Transcript of 1482206110 grids
- 1. S M A R T G R I D S Clouds, Communications, Open Source, and
Automation
- 2. Devices, Circuits, and Systems Series Editor Krzysztof
Iniewski CMOS Emerging Technologies Research Inc., Vancouver,
British Columbia, Canada PUBLISHED TITLES: Atomic Nanoscale
Technology in the Nuclear Industry Taeho Woo Biological and Medical
Sensor Technologies Krzysztof Iniewski Building Sensor Networks:
From Design to Applications Ioanis Nikolaidis and Krzysztof
Iniewski Circuits at the Nanoscale: Communications, Imaging, and
Sensing Krzysztof Iniewski Electrical Solitons: Theory, Design, and
Applications David Ricketts and Donhee Ham Electronics for
Radiation Detection Krzysztof Iniewski Embedded and Networking
Systems: Design, Software, and Implementation Gul N. Khan and
Krzysztof Iniewski Energy Harvesting with Functional Materials and
Microsystems Madhu Bhaskaran, Sharath Sriram, and Krzysztof
Iniewski Graphene, Carbon Nanotubes, and Nanostuctures: Techniques
and Applications James E. Morris and Krzysztof Iniewski High-Speed
Photonics Interconnects Lukas Chrostowski and Krzysztof Iniewski
Integrated Microsystems: Electronics, Photonics, and Biotechnology
Krzysztof Iniewski Integrated Power Devices and TCAD Simulation Yue
Fu, Zhanming Li, Wai Tung Ng, and Johnny K.O. Sin Internet
Networks: Wired, Wireless, and Optical Technologies Krzysztof
Iniewski Low Power Emerging Wireless Technologies Reza Mahmoudi and
Krzysztof Iniewski Medical Imaging: Technology and Applications
Troy Farncombe and Krzysztof Iniewski
- 3. MEMS: Fundamental Technology and Applications Vikas
Choudhary and Krzysztof Iniewski Microfluidics and Nanotechnology:
Biosensing to the Single Molecule Limit Eric Lagally and Krzysztof
Iniewski MIMO Power Line Communications: Narrow and Broadband
Standards, EMC, and Advanced Processing Lars Torsten Berger,
Andreas Schwager, Pascal Pagani, and Daniel Schneider
Nano-Semiconductors: Devices and Technology Krzysztof Iniewski
Nanoelectronic Device Applications Handbook James E. Morris and
Krzysztof Iniewski Nanoplasmonics: Advanced Device Applications
James W. M. Chon and Krzysztof Iniewski Nanoscale Semiconductor
Memories: Technology and Applications Santosh K. Kurinec and
Krzysztof Iniewski Novel Advances in Microsystems Technologies and
Their Applications Laurent A. Francis and Krzysztof Iniewski
Optical, Acoustic, Magnetic, and Mechanical Sensor Technologies
Krzysztof Iniewski Radiation Effects in Semiconductors Krzysztof
Iniewski Semiconductor Radiation Detection Systems Krzysztof
Iniewski Smart Grids: Clouds, Communications, Open Source, and
Automation David Bakken and Krzysztof Iniewski Smart Sensors for
Industrial Applications Krzysztof Iniewski Technologies for Smart
Sensors and Sensor Fusion Kevin Yallup and Krzysztof Iniewski
Telecommunication Networks Eugenio Iannone Testing for Small-Delay
Defects in Nanoscale CMOS Integrated Circuits Sandeep K. Goel and
Krishnendu Chakrabarty Wireless Technologies: Circuits, Systems,
and Devices Krzysztof Iniewski PUBLISHED TITLES:
- 4. FORTHCOMING TITLES: 3D Circuit and System Design: Multicore
Architecture, Thermal Management, and Reliability Rohit Sharma and
Krzysztof Iniewski Circuits and Systems for Security and Privacy
Farhana Sheikh and Leonel Sousa CMOS: Front-End Electronics for
Radiation Sensors Angelo Rivetti Gallium Nitride (GaN): Physics,
Devices, and Technology Farid Medjdoub and Krzysztof Iniewski High
Frequency Communication and Sensing: Traveling-Wave Techniques
Ahmet Tekin and Ahmed Emira High-Speed Devices and Circuits with
THz Applications Jung Han Choi and Krzysztof Iniewski Labs-on-Chip:
Physics, Design and Technology Eugenio Iannone Laser-Based Optical
Detection of Explosives Paul M. Pellegrino, Ellen L. Holthoff, and
Mikella E. Farrell Metallic Spintronic Devices Xiaobin Wang Mobile
Point-of-Care Monitors and Diagnostic Device Design Walter Karlen
and Krzysztof Iniewski Nanoelectronics: Devices, Circuits, and
Systems Nikos Konofaos Nanomaterials: A Guide to Fabrication and
Applications Gordon Harling and Krzysztof Iniewski Nanopatterning
and Nanoscale Devices for Biological Applications Krzysztof
Iniewski and Seila Selimovic Optical Fiber Sensors and Applications
Ginu Rajan and Krzysztof Iniewski Organic Solar Cells: Materials,
Devices, Interfaces, and Modeling Qiquan Qiao and Krzysztof
Iniewski Power Management Integrated Circuits and Technologies Mona
M. Hella and Patrick Mercier Radio Frequency Integrated Circuit
Design Sebastian Magierowski Semiconductor Device Technology:
Silicon and Materials Tomasz Brozek and Krzysztof Iniewski
- 5. FORTHCOMING TITLES: Soft Errors: From Particles to Circuits
Jean-Luc Autran and Daniela Munteanu VLSI: Circuits for Emerging
Applications Tomasz Wojcicki and Krzysztof Iniewski Wireless
Transceiver Circuits: System Perspectives and Design Aspects
Woogeun Rhee and Krzysztof Iniewski
- 6. CRC Press is an imprint of the Taylor & Francis Group,
an informa business Boca Raton London NewYork E D I T E D B Y David
Bakken Washington State University School of Electrical Engineering
and Computer Science M A N A G I N G E D I T O R Krzysztof Iniewski
CMOS Emerging Technologies Research Inc. Vancouver, British
Columbia, Canada S M A R T G R I D S Clouds, Communications, Open
Source, and Automation
- 7. MATLAB is a trademark of The MathWorks, Inc. and is used
with permission. The MathWorks does not warrant the accuracy of the
text or exercises in this book. This books use or discussion of
MATLAB soft- ware or related products does not constitute
endorsement or sponsorship by The MathWorks of a particular
pedagogical approach or particular use of the MATLAB software. CRC
Press Taylor & Francis Group 6000 Broken Sound Parkway NW,
Suite 300 Boca Raton, FL 33487-2742 2014 by Taylor & Francis
Group, LLC CRC Press is an imprint of Taylor & Francis Group,
an Informa business No claim to original U.S. Government works
Version Date: 20140305 International Standard Book Number-13:
978-1-4822-0612-8 (eBook - PDF) This book contains information
obtained from authentic and highly regarded sources. Reasonable
efforts have been made to publish reliable data and information,
but the author and publisher cannot assume responsibility for the
validity of all materials or the consequences of their use. The
authors and publishers have attempted to trace the copyright
holders of all material reproduced in this publication and
apologize to copyright holders if permission to publish in this
form has not been obtained. If any copyright material has not been
acknowledged please write and let us know so we may rectify in any
future reprint. Except as permitted under U.S. Copyright Law, no
part of this book may be reprinted, reproduced, transmit- ted, or
utilized in any form by any electronic, mechanical, or other means,
now known or hereafter invented, including photocopying,
microfilming, and recording, or in any information storage or
retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from
this work, please access www.copyright. com
(http://www.copyright.com/) or contact the Copyright Clearance
Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923,
978-750-8400. CCC is a not-for-profit organization that provides
licenses and registration for a variety of users. For organizations
that have been granted a photocopy license by the CCC, a separate
system of payment has been arranged. Trademark Notice: Product or
corporate names may be trademarks or registered trademarks, and are
used only for identification and explanation without intent to
infringe. Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com and the CRC Press Web site at
http://www.crcpress.com
- 8. ix Contents
Preface....................................................................................................................
xiii
Editors.......................................................................................................................xv
Contributors............................................................................................................xvii
Chapter 1 Mission-Critical Cloud Computing for Critical
Infrastructures...........1 Thoshitha Gamage, David Anderson, David
Bakken, Kenneth Birman, Anjan Bose, Carl Hauser, Ketan Maheshwari,
and Robbert van Renesse Chapter 2 Power Application Possibilities
with Mission-Critical Cloud
Computing...........................................................................................
17 David Bakken, Pranavamoorthy Balasubramanian, Thoshitha Gamage,
Santiago Grijalva, Kory W. Hedman, Yilu Liu, Vaithianathan
Venkatasubramanian, and Hao Zho Chapter 3 Emerging Wide-Area Power
Applications with Mission-Critical Data Delivery
Requirements...............................................................33
Greg Zweigle Chapter 4 GridStat: High Availability, Low Latency,
and Adaptive Sensor Data Delivery for Smart Generation and
Transmission......................55 David E. Bakken, Harald
Gjermundrd, and Ioanna Dionysiou Chapter 5 A Distributed Framework
for Smart Grid Modeling, Monitoring, and
Control....................................................................
115 Alfredo Vaccaro and Eugenio Zimeo Chapter 6 Role of PLC
Technology in Smart Grid Communication
Networks........................................................................................
133 Angeliki M. Sarafi, Artemis C. Voulkidis, Spiros Livieratos,
and Panayotis G. Cottis Chapter 7 Power Grid Network Analysis for
Smart Grid Applications............. 151 Zhifang Wang, Anna
Scaglione, and Robert J. Thomas
- 9. x Contents Chapter 8 Open Source Software, an Enabling
Technology for Smart Grid
Evolution...................................................................................
