives.

­ACKNOWLEDGMENTS The authors would like to acknowledge the con­ tributions of their many professional colleagues working in the wind integration and transmission planning field, who have contributed to the thinking and progress ­reported here. This work was part of the International Energy Agency Wind Task 25 wind integration collaboration.

REFERENCES 1. U.S Wind MARKET update 2010. AWEA. Available at: http://www.awea.org/leamabout/publications/factsheets/ factsheets_windenergybasics.cfm, (Accessed December 20, 2011.) 2. Wind in Power: 2010 European Statistics. EWEA, February 2011. Available at: http://ewea.org/fileadmin/ ewea_documents/documents/statistics/EWEA_Annual_ Statistics_2010.pdf. (Accessed December 20, 2011.) 3. ENTSO‐E Position Paper. Addressing transmission‐ related challenges of renewable energy integration. ENTSO‐E RES Coordinating Committee, November 2011. Available at: https://www.entsoe.eu/resources/ position‐papers. (Accessed December 20, 2011.) 4. Energy Industry Act, Section 17, Subsection 2a. Available at: http://www.gesetze‐im‐internet.de/enwg2005/17.html. (Accessed December 20, 2011.) 5. Offshore Networks: The Connection of Offshore Parks to the National Grid. Dena Fact Sheet 2. German Energy Agency, April 2009. Available at: http://www. dena.de/fileadmin/user_upload/Download/Dokumente/ Publikationen/ESD/02_eng_Offshore_networks.pdf. (Accessed December 20, 2011.) 6. System Adequacy Methodology. UCTE, 2009. Available at: https://www.entsoe.eu/resources/publications/former‐ associations/ucte/system‐adequacy. (Accessed December 20, 2011.) 7. Annual Transmission Expansion Plan (MTEP). MISO, 2011. Available at: https://www.misoenergy.org/Planning/ TransmissionExpansionPlanning/Pages/MTEP11.aspx. (Accessed December 20, 2011.)

198  Advances in Energy Systems

8. Bhavaraju, M.P. and Billinton, R. (1971). Transmission planning using a reliability criterion part 11 – transmission planning. IEEE Trans. Power Apparatus Syst. PAS‐90: 70–78. https://doi.org/10.1109/TPAS.1971.292900. 9. Platts, J.E. and Womeldorff, P.J. (1980). The significance of assumptions implied in long‐range electric utility plan­ ning studies. IEEE Trans. Power Apparatus Syst. PAS‐99: 1047–1056. https://doi.org/10.1109/TPAS.1980.319735. 10. Fulli G, Ciupuliga A, Abbate A, et al. Review of existing methods for transmission planning and for grid connection of wind power plants. Realisegrid Deliverable3.1.1, 2009. Available at: http://realisegrid.rse‐web.it/Publications‐ and‐results.asp. (Accessed December 20, 2011.) 11. Joint Coordinated System Plan. Joint coordinated sys­ tem plan, 2008. Available at: https://www.midwestiso. org/Planning/Pages/StudyRepository.aspx. (Accessed December 5, 2011.) 12. European Wind Integration Study. EWIS Final Report, March 2010. Available at: http://www.wind‐integration. eu/downloads. (Accessed December 5, 2011.) 13. 10 Year Network Development Plan 2010–2020. ENTSO‐E, June 2010. Available at: https://www.entsoe. eu/fileadmin/user_upload/_library/SDC/TYNDP/ TYNDP‐final_document.pdf. (Accessed December 5, 2011.) 14. EU‐IEE Project TradeWind. Available at: http://www. trade‐wind.eu. (Accessed December 5, 2011.) 15. North Sea Countries Offshore Grid Initiative – Memorandum of Understanding, 2010. Available at: https://www.entsoe.eu/fileadmin/user_upload/_library/ news/MoU_North_Seas_Grid/101203_MoU_of_the_ North_Seas_CountriesOffshore_Grid_Initiative.pdf. (Accessed December 20, 2011.) 16. ENTSO‐E Views on the Offshore Grid Development in the North Seas. ENTSO‐E, February 2011. Available at: www.entsoe.eu. (Accessed December 20, 2011.) 17. Woyte A, De Decker J, Thong VV. A North Sea electric­ ity grid [Revolution: electricity output of interconnected offshore wind power, a vision of offshore wind power integration]. Greenpeace ‐3E, 2008. Available at: www. greenpeace.org. (Accessed December 5, 2011.) 18. Fichaux N, Wilkes J, Van Hulle F, et  al. Oceans of opportunity. Harnessing Europe’s largest domestic energy resource. EWEA 2009, Available at: http://www. ewea.org/fileadmin/ewea_documents/documents/ publications/reports/Offshore_Report_2009.pdf. (Accessed December 5, 2011.) 19. Burke DJ. Accommodating wind power characteristics in power transmission planning applications. PhD Thesis, University College Dublin, Dublin, Ireland, 2010. Available at: http://erc.ucd.ie/files/theses/Daniel%20 Burke%20‐% 2 0 Ac c o m m o d a t i n g % 2 0 Wi n d % 2 0 Energy%20Characteristics%20in%20Power%20 Transmission%20Planning%20Applications.pdf. (Accessed December 5, 2011.) 20. Burke, D.J. and O’Malley, M.J. (2011). A study of opti­ mal non‐firm wind power connection to congested trans­ mission networks. IEEE Trans. Sustainable Energy 2: 167–176. https://doi.org/10.1109/TSTE.2010.2094214.

21. Padhy, M.P. (2004). Unit commitment – a bibliographi­ cal survey. IEEE Trans. Power Syst. 19: 1196–1205. https://doi.org/10.1109/TPWRS.2003.821611. 22. Latorre, G., Cruz, R.D., Areiza, J.M. et  al. (2003). Classification of publications and models on transmis­ sion expansion planning. IEEE Trans. Power Syst. 18: 938–946. https://doi.org/10.1109/TPWRS.2003.811168. 23. Romero, R., Monticelli, A., Garcia, A. et  al. (2002). Test systems and mathematical models for transmis­ sion network expansion planning. IEE Proc. Gener. Transm. Distrib. 149: 27–36. https://doi.org/10.1049/ ip‐gtd:20020026. 24. L’Abbate A, Losa I, Migliavacca G, et  al. D3.3.1 Possible criteria to assess technical‐economic and strategic benefits of specific transmission projects. ­ Realisegrid Deliverable. Available at: http://realisegrid. rse‐web.it. (Accessed December 5, 2011.) 25. Ciupuliga AR, Gibescu M, Pelgrum E, et al. Round‐the‐ year security analysis with bottleneck ranking for intercon­ nected power systems with large‐scale wind power. In:  IEEE PES Conference on Innovative Smart Grid Technologies Europe. Gothenburg; October 11–13, 2010. doi:https://doi.org/10.1109/ISGTEUROPE.2010.5638973. 26. Eastern Wind Integration and Transmission Study Website, Available at: http://www.nrel.gov/wind/systemsintegration/ ewits.html. (Accessed December 5, 2011.) 27. Analysis of Transmission Alternatives for Competitive Renewable Energy Zones in Texas. Electric Reliability Council of Texas, 2006, Available at: http://www.ercot. com/news/presentations/2006/ATTCH_A_CREZ_ Analysis_Report.pdf. (Accessed December 5, 2011.) 28. Labra P. Regional Group Continental South West (RG CSW). Results. In: ENTSO‐E Workshop Ten‐year network development plan. Madrid, Spain, 29 ­ November  2011. Available at: https://www.entsoe.eu/ events/workshops. (Accessed December 5, 2011.) 29. Clavero A, Martínez S, Rodrígruez‐Bobada F. Estudios de integración de generación de régimen especial en 2016 en el sistema eléctrico peninsular espabol. In: XIII Encuentro Regional Iberoamericano de Cigre. Puerto de Iguazb; May 24–28, 2009. 30. Estudio de estabilidad eólica de la península Ibérica. Síntesis de criterios y metodologías. REE/REN; May, 2005. 31. Rodríguez‐Bobada F, Reis Rodriguez A, Ceña A, et al. Study of wind energy penetration in the Iberian penin­ sula. In: European Wind Energy Conference. Athens; February 27–March 2, 2006. 32. Criterios generales de protección del sistema eléctrico peninsular español. Red Eléctrica y empresas elec­ tricas, 1995. Available at: http://www.ree.es/ publicaciones/publicaciones_on_line.asp. (Accessed December 5, 2011.) 33. Rodríguez‐Bobada F, Ledesma P, Martínez S, et  al. Simplified wind generator model for transmission sys­ tem operator planning studies. In: 7th International Workshop on Large‐Scale Integration of Wind Power into Power Systems as well as on Transmission Networks for Offshore Wind Farms. Madrid; May 26–27, 2008.

Transmission Planning for Wind Energy in the United States and Europe: Status and Prospects  199

34. Plano de investimento e desenvolvimento da rede de transporte 2009–2014, 2019. REN‐Rede Electrica Nacional, February 2008, 243. Available at: www.ren.pt. (Accessed December 5, 2011.) 35. Simoes T, Costa P, Estanqueiro A. A methodology for the identification of the sustainable wind potential. The  Portuguinese case study. In: IEEE/PES Power Systems  Conference and Exposition. Seattle, WA; March 1518,  2009. doi:https://doi.org/10.1109/ PSCE.2009.4839951. 36. Plano de investimento da rede de transporte 20062011, Vol. 1. REN‐Rede Electrica Nacional, November 2005 180. Available at: www.ren.pt. (Accessed December 5, 2011.) 37. Sucena Paiva JP, Ferreira de Jesus JM, Castro R, et al. Transient stability study of the Portuguese transmission network with a high share of wind power. In: XI Encuentro Regional Iberoamericano de Cigré. Her‐ nandarias; May 22–26, 2005. 38. Dena Grid Study II. Integration of renewable energy sources in the German power supply system from 2015– 2020 with an outlook to 2025. Summary of the main results by the project steering group. German Energy Agency (dena), Berlin, 2010. Avaialable at: http://www. dena.de/fileadmin/user_upload/Download/Dokumente/ Studien_Umfragen/Summary_dena_Grid_Study_II.pdf. (Accessed December 5, 2011.) 39. Trötscher, T. and Korpås, M. (2011). A framework to determine optimal offshore grid structures for wind power integration and power exchange. Wind Energy 14: 977–992. https://doi.org/10.1002/we.461. 40. Pure Power: Wind Energy Targets for 2020 and 2030. European Wind Energy Association (EWEA), 2011.

Avaialbale at: www.ewea.org. (Accessed December 5, 2011.)

FURTHER READING Ackermann, T. (2005). Wind Power in Power Systems. New York: Wiley. Burton, T., Sharpe, D., Jenkins, N. et al. (2001). Wind Energy Handbook. New York: Wiley. Christie, R.D., Wollenberg, B.F., and Wangensteen, I. (2000). Transmission management in the deregulated environ­ ment. Proc. IEEE 88: 170–195. ENTSO‐E. Definitions of Transfer Capacities in Liberalized Electricity Markets. Available at: www.entsoe.eu. (Accessed May 17, 2012). Grainger, J. and Stevenson, W. (1994). Power System Analysis. New York: McGraw‐Hill. Kundur, P. (1997). Power System Stability and Control. New York: McGraw‐Hill. Momoh, J. and Mili, L. (2010). Economic Market Design and Planning for Electric Power Systems. New York: Wiley‐Blackwell. NERC. Documentation of Total Transfer Capability and Available Transfer Capability Calculation Methodologies, Standard MOD‐001–0. Princeton: North American Electric Reliability Corporation; 2005. Seifi, H. and Sepasian, M. (2011). Electric Power System Planning: Issues, Algorithms and Solutions. Germany: Springer‐Verlag. Sullivan, R. (1977). Power System Planning. New York: McGraw‐Hill. Zima, M. and Andersson, G. (2005). On security criteria in power systems operation. Power Eng. Soc. Gen. Meeting.

12

Opportunities and Barriers of High‐Voltage Direct Current Grids: A State‐of‐the‐Art Analysis Debora Coil‐Mayor1 and Jürgen Schmid2 SMA Solar Technology AG, Kassel, Germany Fraunhofer Institute for Wind Energy Systems, Kassel, Germany 1

2

Limiting the consequences of the climate change means that global CO2 emissions are to be reduced very rapidly. Taking into account that the expected economic growth of developing countries will inevitably be accompanied by a further growth in emissions in these countries, there is a strong necessity that the energy systems in the industrialized world be transformed into emission‐free ones. This level of reduction requires, for example, that the European Union achieves a “nearly zero‐carbon power supply,” involving replacing of fossil fuels in most or all buildings and much of the transport sector by nearly zero‐carbon electricity. It is also clear that the current political framework is, especially in Europe, propitiating the large‐scale implementation of renewable energy sources (RES). In order to make this large implementation economically feasible, it is essential that the potential of these sources is maximized, placing them at sites providing the highest possible output. These sites are often located far away from consumption centers, making compulsory the transmission of large amounts of energy from generation to demand. In the global scenario, the necessary capacity of the power infrastructure for doing such exceeds by far the existing infrastructure and therefore new transmission

capacity is necessary. Taking into account that large amounts of RES are to be integrated in the current power system and that most of the generation will take place at remote locations requiring new transmission assets, this chapter answers the question of whether a supergrid is necessary to enable this transformation. ­INTRODUCTION Limiting the consequences of the climate change means that global CO2 emissions are to be reduced very rapidly. By the mid‐century, these emissions must be reduced by 50% compared with the 1990 levels.1 Taking into account that the expected economic growth of developing countries will inevitably be accompanied by a further growth in emissions in these countries, there is a strong necessity that the energy systems in  “Global greenhouse gas emissions need to be reduced by at least 50% below 1990 levels by 2050” was a declaration issued by more than 200 international climate scientists urging politicians at the United Nations Climate Change Conference in Bali to agree on strong targets for tackling climate change (http://unfccc.int/meetings/cop_13/ items/4049.php).

1

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

202  Advances in Energy Systems

the industrialized world be transformed into emission‐ free ones. In 2009, the European Union (EU) and the G8 Heads of Government committed their countries to an 80% reduction in greenhouse gas emissions by 20502. This level of reduction requires the EU to achieve a nearly zero‐carbon power supply, involving replacing of fossil fuels in most or all buildings and much of the transport sector by nearly zero‐carbon electricity[1]. It is also clear that in Europe there is a political consensus that renewable energy will play a major role in such a transformation[2]. Replacing fossil fuels, increasing penetration of renewables, and targeting a transport sector with nearly zero‐­ carbon emissions present an enormous challenge for the current power infrastructure, which will have to assume large amounts of intermittent generation, bidirectional fluxes in the power network, and so on, without decreasing the quality and security of supply. On the contrary, for an economically optimized energy supply system, the individual renewable sources such as wind, solar, and biomass must be placed at sites providing the highest possible output. Most of these sources still have a high unused potential, for example, the European Wind Energy Association foresees 230 GW of wind power in 2020 (including 40 GW offshore, and even 265 GW in their optimistic scenario)[3]. The capacity of the necessary infrastructure exceeds the capacity of existing grids by far in terms of maximum power as well as in terms of distance. Therefore, in some areas, the limits to transmission capacity due to including renewable energy are already being reached, and potential clean energy production is thus being curtailed. In some cases, this limitation in the transmission capacity can be overcome using soft measures, but in many cases, the construction of new transmission lines (also called hard measures) is unavoidable[4]. New technical challenges for the inclusion of large amounts of distributed and/or intermittent generation, along with the necessity for the construction of new transmission lines, make imperative the creation of a superstructure upon existing regional transmission grids with abilities exceeding those of the current power grids. This superstructure is called a supergrid. Since wind and solar power fluctuate, a second function of such a supergrid is to provide a balance between different locations and different sources. A third function of a supergrid is to make accessible the existing hydropower storage capabilities such as the enormous facilities located in Norway, which, to date, cannot be fully exploited because of the lack of adequate interconnections.  G8 summit in L’Aquila (Italy). July 2009.

2

In the European case, the capabilities of this new supergrid[5] have already been defined: (i) accommodating ever‐increasing wind surplus generation in and around the Northern and Baltic Seas and increasing renewable generation in the East and South of Europe and also North Africa; (ii) connecting these new generation hubs with major storage capacities in Nordic countries and the Alps and with the major consumption centers in Central Europe; and (iii) coping with an increasingly flexible and decentralized electricity demand and supply. This article defines priority corridors, describes how these corridors can be constructed (the supergrid vision), identifies the barriers, and facilitators for this supergrid and presents conclusions.

­ RIORITY CORRIDORS: LINKING LARGE P RENEWABLE ENERGY SOURCES (RES) GENERATION WITH CONSUMPTION CENTERS Priority corridors also known as green corridors are interregional power transmission networks able to transmit large amounts of power from renewable source generation to large consumption centers and massively strengthen existing cross‐border interconnections. Priority corridors are seen in Europe as absolutely imperative in achieving the policy goals of increasing security of supply, integrating renewables, and supporting the proper functioning of the energy market[6]. Contrary to this, in the United States, in identifying priority corridors, limited attention to the long‐term drivers of electricity generation such as renewable energy facilities has been given[7]. The drivers in the United States case are mostly based on the need for upgraded and new electricity transmission and distribution facilities to improve reliability, relieve congestion, and enhance the capability of the national grid to deliver electricity[8]. In China, priority corridors are those supplying the megacities with hydropower production from the remote areas. The priority corridors are dispersed in all the regions of the world, linking, for example, large desert areas or offshore wind potential areas to urban areas. As an example, the priority corridors identified in the European case are provided. Case Study: Europe The European Network of Transmission System Operators for Electricity (ENTSOE) in its Development plan[9] identifies the generation and demand development tendencies in Europe, therefore deter-

Opportunities and Barriers of High‐Voltage Direct Current Grids: A State-of-the-Art Analysis  203

Wind energy Solar energy Pump storage Marine energie Offshore super grid Transnational electricity transmission system

Figure 12.1  Potential of green generation for Europe determining the future transmission needs. Source: Fraunhofer‐ Institut für Windenergie und Energiesystemtechnik (IWES).

mining the future transmission needs. Taking this information into account, the European Commission proposes to focus attention on the following priority corridors, which will make Europe’s electricity grids fit for 2020[5]: • Offshore grid in the Northern Seas and connection to northern as well as central Europe. This first priority corridor would integrate and connect energy production capacities in the Northern Seas with consumption centers in northern and central Europe and hydro storage facilities in the Alpine region and in Nordic countries. • Interconnections in southwestern Europe to accommodate wind, hydro, and solar. This second priority corridor would, in particular, strengthen the connection between the Iberian Peninsula and France, and further connect with Central Europe, to make best use of Northern African renewable energy sources (RES) and the existing infrastructure between North Africa and Europe. • The DESERTEC concept is framed in this priority corridor. • Connections in central eastern and southeastern Europe. This third‐priority corridor would strengthen the regional network in North−South and East−West power flow directions, in order to assist market and renewables integration, including connections to storage capacities and the integration of energy islands.