179 Russell Robertson, Fred Elmendorf, and Shawn Williams Chapter 9
Contribution of Microgrids to the Development of the Smart
Grid...............................................................................
191 Tine L. Vandoorn and Lieven Vandevelde Chapter 10
Microgrids.........................................................................................
213 Mietek Glinkowski, Adam Guglielmo, Alexandre Oudalov, Gary
Rackliffe, Bill Rose, Ernst Scholtz, Lokesh Verma, and Fang Yang
Chapter 11 Integrating Consumer Advance Demand Data in Smart Grid
Energy Supply
Chain........................................................................
251 Tongdan Jin, Chongqing Kang, and Heping Chen Chapter 12
Photovoltaic Energy Generation and Control for an Autonomous Shunt
Active Power
Filter..................................................................
275 Ayman Blorfan, Damien Flieller, Patrice Wira, Guy Sturtzer, and
Jean Merckl Chapter 13 Self-Tuning and Self-Diagnosing
Simulation.................................... 311 Jin Ma Chapter
14 A Consensus-Based Fully Distributed Load Management Algorithm
for Smart
Grid.................................................................
333 Yinliang Xu, Wei Zhang, and Wenxin Liu Chapter 15 Expert
Systems Application for the Reconfiguration of Electric
Distribution
Systems.........................................................................
359 Horacio Tovar-Hernndez and Guillermo Gutierrez-Alcaraz Chapter
16 Load Data Cleansing and Bus Load Coincidence
Factors................ 375 Wenyuan Li, Ke Wang, and Wijarn Wangdee
Chapter 17 Smart Metering and
Infrastructure...................................................399
Wenpeng Luan and Wenyuan Li
- 10. xiContents Chapter 18 Vision of Future Control Centers in
Smart Grids............................. 421 Fangxing Li, Pei
Zhang, Sarina Adhikari, Yanli Wei, and Qinran Hu
Index.......................................................................................................................
435
- 11. xiii Preface While electric interconnections have had
different kinds and levels of intelligence in them for many
decades, in the last 6years the notion of the smart grid has come
seemingly out of nowhere to be on the minds of not just power
engineers but policy makers, regulators, rate commissions, and the
general public. Inherent in the notion of the smart grid is the
ability to communicate much more sensor data and have far more
computations at many more locations using these data. The purpose
of this book is to give power engineers, information technology
workers in the electric sector, and others a snapshot of the state
of the art and practice today as well as a peek into the future
regarding the smart grid. There is a special focus on new kinds of
communications and computations enabled or necessitated by the
smart grid. This book is divided into four parts. Part I deals with
cloud computing, whose use is being seriously considered for
planning and operational use in a number of utilities and
independent system operators/regional transmission organizations as
of March 2014. Cloud computing has the potential to deploy massive
amounts of computational resources to help grid operations,
especially under contingency situations. Chapter1 describes the
mission-critical features that cloud computing infrastructures must
sup- port in order to be appropriate for operational use in power
grids. It also describes the Advanced Research Projects
Agency-Energy GridCloud project to develop such technologies.
Chapter 2 describes a handful of killer apps for cloud computing in
power grid operations. It has been written by leading power
researchers. Part II deals with wide-area communications for power
grids. Chapter 3 describes a wide range of power application
programs that have extreme communications require- ments over wide
distances. Such applications are becoming more widely deployed as
grids come under more pressure with every passing year. Chapter 4
describes GridStat, a middleware communications framework designed
from the ground up to meet these challenging requirements. The
chapter includes a detailed analysis of how different technologies
used in todays grids such as multiprotocol label switching,
Internet pro- tocol multicast, IEC 61850, and others are inadequate
for the applications described in Chapter 3 and the requirements
derived from them in Chapter 4. Chapter 5 presents an advanced
framework based on the service-oriented architecture approach for
integrated modeling, monitoring, and control. Chapter 6 analyzes
the role of power line commu- nication, which is also called
broadband over power lines, in the smart grid. Power line
communication/broadband over power lines technologies can provide
additional redun- dant paths for data delivery in a grid, and ones
that have failure characteristics other than traditional network
communications infrastructures. Finally, Chapter 7 describes a
novel approach for estimating the statistical properties of power
grids. This is an impor- tant first step toward having more broadly
reusable power algorithms with greater con- fidence, as computer
scientists and mathematicians have done for centuries. Part III
deals with open source, something common in other industries that
is start- ing to draw great interest from utilities and has great
potential to help stimulate inno- vation in power grids (which
suffer from a far higher degree of vendor lock-in than
- 12. xiv Preface most other industries). Chapter 8 explains what
open source software is and its history. It then overviews a number
of freely available open source power application programs. Part IV
deals with the broad category of automation. Chapter 9 explains how
microgrids fit into the smart grid landscape and how they can
contribute to its operations. Chapter 10 describes in detail the
design and operation of microgrids. Chapter11 introduces a virtual
energy provisioning concept by which utilities can collect and
aggregate advanced demand information in order to better manage
smart grid supply chains. Chapter 12 describes a new technique for
better managing pho- tovoltaic energy while limiting harmonic
pollution. Chapter 13 provides an approach for two-way interactions
between simulations and an operational wide-area measure- ment
system that is both self-tuning and self-diagnosing. Chapter 14
describes an approach for load management in smart grids that is
stable, distributed and employs multiagent techniques. Chapter 15
details the use of an expert system application to enable electric
distribution systems to be reconfigured in new and advantageous
way. Chapter 16 describes an approach for both cleansing the load
curve data and calculat- ing bus load coincidence factors in order
to better exploit smart meter data. Chapter 17 overviews an
advanced metering infrastructure system and its components,
discusses its benefits, and summarizes a variety of applications by
which smart metering and infrastructure supports both planning and
operations. Finally, Chapter 18 offers a vision of how smart grid
control centers may look in the future. David E. Bakken Pullman,
Washington Krzysztof (Kris) Iniewski Vancouver, British Columbia
MATLAB is a registered trademark of The MathWorks, Inc. For product
informa- tion, please contact: The MathWorks, Inc. 3 Apple Hill
Drive Natick, MA 01760-2098 USA Tel: 508 647 7000 Fax: 508-647-7001
E-mail: info@mathworks.com Web: www.mathworks.com MATLAB and
Simulink are trademarks of the MathWorks, Inc. and are used with
permission. The MathWorks does not warrant the accuracy of the text
or exercises in this book. This books use or discussion of MATLAB
and Simulink software or related products does not constitute
endorsement or sponsorship by the MathWorks of a particu- lar
pedagogical approach or particular use of the MATLAB and Simulink
software.
- 13. xv Editors David Bakken is a professor of computer science
in the School of Electrical Engineering and Computer Science at
Washington State University and chief scien- tist at GridStat, Inc.
His research interests include wide-area distributed computing
systems, middleware implementation, and dependable computing. Since
1999, he has been working closely with researchers in his
departments very strong electric power group on helping rethink the
way data delivery is done in power grids over the wide area, and is
considered the worlds leading expert on this. His GridStat
data-delivery software has influenced the shape of the emerging
NASPInet. He is a frequent visitor and lecturer at utilities,
electrical engineering departments, and power meetings worldwide.
Prior to Washington State University, Dr. Bakken was a research
scientist at BBN (Cambridge, MA), which built the first Internet in
1969. There he was coinventor of the Quality Objects middleware
framework, in which the Defense Advanced Research Projects Agency
invested more than 50 BBN person-years, which was integrated with
approximately 10 other research projects in various demonstrations,
and which flew in Boeing experimental aircraft. Dr. Bakken has
worked for Boeing and consulted for Amazon.com, Harris Corp.,
Realtime Innovations, Intel, TriGeo Network Security, and others.
He holds a MS (1990) and a PhD (1994) in computer science from the
University of Arizona, and Bachelor of Science degrees in com-
puter science and mathematics from Washington State University
(1985). He is the author of over 100 publications and coinventor of
three patents. Krzysztof (Kris) Iniewski manages R&D at Redlen
Technologies Inc., a start- up company in Vancouver, Canada.
Redlens revolutionary production process for advanced semiconductor
materials enables a new generation of more accurate, all- digital,
radiation-based imaging solutions. Kris is also president of CMOS
Emerging Technologies Research Inc. (www.cmosetr.com), an
organization of high-tech events covering communications,
microsystems, optoelectronics, and sensors. In his career, Dr.
Iniewski has held numerous faculty and management positions at the
University of Toronto, the University of Alberta, Simon Fraser
University, and PMC-Sierra Inc. He has published over 100 research
papers in international journals and con- ferences. He holds 18
international patents granted in the United States, Canada, France,
Germany, and Japan. He is a frequent invited speaker and has
consulted for multiple organizations internationally. He has
written and edited several books for CRC Press, Cambridge
University Press, IEEE Press, Wiley, McGraw-Hill, Artech House, and
Springer. His personal goal is to contribute to healthy living and
sus- tainability through innovative engineering solutions. In his
leisure time, Kris can be found hiking, sailing, skiing, or biking
in beautiful British Columbia. He can be reached at
kris.iniewski@gmail.com.