• Completion of the Baltic Energy Market Interconnection Plan. This fourth priority corridor aims to improve the integration of the Baltic States into the European market through reinforcement of their internal networks and strengthening the interconnections with Finland, Sweden, and Poland and through reinforcement of the Polish  internal grid and interconnections east and westward. A proposed configuration of those priority corridors performed by the Fraunhofer‐Institut für Windenergie und Energiesystemtechnik (IWES)3 is given in Figure 12.1. ­THE SUPERGRID VISION The supergrid vision is based on the following points: • The supergrid is an interregional supertransmission grid able to transmit large amounts of energy from remote locations at ultrahigh voltages (UHVs). • The supergrid is based on high‐voltage direct current (HVDC) technologies, using multiterminal technologies, sometimes referred to “supernodes,” together with standard point‐to‐point connections.  Schmid J. Konzept fur eine 100% regenerative Energieversorgung in 2050 und die besondere Rolle der Bioenergie (Concept for a 100% renewable energy supply in 2050 and the special role of bioenergy).

3

204  Advances in Energy Systems

• The supergrid transports large amounts of renewable energy from wherever these sources are abundant to the best available markets, for example, energy from solar thermal power stations in ­deserts and other sun‐rich areas or wind energy generated in wind‐rich coastline areas to distant consumption centers. • The supergrid provides a steady supply, balancing the intermittent effect of the renewables’ sources across a wide region, integrating different types of sources. • The supergrid makes use of the potential hydroelectric capacity as a backup for wind and solar power, to ensure reliable supply. • The supergrid is an ally for the new developments in power system configurations at distribution level, such as smartgrids, virtual power plants, mini or microgrids, and so on and therefore the integration of all the technologies related to them: demand side management, vehicles to grid, and so on. Examples of supergrid success initiatives are now given.

Supergrid Initiatives: Examples In this section, three examples are presented: the North Sea supergrid initiative in Europe, the DESERTEC initiative linking North Africa and Europe, and the hydropower case in China.

energy installations with a further 2 GW. Most of this development is expected to occur in the countries participating in the initiative, as well as using the hydro storage capacity available in Norway. In a study performed by the Friends of the Supergrid, a two‐phase project is defined. The first phase would connect the United Kingdom, Germany, Belgium, Norway, and the Netherlands. This study also provides a financing model for this first phase, assuming a total cost of €28 billion (2010 value), a building time of 6 years, and an operation of 40 years. It would result in a transmission charge from 4.67 cents to as low as 1.55 cents per kilowatt‐hour (c kWh−1), depending on gearing and utilization[12]. An example of the possible configuration of this supergrid can be found in Figure 12.2.

DESERTEC Initiative DESERTEC is allowing people in populated areas in Europe to access solar power from the energy‐rich desert areas of North Africa[14] (see Figure  12.3). DESERTEC core technologies are generation based on concentrated solar power (CSP) and transmission based on HVDC across the sea. The analysis of possible scenarios of the electricity supply system until 2050, taking into account present and expected demand for electricity and power capacity, available renewable energy resources and their applicability for power, and socioeconomic and environmental impacts is provided in Ref.[16].

North Sea Supergrid In 2008, the European Commission identified in its 2008 Second Strategic Energy Review the need for a coordinated strategy concerning offshore grid development. In its communication COM(2008) 781[10], it stressed the necessity of a blueprint for a North Sea offshore grid and underlined that this grid would be one of the building blocks of a future European Supergrid. One year later, at the end of 2009 and beginning of 2010, nine EU countries (Belgium, Denmark, France, Germany, Ireland, Luxembourg, the Netherlands, Sweden, and the United Kingdom) plus Norway signed a political declaration on the North Seas Countries Offshore Grid Initiative. In December 2010, the ministers of the North Seas Countries’ offshore grid initiative signed a memorandum together with the European Commissioner for Energy[11], with the objective of coordinating the offshore infrastructure in North Europe. The target objective is to achieve an installed capacity of over 40 GW of offshore wind by 2020, in addition to other marine renewable

Hydropower in China In December 2009, the first HVDC of 800 kV[17] was operative. This supertransmission line goes over 1400 km and transmits a maximum power of 5 GW, connecting the hydropower production from the remote Yunnan province and the rapidly growing industrial region in the Pearl River delta in Guangdong province with its megacities Guangzhou and Shenzhen (see Figure 12.4).

­BARRIERS AND FACILITATORS The analysis of barriers and facilitators has been done based on the following issues: technological, economical/financial, and political/sociopolitical. These issues affecting the wide implementation of the supergrid have been identified by the ENTSOE in its “Ten‐Year Network Development Plan 2010–2020”[9] and are broadly accepted.

Source: map BSH, grafic overlay B, Valov Fraunhofer IWES

Norway

West

East

Center Great Britain

2 GW

Denmark 2 GW

1 GW The Netherlands

Böxlund

Status: March 2011 Cluster of wind farms Point of common coupling on land German offshore power centers AV

Test field alpha ventus in operation

BI

Test field bard offshorel in operation

BI

Bûsum AV

11 5 GW

Approved wind parks Planned wind parks Wind parks not relevant for clustering

31 3 GW

Cuxhaven

Norden

Internal connection in cluster Wilhelmshaven

External connection of cluster International power transmission system

Brunsbûttel

Emden

Bremerhaven

Conneforde

Diele

Figure 12.2  A possible configuration of the North Sea Supergrid provided by the “IWES‐Konzept 2010”: Eine Vision zum zukünftigen Offshore – Übertragungssystem in der Nordsee [13]. Source: Fraunhofer‐Institut für Windenergie und Energiesystemtechnik (IWES).

Figure 12.3  Concept of a “EUMENA Supergrid” based on HVDC power transmission as “Electricity Highways” to complement the conventional AC electricity grid, as developed by TREC in 2003. The symbols for power sources and lines are only sketching typical locations see reference [15]. Source: DESERTEC foundation.

206  Advances in Energy Systems

Figure 12.4  Projected supertransmission lines linking the hydropower stations in the Yunnan province with the megacities Guangzhou and Shenzhen [18]. Source: SIEMENS AG.

Technological Issues Topology Definition The topology of the supergrid is still being defined and several solutions are emerging (some examples can be found in Refs.[15, 19, 20]). The only matter clear to date is that the supergrid will be based on DC technology. UHV alternating current (AC) grid is less desirable because of losses in the cables as well as the investment costs of AC versus DC over long distances. Other advantages of DC compared  with AC are the higher controllability and that there is no necessity for synchronization of the interconnected grids. There are also nontechnical reasons behind this decision such as the environmental impact in terms of visual pollution as well as the electromagnetic fields generated by AC grids. Reference[21] provides three different main alternatives for the supergrid topology: (i) point‐to‐point technology

and several tappings (simple solution); (ii) a meshed DC grid with two converter stations for every DC line (expensive solution but fully controllable); and (iii) a meshed DC grid without converters between DC lines (fully DC transmission grid with redundancy). Since for the development of the supergrid a meshed DC grid is essential, topology (ii) and (iii) are the most suitable. Converter Technology Since all the distribution will remain based on AC, conversion technologies for DC to AC are required. Two kinds of power converters are available for these applications: the current source converter (CSC) and the voltage source converter (VSC). CSC technology based on thyristors, also known as line commutated converters (LCCs), is the most common and widely applied technology. On the contrary, VSC based on

Opportunities and Barriers of High‐Voltage Direct Current Grids: A State-of-the-Art Analysis  207

under DC regime poses major scientific and technological problems[30]. Higher voltage ratings (up to 500 kV) for DC XLPE cables are under development[31].

insulated gate bipolar transistors (IGBTs) has the ability to connect weak AC networks or even dead networks (black‐start capability) and therefore is the most viable solution for offshore wind parks. The reason for this ability is that VSCs utilize self‐commutating switches such as gate turn‐off thyristors (GTOs) or IGBTs, which can be turned on or off at will[22]. This is in contrast to the conventional CSCs, which operate with line‐commutated thyristor switches. The only barrier is that VSC technology is still under development with the first plant taking up operation in 1996[23]. LCC, however, was implemented earlier, during the 1970s.

There is currently no standard for DC voltage levels. Although the technical committee 115: “HVDC transmission for DC voltages over 100 kV” of the International Electrotechnical Commission is working on a standard for HVDC technologies, Ref.[24] proposes a voltage level of 600 kV as being the most suitable for this application.

Configurations

Power Quality

The only current mature configuration suitable for the supergrid is the point‐to‐point scheme. However, this configuration does not permit multiple points of connection. As the ideal topology should have a high degree of meshing, a multiterminal scheme is required. The multiterminal scheme based on VSC is the configuration widely accepted today as being the most suitable for the supergrid[24]. This configuration is, unfortunately, not based on mature technologies. Therefore, several technical challenges have to be faced, including HVDC circuit breakers, protection systems for interconnected cable circuits, and fast‐ acting control systems to manage power flows in the event of a network component failure[25].

VSC‐based HVDC systems possess the advantage of  being able to control active and reactive power independently and can supply a passive network; these advantages are discussed in Refs.[32] and[33]. ­Reference[34] carried out different case studies, the results of which verified the effectiveness of the multiterminal HVDC in improving power quality.

Voltage Level

Level of Reliability

The level of reliability is a matter of concern for the different actors involved in the energy chain. Different studies have been performed concerning the reliability of VSC‐based HVDC. Since the technology is not yet mature, most of these studies are based on numerical studies of defined models along with sensiCables tivity analyses[35], and some of them are based on the Two types of cable systems are commercially avail- performance of current HVDC technologies (LCC)[36]. able for the transmission of HVDC, based on paper‐­ The results indicate similar levels of availability of insulated cables and extruded cables[26]. Different power transfer capacity for HVAC and HVDC current types of paper‐insulated cables exist, but the mass‐ technologies while having similar levels of redunimpregnated cables have been the most popular since dancy. In the case of HVDC based on VSC, however, its introduction in the 1950s. However, they have the the reliability level is expected to be higher because of disadvantage of rather long joining time, which can be the lack of commutation failure problems in the opera drawback for long distances. Extruded cables can ation of VSC converters. use prefabricated joints, considerably reducing  the installation time, and they are used for up to 500 kV[27]. Insulated HVDC cables do not have limitations Economical/Financial Issues in the length of the cable due to the capacitive reacCost−Benefits Analysis tive current component. However, insulation‐related thermal problems pose a limitation on the power The costs of a supergrid are dependent on the amount capability of HVDC cables[28]. Therefore, beyond a of capacity to be built, the distance involved, and the critical voltage limit, the failure of the cable insulation target region. The supergrid should be justified on a is unavoidable[29]. State‐of‐the‐art polymeric isolation cost−benefit approach, the potential benefits being: for DC cables under high electric stress can suffer the (i) increased competition leading to lower energy degradation of polymer insulation. This phenomenon prices; (ii) increased security of supply; (iii) increased is associated with the formation of space charge. In amount of renewable resources in the generation mix; fact, space charge accumulation in polymeric materials and (iv) accelerated decarbonization.