- 14. xvii Sarina Adhikari Department of Electrical Engineering
and Computer Science University of Tennessee Knoxville, Tennessee
David Anderson School of Electrical Engineering and Computer
Science Washington State University Pullman, Washington David
Bakken School of Electrical Engineering and Computer Science
Washington State University Pullman, Washington Pranavamoorthy
Balasubramanian School of Electrical, Computer, and Energy
Engineering Arizona State University Tempe, Arizona Kenneth Birman
Department of Computer Science Cornell University Ithaca, New York
Ayman Blorfan Modelling, Intelligence, Process and Systems
Laboratory Universit de Haute Alsace Mulhouse, France and National
Institute of Applied Science Strasbourg, France Anjan Bose School
of Electrical Engineering and Computer Science Washington State
University Pullman, Washington Heping Chen Ingram School of
Engineering Texas State University San Marcos, Texas Panayotis G.
Cottis School of Electrical and Computer Engineering National
Technical University of Athens Athens, Greece Ioanna Dionysiou
Department of Computer Science University of Nicosia Nicosia,
Cyprus Fred Elmendorf Grid Protection Alliance Chattanooga,
Tennessee Damien Flieller National Institute of Applied Science
Research Group of Electrical and Electronics in Nancy Strasbourg,
France Thoshitha Gamage School of Electrical Engineering and
Computer Science Washington State University Pullman, Washington
Contributors
- 15. xviii Contributors Harald Gjermundrd Department of Computer
Science University of Nicosia Nicosia, Cyprus Mietek Glinkowski ABB
Inc. Raleigh, North Carolina Santiago Grijalva School of Electrical
and Computer Engineering Georgia Institute of Technology Atlanta,
Georgia Adam Guglielmo ABB Inc. Raleigh, North Carolina Guillermo
Gutierrez-Alcaraz Department of Electrical Engineering Instituto
Tecnolgico de Morelia Morelia, Mexico Carl Hauser School of
Electrical Engineering and Computer Science Washington State
University Pullman, Washington Kory W. Hedman School of Electrical,
Computer, and Energy Engineering Arizona State University Tempe,
Arizona Qinran Hu Department of Electrical Engineering and Computer
Science University of Tennessee Knoxville, Tennessee Tongdan Jin
Ingram School of Engineering Texas State University San Marcos,
Texas Chongqing Kang Department of Electrical Engineering Tsinghua
University Beijing, China Fangxing Li Department of Electrical
Engineering and Computer Science University of Tennessee Knoxville,
Tennessee Wenyuan Li School of Electrical Engineering Chongqing
University Chongqing, China and BC Hydro Vancouver, Canada Yilu Liu
Department of Electrical Engineering and Computer Science
University of Tennessee Knoxville, Tennessee Wenxin Liu Klipsch
School of Electrical and Computer Engineering New Mexico State
University Las Cruces, New Mexico Spiros Livieratos School of
Pedagogical and Technological Education Athens, Greece Wenpeng Luan
State Grid Smart Grid Research Institute Beijing, China
- 16. xixContributors Jin Ma School of Electrical and Information
Engineering The University of Sydney New South Wales, Australia
Ketan Maheshwari Argonne National Laboratory Lemont, Illinois Jean
Merckl Modelling, Intelligence, Process and Systems Laboratory
Universit de Haute Alsace Mulhouse, France Alexandre Oudalov ABB
Inc. Raleigh, North Carolina Gary Rackliffe ABB Inc. Raleigh, North
Carolina Robbert van Renesse Department of Computer Science Cornell
University Ithaca, New York Russell Robertson Grid Protection
Alliance Chattanooga, Tennessee Bill Rose ABB Inc. Raleigh, North
Carolina Angeliki M. Sarafi School of Electrical and Computer
Engineering National Technical University of Athens Athens, Greece
Anna Scaglione Department of Electrical and Computer Engineering
University of California Davis, California Ernst Scholtz ABB Inc.
Raleigh, North Carolina Guy Sturtzer National Institute of Applied
Science Research Group of Electrical and Electronics in Nancy
Strasbourg, France Robert J. Thomas Department of Electrical and
Computer Engineering Cornell University Ithaca, New York Horacio
Tovar-Hernndez Department of Electrical Engineering Instituto
Tecnolgico de Morelia Morelia, Mexico Alfredo Vaccaro Department of
Engineering University of Sannio Benevento, Italy Lieven Vandevelde
Department of Electrical Energy Ghent University Ghent, Belgium
Tine L. Vandoorn Department of Electrical Energy Ghent University
Ghent, Belgium
- 17. xx Contributors Vaithianathan Venkatasubramanian School of
Electrical Engineering and Computer Science Washington State
University Pullman, Washington Lokesh Verma ABB Inc. Raleigh, North
Carolina Artemis C. Voulkidis School of Electrical and Computer
Engineering National Technical University of Athens Athens, Greece
Ke Wang School of Computing Science Simon Fraser University
Vancouver, Canada Zhifang Wang Department of Electrical and
Computer Engineering Virginia Commonwealth University Richmond,
Virginia Wijarn Wangdee The Sirindhorn International ThaiGerman
Graduate School of Engineering King Mongkuts University of
Technology North Bangkok Bangkok, Thailand Yanli Wei Department of
Electrical Engineering and Computer Science University of Tennessee
Knoxville, Tennessee Shawn Williams Grid Protection Alliance
Chattanooga, Tennessee Patrice Wira Modelling, Intelligence,
Process and Systems Laboratory Universit de Haute Alsace Mulhouse,
France Yinliang Xu Klipsch School of Electrical and Computer
Engineering New Mexico State University Las Cruces, New Mexico Fang
Yang ABB Inc. Raleigh, North Carolina Pei Zhang Grid Operations and
Planning Electric Power Research Institute Palo Alto, California
Wei Zhang Klipsch School of Electrical and Computer Engineering New
Mexico State University Las Cruces, New Mexico Hao Zhu Department
of Electrical and Computer Engineering University of Illinois
Champaign, Illinois Eugenio Zimeo Department of Engineering
University of Sannio Benevento, Italy Greg Zweigle Schweitzer
Engineering Laboratories,Inc. Pullman, Washington
- 18. 1 1 Mission-Critical Cloud Computing for Critical
Infrastructures Thoshitha Gamage, David Anderson, David Bakken,
Kenneth Birman, Anjan Bose, Carl Hauser, Ketan Maheshwari, and
Robbert van Renesse 1.1 INTRODUCTION The term cloud is becoming
prevalent in nearly every facet of day-to-day life, bring- ing up
an imperative research question: how can the cloud improve future
critical infrastructures? Certainly, cloud computing has already
made a huge impact on the computing landscape and has permanently
incorporated itself into almost all sec- tors of industry. The
same, however, cannot be said of critical infrastructures. Most
notably, the power industry has been very cautious regarding
cloud-based computing capabilities. This is not a total surprise:
the power industry is notoriously conservative about changing its
operating model, and its rate commissions are generally focused on
short-term goals. With thousands of moving parts, owned and
operated by just as many stakeholders, even modest changes are
difficult. Furthermore, continuing to CONTENTS 1.1 Introduction
......................................................................................................1
1.1.1 Cloud
Computing..................................................................................2
1.1.2 Advanced Power
Grid...........................................................................4
1.2 Cloud Computings Role in the Advanced Power
Grid....................................5 1.2.1 Berkeley Grand
Challenges and the Power Grid..................................7
1.3 Model for Cloud-Based Power Grid
Applications............................................8 1.4
GridCloud: A Capability Demonstration Case Study
......................................9 1.4.1
GridStat.................................................................................................9
1.4.2
Isis2......................................................................................................10
1.4.3
TCP-R.................................................................................................
11 1.4.4 GridSim
..............................................................................................12
1.4.5 GridCloud
Architecture......................................................................12
1.5
Conclusions.....................................................................................................13
References................................................................................................................15
- 19. 2 Smart Grids operate while incorporating large paradigm
shifts is neither a straightforward nor a risk-free process. In
addition to industry conservatism, progress is slowed by the lack
of comprehensive cloud-based solutions meeting current and future
power grid application requirements. Nevertheless, there are
numerous opportunities on many frontsfrom bulk power generation,
through wide-area transmission, to residential distribution,
including at the microgrid levelwhere cloud technologies can
bolster power grid operations and improve the grids efficiency,
security, and reliability. The impact of cloud computing is best
exemplified by the recent boom in e-commerce and online shopping.
The cloud has empowered modern customers with outstanding
bargaining power in making their purchasing choices by provid- ing
up-to-date pricing information on products from a wide array of
sources whose computing infrastructure is cost-effective and
scalable on demand. For example, not long ago air travelers relied
on local travel agents to get the best prices on their reser-
vations. Cloud computing has revolutionized this market, allowing
vendors to easily provide customers with web-based reservation
services. In fact, a recent study shows that online travel
e-commerce skyrocketed from a mere $30 billion in 2002 to a
staggering $103 billion, breaking the $100 billion mark for the
first time in the United States in 2012 [1]. A similar phenomenon
applies to retail shopping. Nowadays, online retail shops offer a
variety of products, ranging from consumer electronics, clothing,
books, jewelry, and video games to event tick- ets, digital media,
and lots more at competitive prices. Mainstream online shops such
as Amazon, eBay, Etsy, and so on provide customers with an
unprecedented global marketplace to both buy and sell items. Almost
all major US retail giants, such as Walmart, Macys, BestBuy,
Target, and so on, have adopted a hybrid sales model, providing
online shops to complement the traditional in-store shopping
experience. A more recent trend is flash sale sites (Fab, Woot,
Deals2Buy, Totsy, MyHabit, etc.), which offer limited-time deals
and offers. All in all, retail e-commerce in the United States
increased by as much as 15% in 2012, totaling $289 billion. To put
this into perspective, the total was $72 billion 10 years earlier.
Such rapid growth relied heav- ily on cloud-based technology to
provide the massive computing resources behind online shopping.