208  Advances in Energy Systems

Financial Scenario

Social Acceptance/Public Support of the Supergrid

Electricity consumers will most likely take over the costs of the supergrid, either directly (through the transmission tariffs) or indirectly (through recharging transmission tariffs levied on the suppliers), but the cost−benefit approach would ensure that consumers receive an overall benefit through, for instance, reduced generation prices due to increased competition and improved operation efficiency. The cost of new grids is likely to be significantly higher than the cost of the transmission assets existing today. The main reason for this being that they were built in a time when there was much greater public acceptance of the benefits of overhead high‐voltage lines. New lines may need total or significant undergrounding (with the associated extra expense) in order to be accepted and approved.

Social acceptance is highly dependent on increasing public awareness of the need for a transmission infrastructure, with emphasis on all the benefits shown in the above sections: increase in power quality, reinforcement of the energy market, increase in system reliability, and so on. The supergrid represents a strong support for the integration of large amounts of renewable sources in the power system. While obviously being a positive factor, the installation of new UHV lines is a cause of serious environmental concern, which could reduce acceptance.

Reduction in Energy Prices for Consumers and Businesses Logic dictates that a greater interconnection will provide greater market coupling and therefore reduce the electricity prices. This will provide more stable prices for customers. For generators, it will provide increased market access and more certainty for investors and therefore assist in delivering the new renewable and low‐carbon generation. Overall, these effects should deliver the low‐carbon electricity, which is needed at the lowest possible cost. The expected increase in grid security and security of supply will have an impact on consumer confidence, decreasing their necessity to rely on backup technologies. This will obviously indirectly reduce the price of electricity for the consumer. Ownership of the Supergrid The ownership of the supergrid is still an open question. In the European case, ENTSOE seems to be the most likely owner of the supergrid. In the point of view of Ref.[12], ownership of the Supergrid should fall under a variety of interested investors and not just on existing transmission system operators (TSOs). Political/Sociopolitical Issues Legal and Regulatory Framework Regulation and planning are critical issues for a successful supergrid implementation, and many of the necessary requirements are not fully in place. For example, at the European level, the European Regulator currently has insufficient authority and national regulations remain largely independent in issues such as transmission planning[37].

Governmental Support Governmental support is a prerequisite for the success of a project of this magnitude. Since the supergrid crosses country borders, not only the support of one government is necessary but also political developments between neighboring countries are needed in order to create a proper atmosphere of understanding and confidence between nations. Identifying Barriers and Facilitators Taking into consideration the issues identified above, it becomes evident that technical challenges are no longer a barrier when compared with regulatory, financial, political, and acceptance issues. On the contrary, the availability of technical solutions is a great facilitator to supergrid implementation, since initial steps and projects are undoubtedly feasible using current technology. More advanced concepts (e.g. wide‐area offshore meshed grids) will require demonstration projects, and first projects using the new technology might not be cost‐effective. Inside regulatory barriers, standardization issues (e.g. of HVDC voltages) will need to be weighed against the risk of stifling innovation and giving potential advantage to one manufacturer over another. Social acceptance, governmental support, as well as financial issues today present the greatest barriers.

­CONCLUSION The current political framework is, especially in ­Europe, propitiating the large‐scale implementation of RES. In order to make this large implementation economically feasible, the maximization of the potential of these sources placing them at sites, providing the highest possible output is essential. These sites are often located far away from consumption centers, making compulsory the transmission of large

Opportunities and Barriers of High‐Voltage Direct Current Grids: A State-of-the-Art Analysis  209

amounts of energy from generation to demand. In the different analyzed environments (mostly Europe, Asia, and United States), the necessary capacity of the power infrastructure for doing such exceeds by far the  ­existing and, therefore, new transmission capacity is necessary. These necessary new transmission capacities are often referred to as priority ­corridors. Different technologies are being considered for these priority corridors, but a common vision is emerging. This vision is the supergrid, a supertransmission HVDC grid able to transmit large amounts of energy from remote locations at UHVs, using multiterminal technologies together with standard point‐to‐ point connections, balancing the effect of the different renewable sources and making massive use of hydro storage capacities. This supergrid faces several barriers, mostly nontechnical, but it also has great facilitators as the availability of technical solutions for its implementation.

­REFERENCES 1. Boot PA, van Bree B. A zero‐carbon European power system in 2050: proposals for a policy package; 2010. Available at: http://www.roadmap2050.eu/attachments/ files/Roadmap2050ECNworkingdocument.pdf. 2. Newsroom European Commission. Commission proposes an integrated energy and climate change package to cut emissions for the 21st Century. Brussels, Belgium; 2007. Available at: http://europa.eu/rapid/ pressReleasesAction.do?reference=IP/07/29. 3. European Wind Energy Association. Pure Power. Wind energy targets for 2020 and 2030; 2009. Available at: http://www.ewea.org/fileadmin/ewea_documents/ documents/publications/reports/Pure_Power_Full_ Report.pdf. 4. Akermann, T. (2005). Wind Power in Power Systems. Chichester, UK: Wiley. ISBN: 0‐47085508‐8. 5. European Commission and Directorate General for Energy (2011). Energy Infrastructure. Priorities for 2020 and Beyond—a Blueprint for an Integrated European Energy Network. European Union Press. 6. Newsroom European Commission. Energy infrastructure: commission proposes EU priority corridors for power grids and gas pipelines. Brussels, Belgium; 2010. Available at: http://setis.ec.europa.eu/newsroom‐items‐ folder/commission‐proposes‐eu‐priority‐corridors‐ for‐power‐grids‐and‐gas‐pipelines. 7. Vajjhala SP, Paul AC, Sweeney R, et al. Green Corridors: linking interregional transmission expansion and renewable energy policies; 2008. RFF Discussion Paper No. 08–06. Available at: http://www.rff.org/documents/RFF‐ DP‐08–06.pdf. (Accessed June 23, 2011). 8. US Department of the Interior. Record of Decision (RoD) for Designation of Energy Corridores in the 11  Western States. Ref.: BLM/WO‐GI‐09–005–1800. January 2009. Available at: http://corridoreis.anl.gov/

documents/docs/Energy_Corridors_final_signed_ ROD_1_14_2009.pdf. 9. ENTSOE. Ten‐year Network Development Plan 20102020. June 2010. Available at: https://www.entsoe. eu/fileadmin/user_upload/_library/SDC/TYNDP/ TYNDP‐final_document.pdf. 10. COM (2008) 781 of the European Commission to the  European Parliament, the council, the European Economic and Social committee and the committee of the regions. Second Strategic Energy Review: An EU Energy Security and Solidarity Action Plan. Brussels, Belgium. 11. The North Seas Countries’ offshore grid initiative. Memorandum of understanding. Available at: http://ec. europa.eu/energy/renewables/grid/doc/north_sea_ countries_offshore_grid_initiative_mou.pdf. (Accessed June 23, 2011). 12. Friends of the Supergrid (FOSG). Position paper on the EC Communication for a European Infrastructure Package. December 2010. Available at: http://www. ebcd.org/website%2011/Grids/Supergrid%20Phase%20 1%20Final.pdf. (Accessed June 23, 2011). 13. Valov B, Strauß P, Degner T, et  al. “IWES‐Konzept 2010” zum Offshoreiibertragungssystem 2020. DEWI Magazin No. 37; 2010, 44–51. 14. DESERTEC Foundation. Red Paper: An overview of the  DESERTEC Concept. Available at: http://www. desertec.org/fileadmin/downloads/desertec‐foundation_ redpaper_3rd‐edition_english.pdf. (Accessed June 23, 2011). 15. DESERTEC Foundation (2009). WhiteBook: Clean Power from Deserts. The Desertec Concept for Energy, Water and Climate Security, 4e. Available at: http:// www.dun‐eumena.com/sites/default/files/files/doc/trec_ white_paper.pdf. 16. German Aerospace Center. Trans‐Mediterranean Interconnection for Concentrating Solar Power; 2006. Available at: http://www.dlr.de/tt/desktopdefault.aspx/ tabid‐2885/4422_read‐6588. (Accessed June 23,2011). 17. Press release at http://siemens.com: long line for electricity from hydropower. Reference Number: ­ PN  2010.01; 2010. Available at: http://www.siemens. com/press/en/presspicture/?press=/en/presspicture/ pictures‐photonews/2010/pn201001.php. (Accessed June 19, 2011). 18. Tschulik P. Energie Quo Vadis—Innovationen in der  Energietechnik; 2009. Available at: http://www. ingenieurbueros.at/media/Vpc_Basic_DownloadTag_ Component/49–1064–1066‐downloadTag/default/575d 441c2a12f22cf79569e177ecae9d/1304333150/vortrag_ di_dr_tschulik.pdf. (Accessed June 23, 2011). 19. De Decker J, Woyte A, Schödwell B, Vólker J, Srikan‐ dam C. Directory of offshore grid initiatives, studies & organisations. Tech. Rep. OffshoreGrid Project; 2009. Available at: www.offshoregrid.eu. (Accessed June 10, 2011). 20. Woyte A, De Decker J, Vu Van T. A North Sea electricity grid [Revolution—electricity output of interconnected offshore wind power: a vision of offshore wind power integration, Greenpeace–3E; 2008. Available at: http://www.greenpeace.org/raw/content/eu‐unit/

210  Advances in Energy Systems

press‐centre/reports/A‐North‐Sea‐electricity‐grid‐(r) evolution.pdf. (Accessed June 15, 2011). 21. Van Hertem D, Delimar M. Technical Limitations Towards a SuperGrid—A European Prospective. Bahrein: 2010 IEEE International Energy Conference; 2010. 22. Sood, V.K. (2004). HVDC and FACTS Controllers: Applications of Static Convertors in Power Systems. Kluwer Academic Publishers. 23. Asplund G, Eriksson K, Svensson K. DC Transmission based on Voltage Source Converters. CIGRE SC14 Colloquium in South Africa; 1997. 24. Van Hertem, D. and Ghandhari, M. (2010). Multi‐­ terminal VSC HVDC for the European Supergrid: Obstacles. Renewable Sustainable Energy Rev. 14: 3156–3163. 25. A European Supergrid. Memorandum submitted by E.ON UK (ESG 05). Session 2010–12 of the UK Parliament. March 2011. Available at: http://www.publications. parliament.uk/pa/cm201012/cmselect/cmenergy/ writev/1040/esg05.htm. (Accessed June 14, 2011). 26. Jeroense M. HVDC, the next generation of transmission highlights with focus on extruded cable systems. Electrical Insulating Materials; 2008. 27. Gustafsson, A., Jeroense, M., and Karlstrand, J. (2008). Light, safe and effective. ABB Rev. 2: 52–55. 28. Reddy, C.C. (2009). Theoretical maximum limits on power‐handling capacity of HVDC Cables. IEEE Trans. Power Delivery 24: 980–987. 29. Reddy, C.C. and Ramu, T.S. (2006). On the computation of electric stress and temperature distribution in HV DC

cable insulation. IEEE Trans. Dielectr. Electr. Insul. 13: 1236–1244. 30. Wang Q, Li X, Yin Y. Computer simulation and analysis of electric and temperature fields of HVDC cables. Proceedings of the 9th International Conference on Properties and Applications of Dielectric Materials. Harbin, China; 2009. 31. Maruyama, S., Ishii, N., Shimada, M. et  al. (2004). Development of a 500 kV DC XLPE cable system. Furukawa Rev. 25: 47–52. 32. Ooi, B.‐T. and Wang, X. (1991). Boost type PWM HVDC transmission system. IEEE Trans. Power Delivery 6: 1557–1563. 33. Bahrman, M.P., Johansson, J.G., and Nilsson, B.A. (2003). Voltage source converter transmission technologies – the right fit for the application. IEEE Power Eng. Soc. Gen. Meeting 3: 1840–1847. 34. Chen, H., Zhang, F., and Chang, Y. (2006). Improvement of power quality by VSC based multi‐terminal HVDC. IEEE Power Eng. Soc. Gen. Meeting. 35. Zadkhast, S., Fotuhi‐Firuzabad, M., Aminifar, F. et  al. (2010). Reliability Evaluation of an HVDC transmission system tapped by a VSC Station. IEEE Trans. Power Delivery 25: 19621970. 36. Linden K, Jacobson B, Bollen MHJ, et al. Reliability study methodology for HVDC grids. B4–108 CIGRE 2010. 37. Hirschhausen CV, Egerer J, Wand R. Engineering‐ Economic‐Institutional Analysis of Supergrids: General Issues and DESERTEC as a Case. EUI Academic Roundtable. Florence; 2010.

13

Wireless Power Transmission: Inductive Coupling, Radio Wave, and Resonance Coupling Naoki Shinohara Research Institute for Sustainable Humanosphere, Kyoto University, Kyoto, Japan

This chapter describes various wireless power transmission (WPT) technologies, including inductive coupling, radio waves, and resonance coupling. Theoretically, these three WPT technologies are similar, in that all of them depend on Maxwell’s equations. However, there are pros and cons for each application of the WPT. Therefore, a suitable WPT technology must be selected for each application. Herein, the theory, technologies, and applications of WPT are discussed. WPT is a useful and convenient technology that can be employed to charge the batteries in mobile phones, notebook PCs, electric vehicles, light‐emitting diodes, integrated circuits, and other equipment without the need for a wire connection. For systems that use very low power, the battery can be removed altogether, and the systems can be run on energy harvested from ambient radio frequency and microwave radiation. Therefore, the number of batteries can be reduced when wireless power is available from various locations, and because batteries can be charged wirelessly, concerns about the shortage of batteries can also be reduced. WPT via microwaves, for example, can be applied in the future to stable and CO2‐free space‐based solar power satellites. Overall, WPT

will support both future energy production and the environment.

­INTRODUCTION Approximately 100 years ago, Tesla[1, 2], a famous scientist, dreamt about a wireless technology that required no wire to feed electricity. He was theoretically right. However, he failed technically because he could not concentrate enough wireless power to satisfy the requirements of users, even 100 years ago[1, 2]. Nearly 100 years after Tesla’s proposition, it is now possible to use higher‐frequency radio waves, or microwaves, to focus electric power wirelessly at levels that satisfy current user needs. In case of a short distance, inductive or resonance coupling technologies are effective. Wireless power transmission (WPT) is, in fact, a valuable and convenient technology that can be used to charge the batteries in mobile phones, notebook PCs, electric vehicles (EVs), light‐emitting diodes (LEDs), integrated circuits (ICs), and other equipments. Systems that use very low power do not even require a battery, and can be run on energy harvested from ambient radio frequency and microwave radiation.

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

212  Advances in Energy Systems

Therefore, the number of batteries required can be reduced when wireless power is widespread because they can be charged wirelessly, and hence, the shortage of batteries will not be a concern. This scenario will come true soon. There are already various technologies and commercial applications of WPT. Some WPT commercial products are based on inductive or resonance coupling. These technologies are classified as “short‐range WPT.” Alternatively, WPT via radio waves is applied for long‐range WPT systems. In the near future, stable CO2free electric power from space will be possible with WPT technology. The concept of a space‐based solar power satellite (SPS) is supported by WPT technologies via microwaves. All WPT technologies are based on Maxwell’s equations. However, there are minor differences in their applications. In this chapter, the theory and application of WPT via inductive and resonance coupling are first described, followed by a discussion of the theory and applications of microwave‐based WPT technologies.

­HISTORY OF THE WPT Historically, WPT began with radio waves. Maxwell’s equations, formulated in 1862 and describing all phenomena of radio waves, are essentially the first theoretical basis for WPT. After Maxwell proposed his equations, Poynting described radio waves as an energy flow, which is a well‐known concept of the Poynting vector. After the discoveries of Maxwell and Poynting, Tesla[1, 2], still over 100 years ago, dreamt that all electricity would be provided wirelessly. He conducted the first WPT experiment at the end of the nineteenth century[1, 2]. Unfortunately, he failed because of the diffusion of the wireless power, which depends on the frequency of the operation and size of the transmitting antenna. He used an operating ­frequency of 150 kHz. After Tesla’s failure, the history of radio wave development focused on wireless communication and remote sensing, rather than WPT. However, the technologies for wireless communication and remote sensing helped lead to the development of new WPT techniques. To increase the information contained in radio waves, it is necessary to use higher frequency radio waves. A frequency higher than that used by Tesla, or microwaves, could be generated and used for wireless communication and radar systems only after World War II. Using a higher frequency with the same size antenna enables a greater concentration of radio waves than is possible with lower frequencies. Therefore, the transmitted radio wave power can be increased with microwaves as a higher frequency radio wave than that

used by Tesla. In the 1960s, Brown[3, 4] ­conducted new WPT experiments using high‐efficiency microwave technologies. He used 2.45 GHz microwave tubes (such as magnetrons and klystrons) and succeeded in achieving WPT to a helicopter with 2.45 GHz microwave in 1964 and to a free‐flying helicopter in 1968. He was also the first person to develop a rectifying antenna, which he called a rectenna, for receiving and rectifying microwaves. The efficiency of the first rectenna, developed in 1963, was 50% at an output power of 4 W DC and 40% at an output power of 7 W DC[5]. Finally, he achieved an estimated 90% efficiency rectenna at 2.45 GHz. In 1975, he achieved a total DC–DC efficiency of up to 54% at 495 W DC by using a magnetron in the Raytheon Laboratory. Simultaneously, Brown and Dickinson’s team succeeded with the largest microwave power transmission (MPT) demonstration in 1975 at the Venus Site of the Jet Propulsion Laboratory Goldstone Facility[6]. The distance between the parabolic transmitting antenna, which had a diameter of 26 m, and the rectenna array, which was 3.4 × 7.2 m2, was 1.6 km. The transmitted microwave power from the klystron source was 450 kW at a frequency of 2.388 GHz, and the achieved rectified DC power was 30 kW with 82.5% rectifying efficiency. This MPT experiment is the biggest ever carried out in the world. Although Brown succeeded with WPT field experiments, there was still a big mismatch between the demonstrated system size and cost and what would be practical. Therefore, unfortunately, commercial MPT systems did not become a part of our daily life. However, MPT technologies were developed for an SPS (Figure  13.1), which was proposed by Glaser[7] in 1968. The SPS overcame some of the drawbacks of MPT, such as the low overall system efficiency, which depends on the microwave/dc conversion, and the large size of the antennas. The SPS supplied approximately 10 times more electric power than solar cells on the ground because it was in geostationary orbit in space. There is no nighttime in geostationary orbit. In addition, unlike sunlight, microwaves are not absorbed by clouds and rain. Therefore, power generation is possible 24 hours per day within an SPS. MPT was required for the SPS, and consequently, MPT techniques were developed to meet the needs of the SPS[8]. After the 1980s, the center for MPT development moved to Japan. Prof. Hiroshi Matsumoto of Kyoto University and his group carried out MPT rocket experiments in 1983, which was called the Microwave Ionosphere Nonlinear Interaction experiment, for an SPS. They also carried out several MPT field experiments[8] on the basis of new