1.1.1 CLOUD COMPUTING What truly characterizes cloud computing is
its business model. The cloud provides on-demand access to
virtually limitless hardware and software resources meeting the
users requirements. Furthermore, users only pay for resources they
use, based on the time of use and capacity. The National Institute
of Standards and Technology (NIST) defines five essential cloud
characteristics: on-demand self-service, broad network access,
resource pooling, rapid elasticity, and measured service [2]. The
computational model of the cloud features two key characteristics
abstraction and virtualization. The cloud provides its end users
with well-defined application programming interfaces (APIs) that
support requests to a wide range of hardware and software
resources. Cloud computing supports various configura- tions
(central processing unit [CPU], memory, platform, input/output
[I/O], network- ing, storage, servers) and capacities (scale) while
abstracting resource management
- 20. 3Mission-Critical Cloud Computing for Critical
Infrastructures (setup, startup, maintenance, etc.), underlying
infrastructure technology, physical space, and human labor
requirements. The end users see only APIs when they access services
on the cloud. For example, users of Dropbox, the popular
cloud-based online storage, only need to know that their stored
items are accessible through the API; they do not need any
knowledge of the underlying infrastructure supporting the service.
Furthermore, end users are relieved of owning large computing
resources that are often underused. Instead, resources are housed
in large data centers as a shared resource pool serving multiple
users, thus optimizing their use and amortiz- ing the cost of
maintenance. At the same time, end users are unaware of where their
resources physically reside, effectively virtualizing the computing
resources. Cloud computing provides three service models: software
as a service (SaaS), platform as a service (PaaS), and
infrastructure as a service (IaaS). Each of these service models
provides unique APIs. Services can be purchased separately, but are
typically purchased as a solution stack. The SaaS model offers
end-point business applications which are customizable and
configurable based on specific needs. One good example is the
Google Apps framework, which offers a large suite of end-user
applications (email, online storage, streaming channels, domain
names, messaging, web hosting, etc.) that individuals, businesses,
universities, and other organizations can purchase individually or
in combination. Software offered in this manner has a shorter
development life cycle, resulting in frequent updates and
up-to-date versions. The life-cycle maintenance is explicitly
handled by the service provider, who offers the software on a
pay-per-use basis. Since the software is hosted in the cloud, there
is no explicit installation or maintenance process for the end
users in their native environment. Some of the prominent SaaS
providers include Salesforce, Google, Microsoft, Intuit, Oracle,
and so on (Figure 1.1). The PaaS model offers a development
environment, middleware capabilities, and a deployment stack for
application developers to build tailor-made applications or host
prepurchased SaaS. Amazon Web Services (AWS), Google App Engine,
and Microsoft Azure are a few examples of PaaS. In contrast to
SaaS, PaaS does not abstract development life-cycle support, given
that most end users in this model are application developers.
Nevertheless, the abstraction aspect of cloud computing is Software
as a service (application) Platform as a service (operating system)
Infrastructure as a service (hardware) FIGURE 1.1 Cloud service
models as a stack.
- 21. 4 Smart Grids still present in PaaS, where developers rely
on underlying abstracted features such as infrastructure, operating
system, backup and version control features, development and
testing tools, runtime environment, workflow management, code
security, and collaborative facilities. The IaaS model offers the
fundamental hardware, networking, and storage capa- bilities needed
to host PaaS or custom user platforms. Services offered in IaaS
include hardware-level provisioning, public and private network
connectivity, (redundant) load balancing, replication, data center
space, and firewalls. IaaS relieves end users of operational and
capital expenses. While the other two models also provide these
features, here they are much more prominent, since IaaS is the
closest model to actual hardware. Moreover, since the actual
hardware is virtualized in climate-controlled data centers, IaaS
can shield end users from eventual hardware failures, greatly
increasing availability and eliminating repair and maintenance
costs. A popular IaaS provider, Amazon Elastic Compute Cloud (EC2),
offers 9 hardware instance fami- lies in 18 types [3]. Some of the
other IaaS providers include GoGrid, Elastic Hosts, AppNexus, and
Mosso [4]. 1.1.2 ADVANCED POWER GRID Online shopping is just one of
many instances where cloud computing is making its mark on society.
The power grid, in fact, is currently at an interesting cross-
roads in this technological space. One fundamental capability that
engineers are striving to improve is the grids situational
awarenessits real-time knowledge of grid statethrough highly
time-synchronized phasor measurement units (PMUs), accurate digital
fault recorders (DFRs), advanced metering infrastructure (AMI),
smart meters, and significantly better communication. The industry
is also facing a massive influx of ubiquitous household devices
that exchange information related to energy consumption. In light
of these new technologies, the traditional power grid is being
transformed into what is popularly known as the smart grid or the
advanced power grid. The evolution of the power grid brings its own
share of challenges. The newly introduced data have the potential
to dramatically increase accuracy, but only if pro- cessed quickly
and correctly. True situational awareness and real-time control
deci- sions go hand in hand. The feasibility of achieving these two
objectives, however, heavily depends on three key features: 1. The
ability to capture the power grid state accurately and
synchronously 2. The ability to deliver grid state data reliably
and in a timely manner over a (potentially) wide area 3. The
ability to rapidly process large quantities of state data and
redirect the resulting information to appropriate power
application(s), and, to a lesser extent, the ability to rapidly
acquire computing resources for on-demand data processing Emerging
power applications are the direct beneficiaries of rapid data cap-
ture, delivery, processing, and retrieval. One such example is the
transition from
- 22. 5Mission-Critical Cloud Computing for Critical
Infrastructures conventional state estimation to direct state
calculation. Beginning in the early 1960s, the power grid has been
employing supervisory control and data acquisi- tion (SCADA)
technology for many of its real-time requisites, such as balancing
load against supply, demand response, and contingency detection and
analysis. SCADA uses a slow, cyclic polling architecture in which
decisions are based on unsynchronized measurements that may be
several seconds old. Consequently, the estimated state lags the
actual state most. Thus, state estimation gives very lim- ited
insight and visibility into the grids actual operational status. In
contrast, tightly time-synchronized PMU data streams deliver data
under strict quality of service (QoS) guaranteeslow latency and
high availabilityallowing control centers to perform direct state
calculations and measurements. The capabilities that come with the
availability of status data make creating a real-time picture of
the grids opera- tional state much more realistic [5]. There are
also many myths surrounding the operations of a power grid in con-
junction with big data and its efficient use. The following is a
nonexhaustive list of some of these myths. 1. Timeliness: Real-time
data is a relative term. Often the application require- ments
dictate the timeliness needs. Modern software and hardware tech-
nologies provide many workarounds on the timeliness of data
availability on wide area networks with average bandwidths. One of
them is selective packet dropping. This technique guarantees a
minimum QoS while deliv- ering information to recipients in a
timely manner. Smart power grids will greatly benefit from these
techniques. 2. Security and Safety: Security and safety are
concerns often cited by deci- sion makers when considering new
technologies. While absolute security is impossible, most concerns
arising from data security issues have been technically addressed.
One large factor that affects security is human errors and
oversights. Often, insufficient emphasis is given to this side of
security. More and more emphasis is given to the communication
channels. Securing an already secure channel only results in
performance losses and overheads. 3. Cost: The cost of maintaining
information infrastructures has become a major portion of budgets
for large industries, and is a substantial challenge in running a
sustainable, data-centered architecture. Thanks to data centers and
cloud computing infrastructures, these challenges are being
success- fully addressed. Clouds facilitate outsourcing of
large-scale computational infrastructures while achieving provably
reliable QoS. 1.2 CLOUD COMPUTINGS ROLE IN THE ADVANCED POWER GRID
Cloud computing can play a vital role in improving the advanced
power grids situational awareness and the ability to derive better
control decisions. As men- tioned earlier, emerging power
applications will leverage large amounts of data in making control
decisions affecting the stability and reliability of the grid.
- 23. 6 Smart Grids Analyzingandprocessing such large amounts of
data require data parallelism and massive computational
capabilities well beyond general-purpose computing. Beyond data
analysis, the future grid can greatly benefit from much more
extensive simula- tion and analysis to remediate stressful
situations. These are spontaneous special purpose applications
(e.g., system integrity protection schemes [SIPS], also known as
remedial action schemes [RASs] or special protection schemes
[SPSs]) [6], each with different needsreal time, scaling, and
computationalthat are triggered by grid disturbances such as
frequency oscillations, voltage fluctuations, line overloads, and
blackouts. Moreover, the number of power grid applications and
their computational needs can only be expected to increase as the
grid evolves. Managing this variety of applications and needs
presents a challenge. Keeping these applications running idle on
dedicated hardware until the specific condition is triggered is
both inefficient and expensive. An elegant solution is presented
here which utilizes cloud computing and its rapid elasticity. Power
grid applications can utilize the cloud to rapidly deploy an
application-specific infrastructure using IaaS and PaaS to achieve
new levels of availability and scalability. Availability and
scalability are properties that are much harder to meet in a
piecemeal fashion, but are inherent features of the cloud and
easily adoptable. Cloud-based solutions also benefit entities at
different levels of the control hierarchy, giving them the ability
to perform an independent, replicated analysis on the same sensor
data. The ability to elastically manage resources in the presence
of a grid disturbance is extremely attractive in comparison with
in-house solutions, which could be overprovisioned or
underprovisioned at the time of need. Another area where cloud
computing performs well is in supporting the varying needs of the
growing ecosystem of power applications. Both PaaS and SaaS will be
useful for developing and supporting power applications. PaaS for
the power grid will need to encompass industry best practices,
engineer- ing standards, compliance requirements, and data privacy
and security requirements as properties of the platform itself. The
CAP theorem [7] argues that simultaneously achieving three key
propertiesconsistency, availability, and partition tolerance is
impossible in distributed systems. As a result, and especially
since their apps are not mission critical, present-day commercial
clouds often sacrifice consistency in favor of availability. Cloud
environments that are used for power applications must be able to
guarantee high-assurance properties, including consistency, fault
tolerance, and real-time responsiveness, in order to support the
anticipated needs of power applications. While PaaS enables power
researchers and developers to expand the power appli- cation
ecosystem, SaaS can abstract essential properties and requirements
to provide end-user application solutions. Grid incident-specific
applications can be offered as SaaS, readily deployable by power
utilities. The success of power grid SaaS depends heavily on the
completeness and the richness of power grid PaaS. The overarching
challenge lies in ensuring that power applications delivered across
SaaS/PaaS mod- els inherently carry the necessary high-assurance
properties. The subtle intricacies of high-assurance properties,
which are often outside the power engineering realm, will
necessitate a different approach to cloud computing as well as a
stronger mesh between power engineering and computer science.