Wireless Power Transmission: Inductive Coupling, Radio Wave, and Resonance Coupling  213

Figure 13.1  SPS image.

microwave technologies for wireless communication and radar sensing. An MPT experiment with a phased array was carried out in 1992[9, 10], the first trial to apply the phased‐array technique for this purpose. Kyoto University, Kobe University, and their team flew a fuel‐free airplane powered only with 2.411 GHz of microwave energy using a phased array with 96 GaAs semiconductor amplifiers and 288 antennas in three subarrays. A Canadian group also flew a fuel‐free airplane in 1987 by adopting a parabolic antenna system[11]. The Institute of Space and Astronautical Science implemented a large SPS research project in the 1990s, called SPS2000[12]. Prof. Makoto Nagatomo, one of the leaders of the SPS research in Japan, was the project leader for the SPS2000 project, and believed that an experimental SPS should be launched soon. Therefore, the SPS2000 was designed as a concrete system in low‐ earth orbit. In the United States, some research groups continued MPT and SPS research through the 1980s. Then in 1995, NASA launched a project to take a fresh look at an SPS. The US research on MPT was adopted as a part of the project[13]. In the 2000s, the SPS research project in Japan is still ongoing[14, 15]. The Japanese SPS research project is based on the “Basic plan for space policy,” which was established by the Strategic Headquarters for Space Policy in June 2009 [16]. In this project, the focus is on the development of a high‐efficiency

and lightweight phased array of 5.8 GHz to control the microwave beam direction. Beam‐forming and target‐detecting algorithms and technologies are as important as the development of the high‐efficiency and lightweight phased array. There are various beam‐forming and target‐detecting techniques for the SPS, including retrodirective target detecting with a pilot signal, the rotating electromagnetic vector method[17], and the position and angle correction method[18]. The Japanese SPS project involves the verification of the various beam‐forming and target‐ detecting techniques. Simultaneously, advanced microwave technologies pushed WPT back into consideration for commercial applications. The power requirements for mobile communication technologies are fairly minimum. As a result, enough power can be received via microwaves, just as with wireless communication. There are many users of mobile phones, mobile PCs, sensor networks, LEDs, ICs, and so forth, who can benefit. In addition, significant microwave power can be transmitted with small antennas over a short distance to EVs. Researchers at Massachusetts Institute of Technology (MIT) made a critical discovery when they proposed and demonstrated resonance coupling WPT[19], which is based on band‐pass filter (BPF) and inductive coupling technologies. With the breakthrough at MIT, MPT can now, in the twenty‐first century, be applied not only to an SPS but also to various commercial mobile systems. The development of these other advanced microwave technologies, especially BPF  –  which enable resonance coupling WPT – has resulted in a selection of various WPT systems.

­INDUCTIVE COUPLING AND RESONANCE COUPLING Inductive Coupling WPT Inductive coupling WPT is based on Ampère’s circuital law and Faraday’s law of induction. Ampère’s circuital law describes the relationship between the integrated magnetic field around a closed loop (coil) and the electric current passing through the loop. Faraday’s law of induction describes the relationship between a time‐varying magnetic field and an induced electric field. The electric power is carried through the magnetic field between two coils (Figure 13.2). Ampère’s circuital law and Faraday’s law of induction are both examples of Maxwell’s equations. The efficiency of WPT depends on the coupling coefficient, which, in turn, depends on the distance between the two coils. Therefore, the wireless energy

214  Advances in Energy Systems

H User

products. In 2011, Nissan motor company released an inductive coupling wireless charging system for the charging of its electric cars. Magnetic Resonance Coupling WPT

I

I Supply Figure 13.2  Inductive coupling.

cannot be carried over a distance longer than a few millimeters with high efficiency, and the frequency used in inductive coupling is below some dozen megahertz. Inductive coupling WPT is the oldest type of WPT used in real applications. Some battery chargers adopted inductive coupling as an independent technology. In 1995, for example, the Japanese consumer electronics company Sony proposed and sold “Felica” IC cards with inductive coupling WPT. In Japan, various types of IC cards as well as transportation and electronic money have adopted the Felica system. The frequency of the Felica system is 13.56 MHz. A wireless charging pad for mobile phones based on the Qi standard defined by the Wireless Power Consortium (WPC) was released in 2011 in Japan[20]. The Qi standard is based on an inductive coupling WPT technology. The WPC has been active in popularizing the Qi standard and multipurpose inductive coupling WPT. Inductive coupling has also been applied to the wireless charging of EVs since the 1980s in the United States, when it was introduced as a part of the Partners for Advanced Transit and Highways project. In 2009 in Japan, Hino Motors Ltd.[21] and Showa Aircraft Industry[22] carried out wireless charging field experiments with an electric bus using the inductive power transfer technology developed by Wampfler Co., Germany. They measured the CO2 emissions in this experiment as compared with the wired charging, and reported that the wirelessly charged bus could reduce the amount of CO2 emissions because of high‐ frequency charging versus normal‐frequency wired charging[23]. Some companies – for example, HaloIPT Co., Evatran Co., and UniServices Co. – offer inductive coupling WPT systems for EVs as commercial

A resonator is formed by adding capacitance (C) on an induction (L) coil (Figure 13.3). Two resonators are coupled electromagnetically, and the energy in one resonator is transmitted to the other through an evanescent mode wave. This phenomenon is well known as a coupling theory applied to microwave BPF. However, it was not until 2006 that MIT researchers demonstrated a WPT experiment using resonance coupling[19]. Resonance coupling with coils is called magnetic resonance coupling. The transmitted power is mainly a magnetic field supported by the coil. Resonance coupling is realized with two plane‐shaped conductors passing through the electric field, and is called electric resonance coupling. After the publication of the MIT research, magnetic resonance coupling WPT systems were applied to mobile phones and other mobile devices, TVs, EVs, and so forth. In Korea, a resonant coupling technique for wireless power supply has been used for an online EV[24]. Power from the 60 Hz supply is converted to a frequency of 20 kHz by an inverter stage. Sixty kilowatts of power may be transferred wirelessly from the power lines with 80% efficiency. In 2011, IFEV2011, a topical international forum on EVs with wireless power charging was held in K ­ orea. In Japan, a research group at the University of Tokyo carried out a wireless charging experiment for an EV with resonance coupling. As a first experiment, they used an approximately 10 MHz band for WPT.

H User L C

I

L

C

I Supply Resonance of L and C

Figure 13.3  Magnetic resonance coupling.

Wireless Power Transmission: Inductive Coupling, Radio Wave, and Resonance Coupling  215

­Currently, they use approximately 120 kHz to be able to use the power supply components. Toyota Central R&C Lab. Inc. and Toyohashi University of Technology in Japan proposed a new concept for power transfer through a capacitor composed of a steel belt in a tire and a metal plate attached to the road[25]. Separately, in 2011, Toyota Motor Corporation invested in WiTricity Corporation, the first inventor of resonant coupling for WPT. IHI Corporation was also given a license by WiTricity Corporation in 2011. Furthermore, the Qualcomm Company in the United States began wireless charging experiments on EVs with resonance coupling in London in 2011. Previously, Qualcomm Company had demonstrated a wireless charging technology (the “eZone”) for mobile phones using a resonant coupling WPT technology at the Mobile World Congress in 2009[26], which had a coupling frequency of 13.56 MHz. Sony also released a resonance coupling WPT system for TVs in 2009. As the next step, they proposed a new WPT with a simple receiver coil[27].

Theory of Coupling WPT Recently, for magnetic resonant coupling, a theoretical investigation based mainly on inductive coupling has been carried out[28, 29]. Given the theory of inductive coupling, kQ, where k is the coupling coefficient

and Q is the quality factor of the coil resonator, is established as a critical factor. The maximum coupling transmission efficiency η is calculated using kQ as shown in the following equations: form 2 1 form 2

1 form 2



k 2Q1Q2

On condition that Z0 = 1 + k2Q1Q2 R2

Transmission efficiency

0.6

0.4 form2

ηmax =

0.2

1+

0 0.1

.

R1R2 . 2 M2

where ω is frequency, R is resistance, and M is mutual inductance. The efficiency curve is shown in Figure  13.4. The coupling transmission efficiency is determined by kQ. In inductive coupling, a high Q factor cannot be used; therefore, k should increase as a function of the distance between the two coils. However, in resonance coupling, it is easy to increase kQ with a high Q factor even if the distance between the two coils is large and the k factor is small (note that k contains a wavelength parameter). If a lower frequency is selected, then k increases if the distance and Q factor are the same. As a result, the WPT distance with a high efficiency is expanded using a lower frequency in an inductive or resonance coupling WPT system. A lower frequency indicates that higher efficiency can be achieved using lower cost components for the system.

1

0.8

2

1

10

1+ form2

2

100

Figure-of-merit (kQ) R1R2 Where = k2Q1Q2 = form2 ω2 M 2 K 2 = M2 /L1L2 L1,2 : Self-inductance of coil Q1,2 = ω0L1,2 /R1,2 M : Mutual inductance

Figure 13.4  Coupling transmission efficiency of resonant coupling.

216  Advances in Energy Systems

Transmitting system

Microwave power beam ( 2. This theoretical

Wireless Power Transmission: Inductive Coupling, Radio Wave, and Resonance Coupling  219

curve is similar to that shown in Figure 13.4 for resonance coupling WPT. After collecting the microwave power, it must be converted to DC on the rectenna. The rectenna is composed of an antenna that receives microwave power, a low‐pass filter that prevents higher harmonics from reaching the rectifier, and a rectifier with a diode(s) and an output filter to rectify the received microwave signal to DC. The ideal efficiency of the rectifying circuit is 100%. In the rectenna, a single‐shunt full‐ wave rectifying circuit with one diode, a λ/4 distributed line, and a capacitor is typically used to reduce the loss in the diode. The rectenna can theoretically rectify microwave signals with 100% efficiency with only one diode[39]. A theoretical approach with a real diode was described by Yoo and Chang[40] of Texas A&M University.