- 24. 7Mission-Critical Cloud Computing for Critical
Infrastructures 1.2.1 BERKELEY GRAND CHALLENGES AND THE POWER GRID
The Berkeley view of the cloud [8,9] outlines 10 grand challenges
and opportunities for cloud computing. The following list reviews
some of these challenges and their implications for cloud-based
power grid applications: 1. Ensuring High Service Availability:
Consistency is arguably one of the most critical requirements for
cloud-based power applications [10], but availability is a close
second. Many of the early adopters of cloud technol- ogy support
availability as a liveness property, while smart-grid applica-
tions depend on availability as a safety property. Availability
also relates to the issue of whether cloud-based power grid
applications should follow stateful or stateless models. The
ability to design stateful applications often depends on the
availability of state information. Achieving high availability
requires avoiding single point of failure scenarios and potential
bottlenecks. The general consensus is that the cloud promotes and
provides high availability. However, using cloud ser- vices from a
single service provider allows a single point of failure [9].
Interoperability between different cloud vendors for the sake of
availability for power grid applications merely because of the many
proprietary and market advantages is a far-fetched ambition.
Perhaps one solution would be for the power grid community to
manage and operate its own cloud, either overlaying existing
commercial clouds or as a private infrastructure with built-in
replication at all levels. Such an initiative, however, would be
dictated by the many economic drivers. 2. Eliminating Data Transfer
Bottlenecks: Large amounts of high-frequency data must cross the
cloud boundary to reach power applications running within the
cloud. Application responsiveness is directly tied to the time
lines with which data reach their destination. The outermost layer
of the cloud can have a dual role as a sensor data aggregator for
sources outside the cloud and as a multiplexer toward the
applications within the cloud. Thus, a sufficiently large number of
replicated cloud end points for sensor data must be provided in
order to prevent a potential data transfer bottleneck. 3. Assuring
Data Confidentiality and Availability: For a community histori-
cally notorious for a conservative modus operandi, sharing sensor
data is a frightening proposition. Power grid entities operate
under industry regula- tions and standards that can prevent data
sharing in many circumstances. Additionally, companies are
reluctant to share market-sensitive data that could give away
economic advantage. Power application data travers- ing the cloud
must be secured, meeting all compliance requirements, so that they
cannot be used by unintended parties. Thus, the cloud will need to
provide adequate data sanitization and filtering capabilities to
protect applicationdata. 4. Performance Predictability under
Scaling: The enormous amount of data that some power applications
require, combined with the impacts of vir- tualized I/O channels
and interrupts, leads to unpredictable performance
- 25. 8 Smart Grids during elastic scaling. Different IaaS
vendors exhibit different I/O perfor- mance, resource acquisition,
and release time characteristics. The computa- tional performance
on current cloud computing infrastructures also shows signs of
strain under scaling [11]. High-end batch processing applications
will require improved resource sharing and scheduling capabilities
for vir- tual machines to ensure strict QoS demands. 1.3 MODEL FOR
CLOUD-BASED POWER GRID APPLICATIONS Many of the cloud adoption
challenges outlined in Section 1.2 are essentially about supporting
highly scalable, highly assured behaviors and stringent
communication guarantees. These are properties that are rarely
found in todays commercial cloud infrastructures, which are
optimized to support mobile applications and web-based e-commerce
applications. The notions of real-time responsiveness, guaranteed
con- sistency, data security, and fault tolerance are significantly
more forgiving in these applications than in infrastructure
control, supplying little incentive for current com- mercial clouds
to embrace the type of changes necessary to support critical infra-
structure systems. Figure 1.2 visually represents an abstract
architectural model for cloud-based power applications. The
architecture includes three basic components: 1. A real-time data
collection and transport infrastructure 2. A soft state, elastic
outer cloud tier that supports data collection, data pre-
processing, temporary archiving, data sanitization and filtering,
and multi- plexing to services residing in interior tiers 3.
Interior cloud tiers hosting services and applications, and
supporting data processing, analysis, batch processing, persistent
storage, and visualization functions The data collection and
transportation infrastructure sits between the physical sensors and
the outermost tier of the cloud, and is the communication backbone
of Data sanitization Soft archiving Visualization Hard archiving
Interior cloud tier (applications) Outermost cloud tier (data
collectors) Analysis Computation Data multiplexing Data aggregation
Batch processing Data ltering Data collection and
transportationSensors FIGURE 1.2 An abstract architectural model
for a cloud-based power grid application.
- 26. 9Mission-Critical Cloud Computing for Critical
Infrastructures the overall architecture. This component is
responsible for delivering data that are produced outside the cloud
to the first-tier cloud collectors with strong QoS guaran- tees
such as guaranteed delivery, ultrahigh availability, ultralow
latency, and guar- anteed latency. The soft state, outermost cloud
tier provides the interface to data flowing to the applications
hosted in the interior tiers. The primary objective of this tier is
to provide high availability, to exhibit rapid elasticity, and to
forward correct data to the appropriate applications. To aid in
this process, this tier will also host auxiliary applications that
provide data sanitization, filtering of bad data, buffering (or
soft achieving), data preprocessing, and forwarding capabilities.
Availability and fault tolerance are heightened by replicated
shardsnodes that collect data from a group of sensorsand by mapping
sensor data sources appropriately to the shards. The interior cloud
tiers host the actual applications that consume data from the
shards and perform analysis, computation, and batch processing
tasks. Additionally, the results of these deeper computations may
be delivered at high rates to visualiza- tion applications residing
inside and outside the cloud. 1.4 GRIDCLOUD: A CAPABILITY
DEMONSTRATION CASE STUDY An Advanced Research Projects
Agency-Energy (ARPA-E)-funded, high-profile research collaboration
between Cornell University and Washington State University is
spearheading efforts to develop, prototype, and demonstrate a
powerful and com- prehensive software platform realizing the cloud
computing needs of the future power grid. Appropriately named
GridCloud [12], this research project aims to bring together
best-of-breed, already existing high-assurance distributed system
technolo- gies as a basis to innovate new cloud architectural
models for the monitoring, man- agement, and control of power
systems. The technologies integrated in this effort include
GridStat [13,14], Isis2 [15,16], TCP-R [17], and GridSim [18]. A
brief descrip- tion of each of these technologies is presented
here. 1.4.1 GRIDSTAT GridStat implements a data delivery overlay
network framework designed from the bottom up to meet the
challenging requirements of the electric power grid (see Chapter
4). Power grids today are increasingly populated with high-rate,
time-synchro- nized sensors that include PMUs and DFRs, whose
functionalities are actually blurring. High-rate, time-synchronized
data are expected to form the basis of many monitor- ing and
control applications with a wide variety of delivery requirements
and config- urations across such dimensions as geographic scope,
latency, volume, and required availability[19]. These needs cannot
be met by Internet protocol (IP) multicast, which forces all
subscribers of a given sensor variable to get all updates at the
highest rate that any subscriber requires. They also cannot be met
by multiprotocol label switching (MPLS), which is not designed to
provide per-message guarantees (only overall statisti- cal
guarantees) and also only has three bits (eight categories) with
which to categorize the millions of different sensor flows that
will likely be deployed in 510 years. GridStat delivers rate-based
updates of sensor variables with a wide range of QoS+ guarantees
(latency, rate, availability) that include support for ultralow
latency
- 27. 10 Smart Grids and ultrahigh availability, which are
implemented by sending updates over redundant disjoint paths, each
of which meets the end-to-end latency requirements for the given
subscription. Additionally, GridStat enables different subscribers
to a given sensor variable to require different QoS+ guarantees,
which can greatly reduce bandwidth requirements and improve
scalability. GridStats data delivery plane is a flat graph of
forwarding engines (FEs), each of which stores the state for every
subscription whose updates it forwards. FEs for- ward sensor
updates on each outgoing link at the highest rate that any
downstream subscriber requires. They drop updates that are not
needed downstream, based on the expressed rate requirements of
subscribers. GridStats management plane is implemented as a
hierarchy of QoS brokers that can be mapped onto the natural
hierarchy of the power grid. Each node in the hierarchy is designed
to contain poli- cies for resource permissions, security
permissions, aggregation, and adaptations to anomalies. With these
policies, the management plane calculates the paths required for
the data delivery (with the given number of disjoint paths) and
updates the for- warding tables in the FEs. Applications interact
with GridStat using publisher and subscriber software libraries
through which the applications requirements for QoS are conveyed to
the management plane. GridStat incorporates mechanisms for secur-
ing communication between the management plane entities and those
of the data plane. Security mechanisms for end-to-end message
security between publishers and subscribers are modular and
configurable, allowing different data streams and appli- cations to
fulfill different security and real-time requirements [20].