­CONCLUSIONS Over 100 years ago, Tesla dreamt that the need for wire to transmit electric power could be eliminated. In the twenty‐first century, wireless power can now be received to charge mobile phones and EVs. Theoretically, the technology in Tesla’s dream and that in the present are the same. The only difference is that other electric technologies, including batteries and semiconductors, remain a requirement. Our minds have been changed. At the end of the nineteenth century, no one believed in wireless power. Now, we do. There have been successful WPT experiments and commercial products. Today, I dream that, in the near future, there will be no wires and no batteries needed for electric power.

­REFERENCES 1. Tesla, N. (1904). The transmission of electrical energy without wires. In: The Thirteenth Anniversary Number of the Electrical World and Engineer. Breckenridge, CO: 21st Century Books. 2. Tesla, N. (1904). Experiments with Alternate Current of High Potential and High Frequency. New York, NY: McGraw Publishing Co. 3. Brown, W.C. (1984). The history of power transmission by radio waves. IEEE Trans. Microwave Theory Tech. 32: 1230–1242. 4. Brown WC. The history of the development of the rectenna. Proceedings of SPS Microwave Systems Workshop at JSC‐NASA; 1980, 271–280. 5. Brown WC. Adapting microwave techniques to help solve future energy problems. 1973 G‐MTT International Microwave Symposium Digest of Technical Papers 73.1; 1973, 189–191. 6. Dickinson RM. Performance of a high‐power, 2.388‐GHz receiving array in wireless power transmission over 1.54

km. 1976 MTT‐S International Microwave Symposium Digest; 1976, 139–141. 7. Glaser, P.E. (1968). Power from the sun: its future. Science 162: 857–886. 8. Matsumoto, H. (2002). Research on solar power station and microwave power transmission in Japan: review and perspectives. IEEE Microwave Mag. 3: 36–45. 9. Matsumoto H, Kaya N, Fujita M, et  al. MILAX airplane  experiment and model airplane [in Japanese]. Proceedings of the 11th ISAS Space Energy Symposium; 1993, 47–52. 10. Fujino Y, Itoh T, Fujita M, et al. A rectenna for MILAX. Proceedings of Wireless Power Transmission Conference ’93, 1993, 273–277. 11. Schlesak, J.J., Alden, A., and Ohno, T. (1988). A microwave powered high altitude platform. IEEE MTT‐S Int. Symp. Digest 283–286. 12. Nagatomo M, Itoh K. An evolutionary satellite power system for international demonstration in developing nations. Proceedings of SPS’91; 1991, 356–363. 13. McSpadden, J.O. and Mankins, J.C. (2002). Space solar power programs and microwave wireless power transmission technology. IEEE Microwave Mag. 3: 46–57. 14. Fuse Y, Saito T, Mihara S, et al. Microwave energy transmission program for SSPS. Proceedings of International Union of Radio Science (URSI) General Assembly 2011, Proceeding CD‐ROM CHGBDJK‐2.pdf; 2011. 15. Sasaki, S., Tanaka, K., Kawasaki, S. et  al. (2004). Conceptual study of SSPS demonstration experiment. Radio Sci. Bull. 310: 9–14. 16. Basic plan for space policy  –  wisdom of Japan moves space. Available at: http://www.kantei.go.jp/jp/singi/ utyuu/basic_plan.pdf. (Accessed July 20, 2012). 17. Homma Y, Sasaki T, Namura K, et al. New phased array and rectenna array systems for microwave power transmission research. Proceedings of 2011 IEEE MTT‐S International Microwave Workshop Series on Innovative Wireless Power Transmission: Technologies, Systems, and Applications (IMWS‐IWPT2011); 2011, 59–62. 18. Ishikawa T, Shinohara N. Study on optimization of microwave power beam of phased array antenna for SPS. Proceedings of 2011 IEEE MTT‐S International Microwave Workshop Series on Innovative Wireless Power Transmission: Technologies, Systems, and Applications (IMWS‐IWPT2011); 2011, 153–156. 19. Kurs, A., Karalis, A., Moffatt, R. et al. (2007). Wireless power transfer via strongly coupled magnetic resonances. Science 317: 83–86. 20. Wireless Power Consortium. Available at: http://www. wirelesspowerconsortium.com. (Accessed July 20, 2012). 21. Hino Global. Available at: http://hino.dga.jp/i‐viewer_ s/?p_no=7&m_p=20&p_id=1983&file_ n a m e = h t t p % 3 A % 2 F % 2 F w w w. h i n o ‐ g l o b a l . com%2Fpdf%2Fhinorep2007_e.pdf&t=HINO+Report &kw=IPT+hybrid. (Accessed July 20, 2012). 22. Showa Aircraft Industry Co., Ltd. Available at: http:// www.showa‐aircraft.co.jp/products/EV/kyuuden.html [in Japanese]. (Accessed July 20, 2012). 23. Takahashi, S. (2011). Wireless charging with inductive coupling for electric vehicle [in Japanese]. In: Wireless Power Transfer and Infrastructure Construction for

220  Advances in Energy Systems

Electric Vehicles (ed. Y. Hori and Y. Yokoi), supervisors, 47–55. Tokyo, Japan: CMC Publication. 24. Ahn S, Kim J. Magnetic field design for high efficient and low EMF wireless power transfer in on‐line ­electric  vehicle. Proceedings of EuCAP2011; 2011, 4148–4151. 25. Hanazawa M, Ohira T. Power transfer for a running automobile. Proceedings of 2011 IEEE MTT‐S International Microwave Workshop Series on Innovative Wireless Power Transmission: Technologies, Systems, and Applications (IMWS‐IWPT2011); 2011, 77–80. 26. Fast Company. Available at: http://www.fastcompany. com/blog/kit‐eaton/technomix/qualcomm‐teases‐vision‐ wireless‐charging‐future. (Accessed July 20, 2012). 27. Miyamoto T, Komiyama S, Mita H, et  al. Wireless power transfer system with a simple receiver coil. Proceedings of 2011 IEEE MTT‐S International Microwave Workshop Series on Innovative Wireless Power Transmission: Technologies, Systems, and Applications (IMWS‐IWPT2011); 2011, 131–134. 28. Inagaki N, Hori S. Classification and characterization of wireless power transfer systems of resonance method based on equivalent circuit derived from even and odd mode reactance functions. Proceedings of 2011 IEEE MTT‐S International Microwave Workshop Series on Innovative Wireless Power Transmission: Technologies, Systems, and Applications (IMWS‐IWPT2011); 2011, 115–118. 29. Hirayama, H., Ozawa, T., Hiraiwa, Y. et  al. (2009). A consideration of electro‐magnetic resonant coupling mode in wireless power transmission. IEICE Electron Express 6: 1421–1425. 30. Shinohara N, Mitani T, Matsumoto H. Study on ubiquitous power source with microwave power transmission. Proceedings of International Union of Radio Science

(URSI) General Assembly 2005, Proceeding CD‐ROM C07.5(01145).pdf; 2005. 31. Shinohara, N., Matsumoto, H., and Hashimoto, K. (2004). Phasecontrolled magnetron development for SPORTS: space power radio transmission system. Radio Sci. Bull. 310: 29–35. 32. Mitani T, Yamakawa H, Shinohara N, et  al. Demonstration experiment of microwave power and information transmission from an airship. Proceedings of 2nd International Symposium on Radio System and Space Plasma 2010; 2010, 157–160. 33. Sample AP, Smith JR. Experimental results with two wireless power transfer systems. Proceedings of RWS2009, MO2A‐5, 2009, 16–18. 34. Smith JR. Mapping the space of wirelessly powered systems. Proceedings of IMS 2010 Workshops, WFB‐3; 2010. 35. Powercast Corporation. Available at: http://www. powercastco.com. (Accessed July 20, 2012). 36. Shinohara, N. (2011). Beam efficiency of wireless power transmission via radio waves from short range to long range. J. Electromagn. Eng. Sci. 10: 224–230. 37. Shinohara N. Wireless charging system of electric vehicle with GaN Schottky diodes. IMS2011 Workshop WFA, CD‐ROM; 2011. 38. Shinohara N, Miyata Y, Mitani T, Niwa Net  al. New application of microwave power transmission for wireless power distribution system in buildings. Proceedings of Asia‐Pacific Microwave Conference 2008), CD‐ROM H2–08.pdf; 2008. 39. Gutmann, R.J. and Gworek, R.B. (1979). Yagi–Uda receiving elements in microwave power transmission system rectennas. J. Microwave Power 14: 313–320. 40. Yoo, T.W. and Chang, K. (1992). Theoretical and experimental development of 10 and 35 GHz rectennas. IEEE Trans. Microwave Theory Tech. 40: 1259–1266.

PART III

FLEXIBILITY MEASURES

14

The Role of Large‐Scale Energy Storage Under High Shares of Renewable Energy Shin‐ichi Inage Renewable Energy and Smart Grids, Hitachi Ltd., Tokyo, Japan

This chapter discusses how a high share of renewable energy (referred to as renewables) will influence the power quality of the grid. The mix of power generation varies from country to country. Each power generator has an important role in minimizing total operating costs and maintaining power quality. Conventionally, middle‐scale thermal power plants play a role in mitigating demand and supply variations. Under a high share of renewables, the supply adjustability of thermal power to mitigate the output variations due to renewables will run short. Therefore, energy storage systems will be required as a countermeasure. The energy storage capacity required will depend on the net variation due to a smoothing effect. The magnitude of this net variation is the key for estimating the correct storage capacities. Under net variations of 15% and 30%, the respective global energy storage capacities needed have been predicted to be 189 and 305 GW in 2050 when aiming at 50% carbon dioxide emission reductions globally. In this chapter, several energy storage systems are reviewed. Specifications for each energy storage system will be identified. It is concluded that an optimum mix of different energy storage systems will be essential to realize decarbonized power grids.