GridStat in the power grid provides the opportunity to respond to
different power system operating conditions with different
communication configurations. GridStat provides a mecha- nism by
which communication patterns can be rapidly changed among multiple
pre- configured modes in response to anticipated power system
contingencies. 1.4.2 ISIS2 Isis2 is a high-assurance replication
and coordination technology that makes it easy to capture
information at one location and share it in a consistent,
fault-tolerant, secure manner with applications running at other
locationsperhaps great num- bers of them. This system revisits a
powerful and widely accepted technology for replicating objects or
computations, but with a new focus on running at cloud scale, where
the system might be deployed onto thousands of nodes and supporting
new styles of machine-learning algorithms. Isis2 enables massive
parallelism, strong con- sistency, and automated fault tolerance,
and requires little sophistication on the part of its users. With
Isis2, all computational nodes and applications sharing the same
data see it [the data?] evolve in the same manner and at nearly the
same time, with delays often measured in hundreds of microseconds.
The system also supports repli- cated computation and coordination:
with Isis2 one could marshal 10,000 machines to jointly perform a
computation, search a massive database, or simulate the conse-
quences of control actions, all in a manner that is fast, secure
against attack or intru- sion, and correct even if some crashes
occur. The form of assurance offered by Isis2 is associated with a
formal model that merges two important prior modelsvirtual
synchrony [21] and Paxos [22]. Isis2
- 28. 11Mission-Critical Cloud Computing for Critical
Infrastructures embeds these ideas into modern object-oriented
programming languages. Isis2 is used to create two new components
for GridCloud: a version of the technology spe- cialized for use in
wide-area power systems networks, and support for high-assurance
smart-grid applications that are hosted in cloud computing data
centers. The GridCloud researchers believe that Isis2 can be used
to support services that run on standard cloud infrastructures and
yet (unlike todays cloud solutions) are able to guarantee
continuous availability, automatically adapting under attack so
that intruders cannot disrupt the grid even if a few nodes are
compromised. They are also analyzing and demonstrating the best
options for building cloud services that respond to requests in a
time-critical manner. 1.4.3 TCP-R GridCloud will tie together a
very large number of components, including sensors, actuators,
forwarding elements and aggregators, cloud-based services, and so
on, using Internet standards. For best performance, it is important
that related components communicate using persistent, stateful
connections. Stateful connections reduce retransmissions and
wasteful connections, and provide better flow control. The standard
for stateful connections in the Internet is transmission control
protocol (TCP). TCP provides network connections that provide
reliable first in, first out (FIFO) communication as well as fair
flow provisioning using adaptive congestion windows. Consider a
cloud service that sends commands to a large number of actuators.
The cloud service consists of a cluster of a few hundred servers.
To keep actuators simple, and also to allow flexibility in evolving
the cloud service, the cloud service should appear to the actuators
as a single end point with a single TCP/IP address. While an
actuator will receive commands from a particular server machine in
the cluster, it appears to the actuators (and their software) as if
the cloud service is a single, highly reliable, and fast machine.
It is desirable to maintain this illusion even when connections
migrate between server machines for load balancing, for hardware or
software upgrades, or when rebooting cloud servers. TCP
connections, unfortu- nately, are between socket end points, and,
using current operating systems abstrac- tions, socket end points
cannot migrate or survive process failures. Also, the cloud service
would have to maintain a TCP socket for every actuator. This does
not scale well, as each TCP socket involves storing a lot of state
information. Replacing TCP with a radically different protocol
would not be feasible today. Operating systems and even networking
hardware implement TCP connections very efficiently. TCP is the
dominant communication protocol on the Internet, and Internet
routers have evolved to support TCP efficiently, easily scaling to
many mil- lions of simultaneous TCP connections. TCP-R proposes to
support standard TCP connections, but to extend them with a
technology that addresses the shortcomings mentioned above. The
essential idea is to extend the cloud service with a filter that
intercepts and preprocesses TCP packets. The filter is scalable and
maintains little state per TCP connection (on the order of 32
bytes). It has only soft state (i.e., it does not have to store its
state persistently across crashes, greatly simplifying fault toler-
ance). The filter allows servers to migrate TCP connections, and
TCP connections
- 29. 12 Smart Grids to survive server failure and recovery.
Originally developed to maintain TCP con- nections between border
gateway protocol (BGP) (Internet routing) servers across failures
and subsequent recovery [23], TCP-R is extended into a scalable
technology for a cluster serving client end points and also to park
connections that are not currently live. 1.4.4 GRIDSIM GridSim is a
real-time, end-to-end power grid simulation package that is unique
in its integration of a real-time simulator, data delivery
infrastructure, and multiple applications all running in real time.
The goal of this project is to simulate power grid operation,
control, and communications on a grid-wide scale (e.g., the Western
Interconnection), as well as to provide utilities with a way to
explore new equipment deployments and predict reactions to
contingencies. The ability to simulate opera- tion under different
equipment deployment configurations includes large-scale con-
figurations of PMUs. With the objective of simulating real-world
equipment usage, and usage in conjunction with readily available
industry equipment, the GridSim simulation package uses the
industry standard C37.118 data format for all streaming measurement
data. The first element in the GridSim platform is a transient
power stability simula- tor, specially modified to output streaming
data in real time. The output data are encoded into C37.118 and
sent to a huge number of substation processes. At each of these
processes, the data are aggregated, as would be done in a real
power utility sub- station. The data are also sent to any of the
substation-level power applications that are running. Both the raw
substation data as well as any power application outputs are then
published to GridStat. GridStat allows the substation data to be
distributed as they would be in the real world. Published data can
be sent via redundant paths, 1many communication
(publish-subscribe, whose degenerate version is network-level
multicast), and so on. The flexibility provided by the GridStat
data delivery middleware allows subscrip- tion applications to be
easily integrated into the system with minimal reconfigura- tion.
Published data are available to any subscribers of GridStat,
including the two applications included in the GridSim simulation,
the hierarchical state estimator (HSE), and the oscillation and
damping monitor (Figure 1.3). 1.4.5 GRIDCLOUD ARCHITECTURE
GridCloud was designed with the expectation that the developers of
the advanced power grid will require easy access to large computing
resources. Tasks may require large-scale computation, or may
involve such large amounts of data that simply host- ing and
accessing the data will pose a substantial scalability challenge.
This leads us to believe that cloud computing will play an
important role in the future grid, supplementing the roles played
by existing data center architectures. The compel- ling economics
of cloud computing, the ease of creating apps that might control
household power consumption (not a subject that has been mentioned
yet), and the remarkable scalability of the cloud all support this
conclusion.
- 30. 13Mission-Critical Cloud Computing for Critical
Infrastructures Figure 1.4 shows the architecture of GridCloud. The
representative application used in this case is a HSE [18,24,25].
Data sources represent PMUs, which stream data to data collectors
across a wide area using GridStat. The HSE comprises several
substation-level state estimators that aggregate, filter, and
process PMU data before forwarding them to a control center-level
state estimator. The input and the first-level computation are
inherently sharded at substation granularity. Furthermore, compu-
tations are inherently parallel between substations. Thus, the HSE
has a natural map- ping in GridCloud with substation state
estimators residing in the outermost tier of the cloud while the
control center state estimator is moved to the interior tier. The
substation state estimators are replicated to increase fault
tolerance and availability. The consistency of the replicas is
managed through Isis2. TCP-R is used to provide fail-over
capabilities for connections across the cloud. 1.5 CONCLUSIONS This
chapter presents a roadmap of how cloud computing can be used to
support the computational needs of the advanced power grid. Todays
commercial cloud comput- ing infrastructure lacks the essential
properties required by power grid applications. These deficiencies
are explained and a cloud-based power grid application architec-
ture is presented which overcomes these difficulties using
well-known distributed system constructs. Furthermore, the GridSim
project, which instantiates this model, is presented as a case
study example. Control- level applications OpenPDC Oscillation
monitor State estimator Substation gateway Static data generator
C37.118 generator Measurement generator Powertech TSAT simulator
Substation SE Subs SE Sub SE Su SE Su SE Substation OM Substation
OM Substation OM Substation O Substation O Substation 1 Substation
N Simulated power system GridStat FE FE FE FE FE FESubstation-
level simulation FIGURE 1.3 The GridSim architecture. (From
Anderson, D., Zhao, C., Hauser, C., Venkatasubramanian, V., Bakken,
D., and Bose, A., IEEE Power and Energy Magazine, 10, 4957,
2012.)
- 31. 14 Smart Grids COL N M COL 2 M SKPM P2 Dataorigin P1 S2
TCP-R S1 COL 1 M COL N 2 COL 2 2 COL 1 2 COL N 1 COL 2 1 COL 1 1 FE
FE FE FE FE FE FE FE GridStat (UDP) S-SE1 M S-SE 1 2 S-SE 2 2 S-SE
N 2 S-SE 1 1 S-SE 2 1 EC2cloud GridCloudOutput Localhost S-SE N 1
S-SE 2 M S-SE N M S-SE N S-SE 2 S-SE 1 ControlcenterSE Localclient
and visualizer Computation Results Isis 2
FIGURE1.4GridCloudarchitecture.
- 32. 15Mission-Critical Cloud Computing for Critical
Infrastructures REFERENCES 1. G. Fulgoni, A. Lipsman and I.
Essling, State of the U.S. Online Retail Economy in Q4 2012,
February 2013. Available at: http://goo.gl/BCnbn. 2. P. Mell and T.
Grance, The NIST Definition of Cloud Computing: Recommendations of
the National Institute of Standards and Technology. The National
Institute of Standards and Technology, Gaithersburg, MD, 2011. 3.
A. W. Services, Amazon elastic compute cloud: User guide API
version 2013-02-01.
http://awsdocs.s3.amazonaws.com/EC2/latest/ec2-ug.pdf (accessed 3
June 2013). 4. R. Prodan and S. Ostermann, A survey and taxonomy of
infrastructure as a service and web hosting cloud providers, in
Proceedings of the 10th IEEE/ACM International Conference on Grid
Computing, October 1315, 2009, Banff, AB. 5. D. Novosel, V. Madani,
B. Bhargava, K. Vu and J. Cole, Dawn of the grid synchroniza- tion,
IEEE Power and Energy Magazine, 6(1), 4960, 2008. 6. S. Horowitz,
D. Novosel, V. Madani and M. Adamiak, System-wide protection, IEEE
Power and Energy Magazine, 6(September), 3442, 2008. 7. S. Gilbert
and N. Lynch, Brewers conjecture and the feasibility of consistent,
available, partition-tolerant web services, ACM SIGACT News, 33(2),
5159, 2002. 8. M. Armbrust, A. Fox, R. Griffith, A. D. Joseph, R.
Katz, A. Konwinski, G. Lee, et al., A view of cloud computing,
Communications of the ACM, 53(4), 5058, 2010. 9. M. Armbrust, A.
Fox, R. Griffith, A. D. Joseph, R. Katz, A. Konwinski, G. Lee, et
al., Above the clouds: A Berkeley view of cloud computing,
Department of Electrical Engineering and Computer Sciences,
University of California Berkeley, Report UCB/ EECS 28, 2009. 10.
K. Birman, L. Ganesh and R. V. Renesse, Running smart grid control
software on cloud computing architectures, in Proceedings of the
Computational Needs for the Next Generation Electric Grid Workshop,
Cornell University, April, 2011, New York. 11. A. Iosup, S.
Ostermann, M. Yigitbasi, R. Prodan, T. Fahringer and D. H. J.
Epema, Performance analysis of cloud computing services for
many-tasks scientific computing, IEEE Transactions on Parallel and
Distributed Systems, 22(6), 931945, 2011. 12. K. Birman, A. Bose,
D. Bakken and C. Hauser, GridControl: A software platform to
support the smart grid, A Cornell University and Washington State
University Research Collaboration, 2011. Available at:
http://www.cs.cornell.edu/Projects/gridcontrol/index.
html#gridcloud (accessed 3 June 2013). 13. H. Gjermundrod, D.
Bakken, C. Hauser and A. Bose, GridStat: A flexible QoS-managed
data dissemination framework for the power grid, IEEE Transactions
on Power Delivery, 24(1), 136143, 2009. 14. C. Hauser, D. Bakken
and A. Bose, A failure to communicate: Next generation commu-
nication requirements, technologies, and architecture for the
electric power grid, IEEE Power and Energy Magazine, 3(2), 4755,
2005. 15. K. P. Birman, D. A. Freedman, Q. Huang and P. Dowell,
Overcoming CAP with consis- tent soft-state replication, IEEE
Computer, 45(2), 5058, 2012. 16. K. Birman, Isis2 cloud computing
library, 2013. Available at: http://isis2.codeplex.com/. 17. A.
Agapi, K. Birman, R. M. Broberg, C. Cotton, T. Kielmann, M.
Millnert, R. Payne, R. Surton and R. van Renesse, Routers for the
cloud: Can the Internet achieve 5-nines availability?, IEEE
Internet Computing, 15, 7277, 2011. 18. D. Anderson, C. Zhao, C.
Hauser, V. Venkatasubramanian, D. Bakken and A. Bose, A virtual
smart grid: Real-time simulation for smart grid control and
communications design, IEEE Power and Energy Magazine, 10(1), 4957,
2012. 19. D. E. Bakken, A. Bose, C. H. Hauser, D. E. Whitehead and
G. C. Zweigle, Smart gen- eration and transmission with coherent,
real-time data, Proceedings of the IEEE, 99(6), 928951, 2011.
- 33. 16 Smart Grids 20. E. Solum, C. Hauser, R. Chakravarthy and
D. Bakken, Modular over-the-wire configurable security for
long-lived critical infrastructure monitoring systems, in
Proceedings of the Third ACM International Conference on
Distributed Event-Based Systems (DEBS), Nashville, TN, 2009. 21. K.
Birman, Reliable Distributed Systems: Technologies, Web Services,
and Applications, Springer, New York, 2005. 22. L. Lamport, Paxos
made simple, ACM SIGACT News, 32(4), 1825, 2001. 23. R. Surton, K.
Birman, R. Broberg, T. Marian and R. van Renesse, Man-in-the-Middle
TCP recovery, Cornell University. Available at: http://goo.gl/9S4Jz
(accessed 3 June 2013). 24. W. Jiang, V. Vittal and G. Heydt, A
distributed state estimator utilizing synchronized phasor
measurements, IEEE Transactions on Power Systems, 22(2), 563571,
2007. 25. M. Zhao and A. Abur, Multi area state estimation using
synchronized phasor measure- ments, IEEE Transactions on Power
Systems, 20(2), 611617, 2005.
- 34. 17 2 Power Application Possibilities with Mission-Critical
Cloud Computing David Bakken, Pranavamoorthy Balasubramanian,
Thoshitha Gamage, Santiago Grijalva, Kory W. Hedman, Yilu Liu,
Vaithianathan Venkatasubramanian, and Hao Zho CONTENTS
2.1Overview..........................................................................................................
18 2.2 Robust Adaptive Topology
Control.................................................................
18
References.................................................................................................................19
2.3 Adaptive Real-Time Transient Stability
Controls............................................20
Reference..................................................................................................................21
2.4 Prosumer-Based Power
Grid............................................................................21
2.4.1Introduction.........................................................................................
21 2.4.2 Prosumer-Based Control
Architecture................................................22
2.4.3 Computational
Challenges...................................................................23
2.4.4 Cloud Computing in the Future Electric
Grid.....................................23 2.4.5 Economic Dispatch
of Stochastic Energy Resources..........................23 2.4.6
Cloud-Based
Apps...............................................................................24
2.4.7 Scenario Analysis and Transmission
Planning...................................24 2.4.8 Model
Integration................................................................................24
2.4.9 Exploiting and Abstracting
Self-Similarity.........................................25
References.................................................................................................................25
2.5 Wide-Area Frequency
Monitoring..................................................................26
2.5.1Introduction.........................................................................................26
2.5.2 FNET
Architecture..............................................................................26
2.5.3 FNET
Applications..............................................................................27
2.5.4 Cloud Computing and
FNET...............................................................27
2.5.5 Rapidly Elastic Data
Concentrators.....................................................27
2.5.6 Computational Requirement
Flexibility..............................................28 2.5.7
Cloud-Based
Applications...................................................................28
References.................................................................................................................29
- 35. 18 Smart Grids 2.1OVERVIEW As we have seen in Chapter 1,
not only is cloud computing coming to the grid, but
mission-critical implementations such as GridCloud can provide
mission-critical properties. This chapter explores new applications
enabled by such technology. While this chapter is only scratching
the surface of what is likely to be routine in a decade, we hope
that it provides a tantalizing glimpse of what is possible. What,
then, is mission-critical cloud computing? To recap, in a nutshell,
it Keeps the same fast throughput as generic commercial cloud
platforms Does not deliberately trade off this throughput to allow
inconsistencies, for example, when a replica does a state update on
a copy of the state but this update is forgotten Is much more
predictable (and faster) in terms of ramp-up time, central
processing unit (CPU) performance per node, and number of nodes
Therefore, the question for power application developers is how
they can use: Hundreds of processors in steady state. Thousands or
tens of thousands of processors when a contingency is reached or is
being approached. Note: often there are many minutes of advanced
warning of this, sometimes an hour or more. Data from all
participants in a grid that is enabled quickly when a crisis is
approached (though, for market reasons, not necessarily during
steadystate). With this in mind, we now present groundbreaking
applications that can exploit such mission-critical cloud
platforms. 2.2 ROBUST ADAPTIVE TOPOLOGY CONTROL Balasubramanian and
Hedman The electric power transmission system is one of the most
complex systems available today. Traditionally, bulk power
transmission systems (lines and transformers) are treated as static
assets, even though these resources are controllable. However, it
is known that transmission topology control has been used in the
past and is still being used for corrective-based applications; for
example, PJM uses corrective topology control as a special
protection scheme (SPS) [1]. These switching actions are primarily
taken on an ad hoc basis, determined by the system operators based
on past historical data rather than in an automated way based on
decision support tools. Past research has demon- strated the
ability of topology control to help improve voltage profiles,
increase transfer capacity, improve system reliability, and provide
cost benefits to the system [28]. Even 2.6 Oscillation Mitigation
Strategies.....................................................................29
References.................................................................................................................30
2.7Automatic Network Partitioning for Steady-State
Analysis............................30
References.................................................................................................................
31
- 36. 19Power Application Possibilities with Mission-Critical
Cloud Computing though transmission topology control can provide
these benefits, harnessing such flex- ibility from the transmission
network in existing operational procedures is limited due to the
computational challenges of optimizing the transmission topology.