­INTRODUCTION According to the Assessment Report on Climate Change made in 2007 by the Intergovernmental Panel on Climate Change (IPCC), it appears more likely than previously thought that carbon emissions and other greenhouse gases are causing the global temperature rise. On the other hand, the macroeconomic costs to maintain the global gaseous emissions at 445–710 ppm heavily depend on the goal of the global gross domestic product (GDP), which ranges from +1.0 to −5.5%. These costs are predicted to slow down global GDP growth by 0.12%. At COP3 held in Kyoto in 1997, there was an agreement to reduce emissions during the five years from 2008 to 2012. G8 leaders reached a consensus of a 50% reduction in global CO2 emissions by 2050 at the Group of Eight Summit held in Heiligendamm, Germany in 2007. The International Energy Agency (IEA) formulated a future energy policy scenario in response to the Group of Eight Summit and has released many reports[1]. Especially, its Energy Technology Perspective 2008 proposed a blueprint to achieve 50% CO2 emission reductions by 2050 with breakthrough technologies to mitigate CO2 emissions such as

Advances in Energy Systems: The Large-scale Renewable Energy Integration Challenge, First Edition. Edited by Peter D. Lund, John A. Byrne, Reinhard Haas and Damian Flynn. © 2019 John Wiley & Sons Ltd. Published 2019 by John Wiley & Sons Ltd.

224  Advances in Energy Systems

60 Baseline emissions 57 Gt

55

End-use fuel and electricity efficiency 38%

50

Gt CO2

45 40

End-use fuel switching 15%

35

Power generation efficiency and fuel switching 5% Nuclear 6%

30 25

Renewables 17%

20

CCS 19%

15

BLUE map emissions 14 Gt

10 5 0 2010

2015

2020

2025

2030

2035

2040

2045

2050

Figure 14.1 CO2 reduction during 2005–2050 based on the BLUE Map scenario[1]. Source: Reproduced with permission from Ref[2] © OECD/IEA, 2009, Figure 1, p. 7.

renewable energy sources (referred to as renewables), saving energy, and carbon capture and storage (CCS). This blueprint is called the BLUE Map scenario and, as defined by IEA, aims to cut energy‐related CO2 emissions in half between 2005 and 2050. Figure 14.1 shows CO2 reduction based on the BLUE Map scenario. Renewables account for 17% of the total global CO2 emission mitigation target in 2050. This contribution comes on top of significant renewable growth in the baseline scenario. For example, the share of renewables in power generation is expected to rise to 46% in 2050 compared with around 19% today. The bulk of the growth of renewables is based on variable renewable supply options: wind, solar, and hydropower, which will each grow to around 5000 TWh. Given the high share of variable renewables in the total global power supply, power system planning faces an emerging challenge that will require engineering solutions to maintain electricity quality because a power supply based on variable renewables will always be subject to weather conditions. To maintain electricity quality, balancing the demand and supply is essential. Middle load, usually supplied by natural gas combined‐cycle (NGCC) plants, can play an important role in balancing the demand and supply, and can also serve as a backup capacity in the case of a renewable power supply shortfall. Figure  14.2 shows an example of output

­variations of wind speed. The difference is remarkable between the long‐term output, on the scale of days or months, and the short‐term output, on the scale of seconds or minutes[2]. A much higher share of renewable power with variable generation will raise a number of engineering issues in the future: 1. Short‐term variation. Variability on the scale of seconds or minutes will cause larger deviations of power system frequency. 2. Hourly variation. Variability on the scale of hours will increase the difficulty of hourly generation dispatch and unit commitments, and will influence the electricity trade between power systems. 3. Long‐term variation. Variability on the scale of days or months influences the stable supply of wind power. Hourly and long‐term variations are relatively forecastable because they depend on overall weather conditions. Therefore, they might be mitigated by flexible generation capacity, transmission interconnection, and load leveling. On the other hand, short‐term ­variations are quite random and difficult to forecast. Under a high share of renewables, both variations should be discussed, but this chapter focuses on short‐ term variations of power system frequency. Several

Intensity of variation of wind speed

The Role of Large‐Scale Energy Storage Under High Shares of Renewable Energy  225

Long term 8 days

10–2

Short term 1h

10–1

1

0.1 h

10 Cycles/h

102

30 seconds

103

Figure 14.2  Example of long‐ and short‐term variations of wind power[3]. Source: Created using data from Ref[3].

100

North Europe

Demand (%)

80

France

PJM

Germany

60 Italia

40

Japan

20 0

1

5

9

13

17

21

1

Hours Figure 14.3  Comparison of daily load curves[5]. Source: Reproduced with permission from Ref[5]. Copyright 2005, IEEJ.

possibilities are explored to maintain electricity quality, particularly energy storage technologies. Energy storage is not the only option able to mitigate output variations due to renewables; the prospects for its growth depend on how it compares to other methods[3]. In this paper, considerations were made from the viewpoint of mitigation of output variations of renewables. First, general characteristics of the power grid and renewables were described to well understand the issues of high share of renewables. Second, the required energy storage capacity was predicted based on the BLUE Map scenario of IEA. Further, several energy storage systems were reviewed. ­GENERAL CHARACTERISTICS OF THE POWER GRID Load duration curves can be split into curves with base and peak loads. Base loads are generated by plants for which output is difficult to change; they therefore operate most of the time at full capacity. Base loads

are generally served by either high‐efficiency fossil‐ fired or nuclear reactor power plants with low production cost. Peak loads are usually served by NGCC plants, gas turbine (GT) generation, or hydropower plants that can change their output in a short time, although with high production cost. Daily load curves for six electricity systems are shown in Figure 14.3[4]. Each load curve represents the day of the year with the maximum load variation and is normalized by the maximum load, which is equal to 100%. In Japan, the peak demand ranges from 50% to 100% of maximum. However, in Northern Europe, the Mid‐Central Atlantic region of the USA (PJM), Italy, France, and Germany, the difference is much smaller. The different load curve shapes depend on climate conditions (e.g. demand for air conditioning), lifestyle, and demand structure (e.g. industry demand is often flat when plants operate full time). In Japan, the large load variations are met by a number of peak load stations such as pumped hydropower and thermal power stations fueled by natural gas or oil. In particular, NGCC and pumped hydropower plants are suitable for following

226  Advances in Energy Systems

Italy UK Coal

France

Oil

Canada

LNG

Germany

Nuclear

Japan

Hydro

Russia

Others

China USA 0%

20%

40%

60%

80%

100%

Figure 14.4  Comparison of power generation mix[6]. Source: Reproduced with permission from “Prospects for Large Scale Energy Projects in Decarbonised Power Grids” © OECD/IEA 2009, Figure 6, p. 14.

the short cycle of demand variation. Variable renewable power generation options may or may not be suited to match fluctuating demand. For example, solar photovoltaic (PV) supplies broadly match air conditioning demand, because air conditioning is needed on hot days when the sun shines. Wind power may match such demand less favorably, because it depends only on wind speed that may be independent of air temperature. For countries with comparatively small demand variations, a high share of coal‐ or nuclear‐based load stations may be acceptable. The mixes of power generation resources found in different countries are compared in Figure 14.4[5]. These shares are a result of demand structure, energy resources, technology availability, and energy policy strategy. It should be noted that in France, Germany, Japan, China, and the United States, the total share of base load contributed by coal and nuclear power ranges from 50% to 90%, which will limit their potential to increase the share of variable renewables. ­POWER QUALITY While related issues include voltage and frequency stability, among others, this chapter limits itself to the issue of frequency stability in systems, with increasing shares of variable renewable generation assets. As shown in Figure 14.5 (left), daily demand constantly changes. The actual demand variations consist of the superposition of short‐term and long‐term variations, which depend on each user’s situation. Without a balance of demand and supply, frequency is not stable. Frequency falls when demand exceeds

supply and rises when supply exceeds demand. Therefore, the power supply must be completely balanced to produce stable power system frequency. Middle‐load power plants have a control system to minimize the frequency change. To balance the load between supply and demand, hydropower and thermal power plants modify their output through governor‐ free (GF) controls for short‐cycle demand variations. For an intermediate cycle, a load frequency control (LFC) system is used. For a long cycle, an economic load dispatching control (EDC) system controls the balance. Figure 14.5 (right) compares the time scales and functions of the three different frequency control techniques. Electric frequency is controlled within a small deviation: for example, in Japan the standard is 0.2– 0.3 Hz; in the United States it is 0.018–0.0228 Hz; and in the European UCTE it is 0.04–0.06 Hz. As renewables increase, the potential for fatal frequency changes grows, because renewable generators rarely have frequency control systems and can produce large variations in output as weather conditions change. Figure  14.6 compares the adjustable load rates of several types of power plants. In general, hydro plants have the fastest response times and are able to change from full power to zero and vice versa within one minute. On the other hand, coal thermal power plants respond comparatively slowly. Using a suitable combination of these power plants yields the optimal frequency. As mentioned in General Characteristics of the Power Grid, hourly and long‐term variations are dealt with by EDC, while short‐term variation that is quite random should be mitigated by LFC or GF. In existing power systems with a low share of renewables, the total capacity of resources needed to

Magnitude of demand fluctuation

The Role of Large‐Scale Energy Storage Under High Shares of Renewable Energy  227

Short-term fluctuation

Demand

Long-term fluctuation

Daily load curve

Daily load curve

Short-term fluctuation

Long-term fluctuation EDC LFC

GF

Time