More recently, sensitivity-based methods have been proposed as a
mechanism to reduce the computational complexity [912]. The robust
adaptive topology control method develops a sensitivity-based
heuristic, which reduces the computational time of the topology
control problem. An expression is derived indicating the impact of
changing the state of a transmission line on the objective. This
expression is used to generate a line-ranking system with the
potential candidate lines for switching based on a direct current
(dc) optimal power flow, which builds on the work of [12]. This
approach selects a single feasible switching action per iteration,
which provides an improvement to the system. The advantage of this
method is that it solves linear programs iteratively to come up
with a beneficial line-switching solution, which is computationally
simple as compared with other methods employing mixed integer
programming. All the possible switching solutions are lined up in
the ranked list, with the switching action most likely to be
beneficial placed at the top of the list. As the list is formed
based on a sensitivity study, the switching action is not
guaranteed to improve the system. Hence, the switching actions need
to be checked for alternating current (ac) feasibility and whether
they truly provide an improvement in the objective before they are
implemented. This is done by selecting the first action from the
ranked list and simulating the switching to find the improvement in
the system. If the switch- ing is not beneficial, the next action
in the ranked list is checked for improvement. This process is
continued until a beneficial switching action is found. While such
a procedure is a heuristic, prior work has shown substantial
economic savings [9] as well as strong performance in comparison
with global optimization techniques [12]. The processing time taken
to come up with a beneficial switching action could be
significantly reduced if this process were parallelized so that all
the proposed switching solutions could be checked at once. This
opens up enormous opportunities for the application of cloud
computing to transmission-switching applications, which would
drastically reduce the computational time and improve the solution
quality, as the best solution from the ranking list could be
identified very quickly. With prior research demonstrating cost
savings of close to 4% for a $500 billion industry [3], there is a
great opportunity for advanced decision support tools to fill this
tech- nological need, in terms of both algorithm sophistication and
advanced computing capabilities, such as cloud computing.
REFERENCES 1. PJM, Manual 3: Transmission Operations, Revision: 40,
2012. Available at: http://www.
pjm.com/~/media/documents/manuals/m03.ashx. 2.W. Shao and V.
Vittal, Corrective switching algorithm for relieving overloads and
volt- age violations, IEEE Transactions on Power Systems, 20(4),
18771885, 2005. 3. K. W. Hedman, M. C. Ferris, R. P. ONeill, E. B.
Fisher, and S. S. Oren, Co-optimization of generation unit
commitment and transmission switching with N-1 reliability, IEEE
Transactions on Power Systems, 25(2), 10521063, 2010. 4. K. W.
Hedman, R. P. ONeill, E. B. Fisher, and S. S. Oren, Optimal
transmission switching with contingency analysis, IEEE Transactions
on Power Systems, 24(3), 15771586, 2009.
- 37. 20 Smart Grids 5. A. Korad and K. W. Hedman, Robust
corrective topology control for system reliability, IEEE
Transactions on Power Systems, 28(4), 40424051, 2013. 6. K. W.
Hedman, R. P. ONeill, E. B. Fisher, and S. S. Oren, Smart flexible
just-in-time trans- mission and flowgate bidding, IEEE Transactions
on Power Systems, 26(1), 93102, 2011. 7.E. B. Fisher, R. P. ONeill,
and M. C. Ferris, Optimal transmission switching, IEEE Transactions
on Power Systems, 23(3), 13461355, 2008. 8.K. W. Hedman, S. S.
Oren, and R. P. ONeill, A review of transmission switching and
network topology optimization, in Proceedings of IEEE Power and
Energy Society General Meeting, July 2011, Detroit, MI. 9.P. A.
Ruiz, J. M. Foster, A. Rudkevich, and M. C. Caramanis, On fast
transmission topology control heuristics, in Proceedings of IEEE
Power and Energy Society General Meeting, July 2011, Detroit, MI.
10. J. M. Foster, P. A. Ruiz, A. Rudkevich, and M. C. Caramanis,
Economic and corrective applications of tractable transmission
topology control, in Proceedings of 49th Annual Allerton Conference
on Communication, Control, and Computing, pp. 13021309, September
2011, Monticello, IL. 11.P. A. Ruiz, J. M. Foster, A. Rudkevich,
and M. C. Caramanis, Tractable transmission topology control using
sensitivity analysis, IEEE Transactions on Power Systems, 27(3),
15501559, 2012. 12.J. D. Fuller, R. Ramasra, and A. Cha, Fast
heuristics for transmission line switching, IEEE Transactions on
Power Systems, 27(3), 13771386, 2012. 2.3 ADAPTIVE REAL-TIME
TRANSIENT STABILITY CONTROLS Venkatasubramanian The power system is
expected to undergo major changes in the next decade, resulting
from rapid growth in system loads (such as electric cars) and from
increased depen- dence on renewable intermittent generation. To
face up to these challenges, power utilities are making major
upgrades to wide-area monitoring and control technolo- gies, with
impetus from major federal investments in the past few years. Power
system operation is designed to withstand small- and large-scale
distur- bances. However, when the system is subjected to several
large disturbances in a short span of time, it may become
vulnerable to blackouts. Some recent events, such as the 2012 San
Diego blackout and the 2003 Northeastern blackout, point to the
need for adaptive real-time transient stability control designs
that are specifically designed on an adaptive premise of making
control decisions during the evolution of the event. In the
present-day power system, wide-area transient stability controls
such as reme- dial action schemes (RAS) or SPS are hard-coded
control algorithms that are triggered by a central controller in
response to the occurrence of specific contingencies based on
preset switching logic. When the system is subject to any unknown
set of contingen- cies that is not part of the RAS controller
logic, the system operation typically switches to a safe mode
whereby interarea power transfers are limited to low conservative
settings. The tie-line transfers remain at safe low values until
the reliability coordinator completes a new set of transient
stability simulation studies, which results in significant economic
losses due to operation at nonoptimal power transfer levels. Cloud
computing emerges as an ideal platform for handling transient
stabil- ity mitigation issues, both for the present-day power
system and for future control designs. In present-day operation,
whenever the system operation is found to be in
- 38. 21Power Application Possibilities with Mission-Critical
Cloud Computing one of the unknown operating conditions, the
reliability coordinator can dial in a vast amount of cloud-based
processing power to carry out the massive number of new transient
stability simulations needed for determining the safe transfer
limits. In the future, we need to rethink the design of transient
stability controls such as RAS or SPS schemes. The massive
computational capability offered by cloud com- puting opens up
truly novel futuristic control schemes for mitigating transient
stabil- ity events, as proposed in [1]. In the present-day power
system, simulation studies are performed off-line for a
guesstimated list of potential contingencies, and RAS schemes are
implemented for a subset of problematic N2 or higher-order
contingen- cies whenever needed. Such RAS schemes, then, only work
for a limited number of potential scenarios. Moreover, the
respective control actions in these RAS schemes are also designed
to be conservative, being based on off-line studies. Zweigle and
Venkatasubramanian [1] propose to select and implement transient
sta- bility controls based on simulations of the system in real
time during the evolution of the events themselves. Wide-area
monitoring from an abundance of phasor measurement units (PMU) in
the future will pave the way for real-time monitoring of the state
and sys- tem topology of the full-size power system. Combining this
real-time state information with real-time simulations will allow
us to evaluate which control actions are optimally suited to the
system at the present time, and the decisions are fully adaptive to
what- ever the system conditions are. Since the controller
continues to monitor the system in a closed-loop fashion, the
proposed control schemes are also robust with respect to simula-
tion errors and communication or actuation failures. The
formulation is not restrictive to any subset of contingencies, and
can handle low-probability events consisting of multiple outages,
such as those that have served as precursors to large blackouts in
the past. In this proposed formulation, denoted as adaptive
real-time transient stability con- trols, massive processing power
is needed to carry out what if simulations of many potential
control candidates in parallel before deciding on whether any
control action is needed and which specific action(s) will be
implemented. The system monitoring and simulations of what if
scenarios will continue throughout the event until the system has
been stabilized. Once the controller recognizes that the system has
returned to its normal state, the controller returns to dormant
system monitoring mode, and cloud resources can be released.
Details of the control algorithms can be seen in [1]. REFERENCE
1.G. Zweigle and V. Venkatasubramanian, Wide-area optimal control
of electric power systems with application to transient stability
for higher order contingencies, IEEE Transactions on Power Systems,
28(3), 23132320, 2013. 2.4 PROSUMER-BASED POWER GRID Gamage and
Grijalva 2.4.1 Introduction The electric power grid, in a bid to
improve its sustainability, is aggressively explor- ing ways to
integrate distributed renewable energy generation and storage
devices
- 39. 22 Smart Grids at many levels. The most obvious integration
is at the level of generation, where renewable generation sources
such as large wind turbine and solar panel farms will supplement
and eventually (it is hoped) supplant traditional nonrenewable
power generation sources. Another natural integration is at the
distribution level, where relatively smaller-scale renewable energy
generation by utilities and other power dis- tribution entities
offers cheaper and greener energy options to customers. While not
on the same bulk scale as the generation or distribution level, an
emerging trend in recent years is for end consumers who are
typically below the distribution level (e.g.,households,
microgrids, and energy buildings) to generate their own power using
renewable sources and become self-sustainable and
energy-independent of the grid. A fascinating aspect of this
changing energy landscape is the drastic changes in the roles of
the players involved. For example, end consumers, in addition to
their typical energy consumption role, are economically motivated
to sell excess energy and provide energy storage services to the
grid. Modern utilities also go beyond their traditional energy
distribution role in buying energy from end consumers when
available. Similar role augmentations can be observed at all levels
of the modern electric power grid [1]. As a consequence, what
traditionally has been a one-way energy transferfrom bulk
generation, through transmission and distribution, to end-user
consumptionis transfor